61811 v1 BACKGROUND PAPER 5 (PHASE II) Powering Up: Costing Power Infrastructure Spending Needs in Sub-Saharan Africa Orvika Rosnes and Haakon Vennemo Econ Pöyry, in association with Norplan and Power Planning Associates MARCH 2009 © 2009 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 USA Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved A publication of the World Bank. The World Bank 1818 H Sreet, NW Washington, DC 20433 USA The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the Executive Directors of the International Bank for Reconstruction and Development / The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923 USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street, NW, Washington, DC 20433 USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. About AICD This study is a product of the Africa Infrastructure Country Diagnostic (AICD), a project designed to expand the world’s knowledge of physical infrastructure in Africa. AICD will provide a baseline against which future improvements in infrastructure services can be measured, making it possible to monitor the results achieved from donor support. It should also provide a better empirical foundation for prioritizing investments and designing policy reforms in Africa’s infrastructure sectors. AICD is based on an unprecedented effort to collect detailed economic and technical data on African infrastructure. The project has produced a series of reports (such as this one) on public expenditure, spending needs, and sector performance in each of the main infrastructure sectors—energy, information and communication technologies, irrigation, transport, and water and sanitation. Africa’s Infrastructure—A Time for Transformation, published by the World Bank in November 2009, synthesizes the most significant findings of those reports. AICD was commissioned by the Infrastructure Consortium for Africa after the 2005 G-8 summit at Gleneagles, which recognized the importance of scaling up donor finance for infrastructure in support of Africa’s development. The first phase of AICD focused on 24 countries that together account for 85 percent of the gross domestic product, population, and infrastructure aid flows of Sub- Saharan Africa. The countries are: Benin, Burkina Faso, Cape Verde, Cameroon, Chad, Côte d'Ivoire, the Democratic Republic of Congo, Ethiopia, Ghana, Kenya, Lesotho, Madagascar, Malawi, Mozambique, Namibia, Niger, Nigeria, Rwanda, Senegal, South Africa, Sudan, Tanzania, Uganda, and Zambia. Under a second phase of the project, coverage is expanding to include as many other African countries as possible. Consistent with the genesis of the project, the main focus is on the 48 countries south of the Sahara that face the most severe infrastructure challenges. Some components of the study also cover North African countries so as to provide a broader point of reference. Unless otherwise stated, therefore, the term “Africa” will be used throughout this report as a shorthand for “Sub-Saharan Africa.” The World Bank is implementing AICD with the guidance of a steering committee that represents the African Union, the New Partnership for Africa’s Development (NEPAD), Africa’s regional economic communities, the African Development Bank, the Development Bank of Southern Africa, and major infrastructure donors. Financing for AICD is provided by a multidonor trust fund to which the main contributors are the U.K.’s Department for International Development, the Public Private Infrastructure Advisory Facility, Agence Française de Développement, the European Commission, and Germany’s KfW Entwicklungsbank. The Sub-Saharan Africa Transport Policy Program and the Water and Sanitation Program provided technical support on data collection and analysis pertaining to their respective sectors. A group of distinguished peer reviewers from policy-making and academic circles in Africa and beyond reviewed all of the major outputs of the study to ensure the technical quality of the work. The data underlying AICD’s reports, as well as the reports themselves, are available to the public through an interactive Web site, www.infrastructureafrica.org, that allows users to download customized data reports and perform various simulations. Inquiries concerning the availability of data sets should be directed to the editors at the World Bank in Washington, DC. Contents Acknowledgments v Abbreviations and acronyms v Summary vi A model to inform energy policy decisions vi A high price tag viii What is the cost of expanding electrification? xi How sensitive are power investments to economic growth? xii Why trade power? xii What patterns of trade would emerge? xiii Who gains most from power trade? xv How will less hydropower development influence the trade flows? xvi What are the environmental impacts of trading power? xvii How would CDM affect generation technology choices? xvii How might climate change affect power investment patterns? xviii 1 Introduction 1 2 Estimating demand 3 Market demand 3 Suppressed demand 7 Social demand 9 Total demand in 2015 17 3 Estimating supply 19 Generation capacity 19 Transmission and distribution 23 Supply-side assumptions related to social demand 27 4 Least-cost expansion model 30 Model assumptions and simplifications 30 The mathematical formulation of the model 32 5 Results: Spending needs through 2015 36 Maintaining electricity access at 2005 rates 37 Cost effects of raising access rates 38 The effects of trade 38 SAPP 40 EAPP 50 WAPP 61 CAPP 71 CO2 emissions and their reduction in Sub-Saharan Africa 80 6 Sensitivity analyses 82 Low economic growth 82 The impact of fuel prices in EAPP 83 Climate change and less reliable hydropower in EAPP 86 Clean Development Mechanism (CDM) in the SAPP 87 Barriers to hydropower development in WAPP: Less hydropower in Guinea 90 Unit investment cost of coal-fired power generation in SAPP 91 iii Unit investment cost of hydropower in WAPP 92 More imports from the Democratic Republic of Congo to CAPP 92 References 94 Appendix 1Documentation of asset stock and refurbishment needs of thermal power generation 96 CAPP Region 96 EAPP – Nile basin region 98 SAPP region 105 WAPP region 110 Island states 117 Appendix 2 Documentation of hydropower refurbishment needs and investments 118 Hydropower characteristics 118 Determination of potential production capacity of new projects 119 Operation, refurbishment and reconstruction 120 Sources of data 120 Data sets 121 Investment plans and costs for hydropower projects 122 Refurbishment needs and costs of existing hydropower projects 131 Appendix 3 Documentation of asset stock and refurbishment needs of transmission and distribution 137 Assumptions and approximations 139 CAPP: Country-specific assumptions and unit costs 143 EAPP – Nile Basin: Country-specific assumptions and unit costs 145 SAPP: Country-specific assumptions and unit costs 146 WAPP: Country-specific assumptions and unit costs 148 Island states: Country-specific assumptions and unit costs 150 Unit investment costs outputs and analysis 150 Refurbishment of existing structures 152 Maintenance of existing and new infrastructure 154 Conclusion 155 Appendix 4 Comparison of Econ Pöyry cost estimates with those of Nexant study 156 Different methodologies: generic least cost model vs. costs related to existing plans 156 Differences in the treatment of trade 156 Different scope: transmission, distribution and connection costs 157 Comparison of results 157 iv COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Acknowledgments The authors wish to thank the following individuals for excellent research assistance: Halvor Aarrestad, Ryan Anderson, Sarah Carter, Jonathan Hedgecock, Sudeshna Ray, Ståle Skrede, and Fredrik Wilander. We also thank Vivien Foster of the World Bank for her patience and astute comments, and Steven Kennedy for improving the language. Their efforts significantly improved the report. Abbreviations and acronyms African Development Bank ADB Agence Française de Développement AFD Africa Infrastructure Country Diagnostic AICD Central Africa Power Pool CAPP Clean Development Mechanism CDM certified emission reduction credits CERs Nile Basin–East Africa Power Pool EAPP Energy Information Administration EIA gross domestic product GDP gigawatt GW gigawatt-hour GWh heavy fuel oil HFO high voltage HV International Energy Agency IEA independent power producer IPP kilovolt kV kilowatt-hour kWh long-run marginal cost LRMC low voltage LV megawatt MW New Partnership for Africa’s Development NEPAD operations and maintenance O&M Southern Africa Power Pool SAPP solar photovoltaic energy solar PV transmission and distribution T&D terawatt-hour TWh United States Agency for International Development USAID United States Department of Agriculture USDA Western Africa Power Pool WAPP v COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Summary S ub-Saharan Africa will require substantial investments in the power sector—on the order of 4 percent of the region’s gross domestic product (GDP) annually before 2015—if it is to meet the demands of economic development, keep pace with population growth, and expand electrification beyond the 2005 regional average of just 34 percent. Developing a regional power-trading market that exploits the vast hydropower potential of the subcontinent may be the best way to bring those costs down while also protecting against increases in oil prices and curbing carbon emissions. Expanding electrification is a daunting challenge, but the costs associated with extending the transmission network are minor in comparison with the investments in generation needed to accompany the demand of Africa’s growing economies. A model to inform energy policy decisions Nowhere in the world is the gap between available energy resources and access to electricity greater than in Sub-Saharan Africa. The region as a whole is rich in oil, gas, and hydropower potential, yet 66 percent of its population lacks access to electricity, with coverage especially low in rural areas. National authorities and international organizations have drawn up plans to increase access, but key policy choices underpin these plans. Which type of power generation is right in which settings? Should individual countries move ahead independently, or should they aim for coordinated development? What are the benefits of regional trade in power, and who are the main beneficiaries? How should major global trends, such as rising oil prices and looming climate change, affect decisions about power generation in Africa? How rapidly can Africa electrify? How sensitive are power investment decisions to broader macroeconomic conditions? To answer these questions, we developed a model to analyze the costs of expanding the power sector over the course of 10 years under different assumptions. The model simulates optimal (least-cost) strategies for generating, transmitting, and distributing electricity in response to demand increases in 43 countries of Sub-Saharan Africa, grouped into four power pools. The Southern Africa Power Pool (SAPP) consists of Angola, Botswana, Democratic Republic of Congo, Lesotho, Mozambique, Malawi, Namibia, South Africa, Zambia, and Zimbabwe. Within SAPP, South Africa clearly occupies a dominant position, accounting for 80 percent of overall power demand. The Nile Basin–East Africa Power Pool (EAPP) consists of Burundi, Djibouti, Egypt, Ethiopia, Kenya, Rwanda, Sudan, Tanzania, and Uganda. Here, Egypt is the driving force, accounting for 70 percent of power demand within EAPP. The Western Africa Power Pool (WAPP) consists of Benin, Burkina Faso, Côte d’Ivoire, Gambia, Ghana, Guinea, Guinea- Bissau, Liberia, Mali, Mauritania, Niger, Nigeria, Senegal, Sierra Leone, and Togo. In WAPP, Nigeria dominates, with two-thirds of electricity consumption in the region. The Central Africa Power Pool (CAPP) consists of Cameroon, the Central African Republic, Chad, the Republic of Congo, Equatorial Guinea, and Gabon. In CAPP, the Republic of Congo and Cameroon are the major players, sharing 90 vi COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA percent of power demand. Finally, Cape Verde, Madagascar, and Mauritius are included in our study as island states. The exercise begins with a projection of power demand over 10 years, 2005–15. Demand consists of (a) market demand associated with different levels of economic growth, structural change and population growth; (b) suppressed demand created by frequent blackouts and ubiquitous power rationing; and (c) social demand, as expressed in political targets for increasing access to electricity. Based on historic trends, demand is projected to grow at 5 percent per year in Sub-Saharan Africa, reaching levels of 680 terawatt-hours (TWh) by 2015. Demand is projected to grow at 4–5 percent per year in SAPP and EAPP to reach levels of 400 and 170 TWh, respectively. The other regions have even higher electricity demand growth: 7 percent per year in CAPP, 9 percent per year in the island states and 12 percent per year in WAPP. The absolute demand levels, however, are lower in these regions: 20 TWh in CAPP, 3 TWh in the island states combined, and about 100 TWh in WAPP. In all cases except the islands, market demand accounts for the great bulk of demand growth over the period. The model then looks for the least costly way of meeting the new demand based on investments in electricity generation, transmission, and distribution. Those investments include refurbishment of existing capacity for electricity generation and construction of new capacity for cross-border electricity transmission. Our analysis covers four modes of thermal generation (natural gas, coal, heavy fuel oil, and diesel) and four renewable generation technologies (large hydropower, mini-hydro, solar photovoltaic [PV], and geothermal). Minihydro, diesel, and solar PV are off-grid alternatives; that is, they are not connected to the central power grid. Operation, but not new investment, of current nuclear power is considered. The main value of the model is that it can be run under a number of different scenarios to highlight the implications of various policies. For example, by comparing a “trade-stagnation” scenario under which no further cross-border transmission capacity is built with a trade-expansion scenario under which all economically viable cross-border transmission capacity is developed, we can quantify the gains from trade. The model can also be used to evaluate the feasibility of alternative electrification targets, ranging from maintaining constant access rates, to raising electrification to a uniform level over ten years, to pursuing a range of national electrification targets. The impact of higher oil prices, higher investment costs and lower rainfall can be gauged through their effects on the relative cost of different generation technologies, while the consequences of slower economic growth on power sector investment needs can also be readily quantified. vii COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA A high price tag How much will it cost to meet market demand for power in 2105 while eliminating power shortages and achieving national policy targets for access to electricity? It is clear that these achievements will require substantial investments in the power sector, demanding about 82,000 MW new generation capacity in total. This entails almost a doubling of current capacity, which for the whole of Sub-Saharan Africa stands at 87,000 MW (2005 data). Since many power installations in Africa are old, much of the capacity operational in 2005 needs to be refurbished before 2015. In the SAPP region, a 2005 capacity of 48,000 MW is expected to be reduced to 17,000 MW, and some 28,000 MW of generating capacity will have to be refurbished (Table A, column “National targets for access rates”). In addition, more than 33,000 MW of new generating capacity will have to be built, an increase of about 70 percent over the 2005 level. In EAPP, the needs for refurbishment are minimal, but 26,000 MW of new generation will be required, essentially doubling the installed capacity of the region. The investment requirements are even larger in WAPP and CAPP: 18,000 MW new capacity will have to be built in WAPP, corresponding to 180 percent of current capacity, while CAPP requires investments of more than 2.5 times 2005 capacity, or 4,400 MW in total. More than half of current capacity must be refurbished both in WAPP and CAPP (7,000 MW and 900 MW, respectively). It is clear that each region, particularly West and Central Africa, require significant investment. The good news is that economic growth drives most demand. Therefore, at least according to the projections, the financial strength to finance investments should emerge alongside new investment needs. The annualized capital investment costs in Sub-Saharan Africa are 2.2–2.4 percent of the region’s GDP in 2015. There is, however, considerable variation between the different regions. The annualized capital investment costs are 2 percent of GDP for the SAPP region, and 2 to 3 percent of GDP for the EAPP and WAPP regions, but below 2 percent in CAPP (table B). The costs of operating the entire power system are similar to investment costs, around 1.7–2.1 percent of GDP in total. The variation between regions is even more pronounced here: the costs are just under 2 percent of GDP for the SAPP region, and about 3 percent of GDP in the EAPP region, while they are about 1.5 percent in WAPP and a negligible amount in CAPP (0.2–0.4 percent of GDP). Thus, total spending amounts to 4.2–4.4 percent of GDP in Sub-Saharan Africa. The total spending is about the same magnitude in SAPP and WAPP, but around 6 percent of GDP in EAPP and 2 percent in CAPP. Around two-thirds of overall system costs are associated with generation infrastructure, and the remaining third with transmission and distribution infrastructure. viii COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table A Generating capacity in Sub-Saharan Africa in 2015, under various trade, access, and growth scenarios Trade-stagnation Low-growth Trade-expansion scenario scenario scenario National targets Regional target National targets National targets for access rates, Generation capacity (MW) 2005 access rate access rate for access rates for access rates trade expansion Southern Africa Power Pool Installed capacity a 17,136 17,136 17,136 17,136 17,136 Refurbished capacity 28,029 28,035 28,046 28,148 28,046 New capacity 31,297 32,168 33,319 32,013 20,729 Hydropower share (%) 33 33 34 25 40 Eastern Africa Power Pool Installed capacity a 22,132 22,132 22,132 22,132 22,132 Refurbished capacity 1,369 1,375 1,375 1,381 1,375 New capacity 23,045 24,639 25,637 17,972 23,540 Hydropower share (%) 49 47 46 28 48 Western Africa Power Pool Installed capacity a 4,096 4,096 4,096 4,096 4,096 Refurbished capacity 5,530 6,162 6,972 6,842 5,535 New capacity 15,979 16,634 18,003 16,239 17,186 Hydropower share (%) 82 79 77 73 80 Central Africa Power Pool Installed capacity a 260 260 260 260 260 Refurbished capacity 906 906 906 1,081 906 New capacity 3,856 4,143 4,395 3,833 3,915 Hydropower share (%) 97 97 97 83 97 Island states Installed capacity a 282 282 282 282 282 Refurbished capacity 83 83 83 83 83 New capacity 189 369 368 368 353 Hydropower share (%) 25 19 19 19 20 Total Sub-Saharan Africa Installed capacity a 43,906 43,906 43,906 43,906 43,906 Refurbished capacity 35,917 36,561 37,382 37,535 35,945 New capacity 74,366 77,953 81,722 70,425 65,723 Hydropower share (%) 48 47 47 36 52 a. “Installed capacity” refers to installed capacity as of 2005 that will not undergo refurbishment before 2015. Existing capacity that will be refurbished before 2015 is not included in the installed capacity figure, but in the refurbishment figure. ix COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table B Estimated annualized cost of meeting power needs of Sub-Saharan Africa under two trade scenarios (national targets for electricity access) US$ billions and % of GDP Southern Eastern Western Central Africa Island states Total Sub- Africa Power Africa Power Africa Power Power Pool Saharan Pool Pool Pool Africa Trade expansion Total estimated cost 18.4 3.7% 15 5.7% 12.3 4.2% 1.4 2.0% 0.6 3.1% 47.6 4.2% of which total  Capital costs 10.0 2.0% 8.2 3.1% 8.2 2.8% 1.2 1.8% 0.2 1.4% 27.9 2.4%  Operating costs 8.4 1.7% 6.8 2.6% 4.0 1.4% 0.2 0.2% 0.3 1.7% 19.7 1.7% of which total  Generation 11.1 2.2% 10.5 4.0% 6.5 2.2% 1.0 1.4% 0.4 2.0% 29.5 2.6%  Transmission and distribution 7.3 1.5% 4.5 1.7% 5.8 2.0% 0.4 0.6% 0.2 1.1% 18.1 1.6% Trade-stagnation Total estimated cost 19.5 3.9% 16 6.0% 12.7 4.4% 1.5 2.2% 0.6 3.1% 50.3 4.4% of which total  Capital costs 10.0 2.0% 6.3 2.4% 8.0 2.7% 1.1 1.6% 0.2 1.4% 25.6 2.2%  Operating costs 9.4 1.9% 9.7 3.7% 4.8 1.6% 0.4 0.6% 0.3 1.7% 24.7 2.2% of which total  Generation 12.6 2.5% 11.6 4.4% 7.1 2.4% 1.2 1.7% 0.4 2.0% 32.8 2.9%  Transmission and distribution 6.9 1.4% 4.4 1.7% 5.7 1.9% 0.3 0.5% 0.2 1.1% 17.5 1.5% Note: Subtotals may not add up to the totals because of rounding. The overall cost of developing the power system appears high, but not unattainable relative to the GDP of each regional trading area. But both GDP and power investment requirements are very unevenly distributed within the regional pools. As a result, under certain scenarios, some countries face power spending requirements that are burdensome relative to the size of their economies (figure A). In SAPP, depending on the electrification target and other variables, spending requirements may exceed 6 percent of GDP in the Democratic Republic of Congo, Mozambique, and Zimbabwe. In EAPP, countries such as Egypt, Burundi, and Ethiopia may require similar levels of spending. About half of the countries in WAPP have investment requirements of almost 10 percent of GDP, and Guinea and Liberia stand out at almost 30 percent. In CAPP, only the Republic of Congo requires investments of more than 5 percent of GDP. Some of these countries have the potential to become major exporters of power, provided they receive cross-border injections of capital to develop their power infrastructure. The necessary capital is not likely to materialize, however, unless trade in power expands. x COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure A Overall power spending needed to reach national targets for electricity access under alternative trade scenarios by country % of GDP in 2015 (a) Southern Africa Power Pool (b) Eastern Africa Power Pool (c) Western Africa Power Pool (d) Central Africa Power Pool What is the cost of expanding electrification? We considered the impact of raising electrification levels from 2005 access levels to a uniform level across each region or to the levels specified in national electrification targets. The regional target levels roughly add 1 percent access every year over a ten-year period. Due to relatively low power consumption by households, the impact of expanding electrification is quite modest. For instance, if national access targets are reached across Sub-Saharan Africa, the region’s power generating requirement (in terms of MW) will only increase 10 percent, or $4 billion per year. The development of transmission and distribution networks, however, would require significant additional investment amounting to $5 billion per year across the region. The cost of transmission and distribution for access is particularly high in EAPP. As a result, raising electrification levels to meet national electrification targets would entail a commitment of $9 billion per year, or 0.8 percent of GDP of Sub-Saharan Africa. There are, however, xi COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA regional differences. In SAPP and CAPP the cost would be 0.3 and 0.4 percent of GDP, compared to 1 percent in WAPP and as much as 1.5 percent in EAPP. The 0.8 percent of GDP needed to increase access is included in the 4.2 percent estimate that we gave above on the cost of providing electricity for market and social needs. This shows that the majority of electricity needs is driven by market demand needs. How sensitive are power investments to economic growth? Economic growth creates greater demand for electricity, while also providing some of the resources needed to pay for it. Lower growth reduces demand. We explored a low-growth scenario in which economic growth per capita was assumed to be 50 percent lower than assumed in our base case. In SAPP, the largest reduction in power demand would occur in South Africa, where investments in new coal-fired plants would be put on hold. In EAPP, lower demand growth would first reduce investments in gas-fired power plants in Egypt. Hydropower investments would be only slightly reduced under the low-growth scenario, implying that even with slower economic growth the market remains large enough to justify the expansion of almost all the hydropower capacity considered in the base case. In WAPP, less of the old gas-fired capacity in Nigeria would be refurbished and less of the hydropower in Côte d’Ivoire would be exported to Ghana. In CAPP, hydropower investments in the Republic of Congo would be reduced by almost 30 percent, but imports from Cameroon would increase to partly replace them. Overall, the reductions in annual power spending needs resulting from lower growth are 10 percent in EAPP and CAPP, about 15 percent in WAPP, and almost 25 percent in SAPP. For all of Sub-Saharan Africa the reduction is 20 percent. The cuts may seem modest compared to a 50 percent cut in economic growth per capita, but keep in mind that economic growth for the country is much higher than economic growth per capita since the population is continually increasing. In fact, the low growth scenario lowers spending needs more than it lowers GDP, and spending needs as a share of GDP would decrease from 4.2 to 4.0 percent. Why trade power? African countries have different endowments of natural resources: some have abundant hydropower resources, while others have domestic resources of coal or natural gas. Some have no domestic energy resources but depend on imported diesel fuel to generate power. Trade with neighboring countries enables power production from the cheapest sources in the region. By stimulating the development of hydropower, expanded regional trade in power would lower the generation costs, reduce carbon emissions from power plants, and insulate countries from hikes in the price of fossil fuels. Expanded trade would also encourage investment. For example, the optimal size of a new hydropower plant is often so large that domestic demand cannot absorb the large capacity expansion, so the new plant will not be built. Further development of power trade will incur significant infrastructure costs to develop cross-border transmission capacity. It is estimated that some 12 GW of needed interconnectors are lacking in SAPP and 14 GW in EAPP. The interconnector needs are less in the other areas: some 5.5 GW in WAPP and only 800 MW in CAPP. Building those lines would cost around $380 million per year in SAPP, $130 million in EAPP, $120 million in WAPP, and $40 million in CAPP. xii COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The benefits of building the interconnectors would be substantial, reducing annualized power system costs by between 5 and 11 percent in the trading regions. The savings would be the largest in CAPP at 11.5 percent, compared to 7–8 percent in SAPP and EAPP and 5.1 percent in WAPP. Keep in mind, however, that operation of existing equipment contributes to the annual cost in 2015. The cost of running this equipment adds to the annualized system cost, but it cannot reasonably be expected to be influenced by future trade. Power trade would save Sub-Saharan Africa an estimated $2.7 billion annually, or 5.3 percent of the annual cost of meeting power needs (or 7.2 percent of the cost when the operation of existing equipment is deducted; see table B). The savings come largely from substituting hydro for thermal plants, which substantially reduces the operating cost of the power system, although it requires more investment in the short run. For example, power trade would provide operating cost savings of 1 percent of the area’s GDP in EAPP and almost 0.5 percent of the area’s GDP in CAPP. In EAPP and WAPP the hydropower plants substitute for gas-fired power plants, while in CAPP the new hydropower replaces thermal power fueled by heavy fuel oil (HFO), which is more expensive and more polluting than gas-fired plants. The savings on operating costs can be considered a return on the additional capital investments made under the trade scenario. In SAPP, the additional investment cost under trade is recouped in less than a year, yielding an annual return of 167 percent. The return is lower in the other three regions, but still generous at around 20–34 percent; the additional investment cost of the trade scenario is recouped over three to four years. For Sub-Saharan Africa as a whole the return on trade investment is 27 percent. Moreover, the gains from trade increase as fuel prices rise, since trade reduces the use of thermal power plants and thus saves fuel. As fuel prices rise, hydropower projects become more profitable. At an oil price of $75/barrel (instead of $46/barrel in the base case), the gains from trade in EAPP amount to almost $3 billion annually. What patterns of trade would emerge? The expansion of power trade, as in our trade-expansion scenario, would allow countries with significant hydropower potential to develop their capacity and meet demand elsewhere. In SAPP, the hydropower share would rise from 25 to 34 percent of the generation capacity portfolio. The Democratic Republic of Congo becomes the major exporter of hydropower, exporting three times as much as its domestic consumptions, while Mozambique continues to be a significant exporter. Hydropower from the Democratic Republic of Congo flows southward along three parallel routes through Angola, Zambia, and Mozambique (figure B). Countries such as Angola, Botswana, Lesotho, Malawi, and Namibia would become reliant on imports to meet more than 50 percent of power demand. In addition, South Africa would import large volumes of power, which would still account for only 10 percent of domestic demand. xiii COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure B Maximum potential for cross-border power trading in Sub-Saharan Africa, 2015 (TWh) A similar shift from thermal to hydropower would occur in the EAPP region, pushing hydropower from 28 to 48 percent of the generation capacity portfolio and displacing gas-fired power capacity in Egypt and Kenya. Ethiopia and Sudan would become the major power exporters, trading more than what they produced for domestic consumption and sending their power northward into Egypt (figure B). While Egypt and Kenya would import significant volumes of power, Burundi would be the only country to become largely dependent on traded electricity. xiv COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA In WAPP, the increase in trade would not significantly increase reliance on hydropower, but hydropower in Guinea would replace hydropower projects dispersed throughout the other countries. Also, trade would reduce the development of gas-fired power plants in countries such as Ghana, Benin, Togo and Mauritania. Guinea would emerge as the major exporter of hydropower, exporting more than 5 times domestic consumption. In CAPP, the share of power production from hydropower would increase from 83 percent to 97 percent. Cameroon emerges as the major power supplier in CAPP, exporting about half of its production. Hydropower capacity in Cameroon replaces the HFO-fired thermal capacity in the other countries, in addition to some hydropower in the Republic of Congo. The other countries in the region, except Central African Republic, import a considerable share of their consumption: Chad and Equatorial Guinea import all of their power from Cameroon, while the Republic of Congo imports about one third and Gabon almost half of its consumption. The Central African region borders the Democratic Republic of Congo and therefore could be expected to benefit from hydropower development there. In this study, the Democratic Republic of Congo is part of the SAPP region and only the 2005 level of imports from the Democratic Republic of Congo to the Republic of Congo was included in the base case. Even if higher imports from the Democratic Republic of Congo to the CAPP region are possible, Cameroon still remains the major supplier in CAPP (depending on the level of imports from the Democratic Republic of Congo). Instead, investments and production in the Republic of Congo are replaced by the increased imports. Who gains most from power trade? There are substantial differences in the long-run marginal cost (LRMC) of power across power pool areas, and those differences are differentially affected by trade (table C). The SAPP and CAPP regions have considerably lower average LRMC ($0.07 per kWh) than the EAPP and WAPP regions: the average LRMC of power in the EAPP region is around $0.12 per kWh and $0.18 per kWh in WAPP. The LRMC is also quite high on islands, between $0.14 and 0.19 per kWh. Of course, these numbers are estimates, with a considerable degree of uncertainty at the country level. The range within each power pool is also wide, though trade tends to narrow that range. Two types of countries benefit from trade. Countries with very high domestic power costs can obtain significantly cheaper electricity by importing. Perhaps the most striking examples are in WAPP, where Guinea-Bissau, Liberia and Niger each can save up to $0.06–$0.07 per kWh by importing electricity. Countries in other regions also benefit from considerable savings: Angola in SAPP, Burundi in EAPP, and Chad in CAPP can all save up to $0.04–$0.05 per kWh by importing electricity. But even countries with smaller unit cost differentials, such as Burundi, Malawi, Ghana, Sierra Leone, and Togo, can generate important savings by moving from self-reliance to heavy imports. On the other hand, countries with very low domestic power costs can also generate substantial revenues by exporting power. The most salient examples are the Democratic Republic of Congo for SAPP, Ethiopia for EAPP, Guinea for WAPP, and Cameroon for CAPP. Power export revenues could amount to 6 percent of GDP for Ethiopia and 9 percent of GDP for the Democratic Republic of Congo. xv COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table C Long-run marginal costs of power in Sub-Saharan Africa Trade Trade- Trade Trade- U.S. dollars per kWh expansion stagnation U.S. dollars per kWh expansion stagnation SAPP average 0.06 0.07 WAPP average 0.18 0.19 Angola 0.06 0.11 Benin 0.19 0.19 Botswana 0.06 0.06 Burkina Faso 0.25 0.26 Congo, Dem. Rep. 0.04 0.04 Cote d’Ivoire 0.15 0.15 Lesotho 0.06 0.07 Gambia 0.08 0.07 Malawi 0.05 0.05 Ghana 0.10 0.10 Mozambique 0.04 0.06 Guinea 0.07 0.06 Namibia 0.11 0.12 Guinea-Bissau 0.09 0.16 South Africa 0.06 0.07 Liberia 0.08 0.14 Zambia 0.08 0.08 Mali 0.25 0.28 Zimbabwe 0.08 0.09 Mauritania 0.14 0.15 EAPP average 0.12 0.12 Niger 0.25 0.30 Burundi 0.11 0.15 Nigeria 0.13 0.13 Djibouti 0.07 0.07 Senegal 0.43 0.47 Egypt 0.09 0.09 Sierra Leone 0.09 0.10 Ethiopia 0.19 0.16 Togo 0.10 0.11 Kenya 0.12 0.13 CAPP average 0.07 0.09 Rwanda 0.12 0.12 Cameroon 0.07 0.06 Sudan 0.13 0.13 Central African Republic 0.11 0.11 Tanzania 0.10 0.08 Chad 0.07 0.11 Uganda 0.12 0.11 Congo, Rep. 0.06 0.08 Island states Equatorial Guinea 0.08 0.10 Cape Verde 0.19 0.19 Gabon 0.07 0.07 Madagascar 0.14 0.14 Mauritius 0.18 0.18 Note: In some cases power exporting countries report higher LRMC under trade expansion. Even if the cost of meeting domestic power consumption may be higher with trade than without, the higher revenues earned from exports would more than compensate for that increment. How will less hydropower development influence the trade flows? In the trade-expansion scenario, cheap hydropower from Guinea supplies much of the power in the WAPP region (except Nigeria). But it might be unrealistic to develop such a huge amount of hydropower in one country in a short time span. In a reduced hydropower development sceanrio, we assume that only three projects, 375 MW in total, can be completed in Guinea before 2015 (instead of the 4,300 MW in the base case). If hydropower development in Guinea is restrained, new trade patterns emerge in the WAPP region. Côte d’Ivoire emerges as the major power exporter, while Ghana increases domestic production considerably to reduce net imports. Mauritania and Sierra Leone also become exporters. xvi COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Hydropower investments in Côte d’Ivoire only increase by just below 200 MW, but production from existing gas-fired power plants also increases, so total power production increases by 3 TWh. Production in Ghana almost doubles to 16 TWh. Like in Côte d’Ivoire, the increase comes partly from new gas-fired power plants and partly from existing plants. Total annualized costs increase by only 3 percent (just above $300 million). There is, however, a huge trade-off between lower capital costs and higher variable costs: while capital costs are $500 million less (mainly due to lower generation investments), variable operation costs of production are $850 million (30 percent) more. This clearly illustrates the trade-off between hydro and thermal capacity. With less hydropower capacity, more of the existing thermal capacity is used with lower efficiency and higher costs than new plants. What are the environmental impacts of trading power? Trade in power also offers potential environmental benefits. In the SAPP region, our model predicts that trade would increase the share of hydropower generation capacity from 25 to 34 percent, reducing CO2 emissions by about 40 million tons. In the EAPP region, CO2 emissions would drop by 20 million tons, even as power production rose by 2.4 TWh. In the WAPP and CAPP regions, the CO2 savings are smaller, 5.2 and 3.6 million tons respectively, since the regions’ total production is smaller. The combined savings are 70 million tons of CO2 annually. By comparison, the Energy Information Administration (EIA) estimates current emissions from power and heat production in Africa to be 360 million tons. The savings from trade is therefore 20 percent of this volume. Our estimates do not, however, include greenhouse gas emissions from hydropower in the form of methane from dams. How would CDM affect generation technology choices? Created pursuant to the Kyoto Protocol, the Clean Development Mechanism (CDM) allows industrialized countries that have made a commitment under the protocol to reduce greenhouse gases to invest in projects that reduce emissions in developing countries instead. The investment covers the difference in cost between a polluting technology and a cleaner but more expensive alternative. The CDM difference in cost is divided by emissions saved to work out the cost of certified emission reduction credits (CERs) associated with a given project. Focusing on SAPP, we analyzed the potential for CDM in the power sector of Sub-Saharan Africa, operating under the trade-expansion scenario. CDM has not been widely used in the power sector in Sub-Saharan Africa. An illustrative simulation shows that at a CER price of $15/ton CO2, investments in the Democratic Republic of Congo, Malawi, Zambia, and Namibia would lead to the development of an additional 8,000 MW (producing 42 TWh) of hydropower. A CER price of $15 has the potential to reduce CO2 emissions by 36 million tons—equivalent to 10 percent of Africa’s emissions from power and heat production. That amount is significant but still less than the carbon reduction brought about by trade, which reduces CO2 emissions by 40 million tons in SAPP. Trade and CDM are not mutually exclusive, of course. Starting from a trade-stagnation position, moving to a trade-plus-CDM position could reduce CO2 emissions by 76 million tons. xvii COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA One facet of the CDM model limits its contribution. System costs for Africa after CDM finance are still higher than before CDM finance. The reason seems to be that transmission and distribution costs increase after CDM (because hydropower plants are located far from consumption centers), but those costs are not addressed by the mechanism. How might climate change affect power investment patterns? By affecting weather patterns and making hydropower less reliable, climate change could increase the costs of generating and delivering power in Africa. Focusing on the EAPP region, we performed an illustrative analysis to examine some of the key issues posed by climate change. Because exact numbers are lacking, we performed simulations in which climate change was assumed to reduce so-called firm hydropower production (in GWh per MW of installed capacity) by up to 25 percent. The reduction was assumed to apply both to existing capacity and to new capacity. Lower firm power would increase the unit cost of hydropower, causing gradual substitution away from hydropower and increasing the total annualized cost of the power sector. It is perhaps some comfort that a reduction of 25 percent in firm hydropower availability would increase the annual costs of satisfying the region’s power needs by only 9 percent. But it is decidedly not comforting that climate change would increase East Africa’s dependency on thermal power—production in gas-fired power plants would increase 40 percent in EAPP. In other words, the solution to the power supply problem brought about by climate change implies an acceleration of the climate problem. xviii 1 Introduction The aim of this study is to analyze the investment requirements and costs associated with electrification in Sub-Saharan Africa, grouped into the Southern Africa Power Pool (SAPP), the Nile Basin–East Africa Power Pool (EAPP), the Western Africa Power Pool (WAPP), and the Central Africa Power Pool (CAPP). The island nations of Cape Verde, Madagascar, and Mauritius are also included. We have analyzed the power sector investments necessary for three different scenarios (defined by the share of the population with access to electricity): (a) keeping electricity access rates constant at the 2005 level; (b) increasing them to an average of 35 percent in SAPP and EAPP, 44 percent in CAPP, and 53 percent in WAPP; and (c) increasing them to the national target rates of individual countries. In addition, we assess the potential gains from electricity trade. We also provide sensitivity analyses of our findings’ robustness. In the analysis, we use a least-cost expansion model developed for the project. The model is populated with data from 43 Sub-Saharan African countries, grouped into regions (as shown in table 1.1). The model will be available online for use by researchers. Table 1.1 Countries covered in the study Southern Africa Power Nile Basin–East Africa Western Africa Power Central Africa Power Island states Pool Power Pool Pool Pool Angola Burundi Benin Cameroon Cape Verde Botswana Djibouti Burkina Faso Central African Republic Madagascar Democratic Republic of Egypt Côte d'Ivoire Chad Mauritius Congo Lesotho Ethiopia Gambia Republic of Congo Mozambique Kenya Ghana Equatorial Guinea Malawi Rwanda Guinea Gabon Namibia Sudan Guinea-Bissau South Africa Tanzania Liberia Zambia Uganda Mali Zimbabwe Mauritania Niger Nigeria Senegal Sierra Leone Togo Each country is represented by a demand side and supply side. The demand side consists of market demand, suppressed demand and social demand. Market demand for electricity is defined as the demand resulting from economic growth and structural change; that is, the growth of existing consumers’ demand due to increase in income. Social demand is related to social targets for new electricity access. Suppressed 1 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA demand is that portion of electricity demand not currently met because of blackouts and brownouts. The next chapter presents the data and methodology used to estimate demand. These projections are then used in the least-cost expansion model. The supply system consists of power generation plants and transmission (between and within countries) and distribution grids (within countries). The available technologies for power generation are hydropower (large scale, including pumped storage, and small scale), thermal (fueled by coal, natural gas, diesel, heavy fuel oil [HFO], and nuclear), solar photovoltaic energy (solar PV) and geothermal power. In addition to grid-based electricity supply, off-grid supply (provided by minihydro, diesel, and solar PV) is an option in rural areas. Each technology is described in terms of capacity and cost. We take the countries’ existing assets (as of 2005) as a starting point. Capacity expansion is possible through investment. In addition, refurbishment can extend the life span of existing capacity. Costs consist of the capital cost and variable operating cost (that is, the cost of fuel and operation and maintenance). Investments are also necessary to extend transmission and distribution grids. These grids also require refurbishment to keep them in working condition. Chapter 2 presents the demand-side data. Chapter 3 presents the supply-side data for generation, as well as for transmission and distribution. Chapter 4 gives an overview of the least-cost expansion model. Chapters 5 and 6 present the results of the analysis. Chapter 5 contains results of the major scenarios for electricity access and trade. Chapter 6 investigates sensitivities with respect to demand and economic growth, fuel prices, investment unit costs, clean development mechanism (CDM) finance, the impacts of climate change, and other barriers to hydropower development. Annex 1 (separately bound) contains detailed country-by-country outputs. Four appendixes accompany this report. Appendixes 1–3 present details of input data for the supply side. Appendix 4 compares results of this study with those of a recent study on the Southern African Power Pool, Nexant (2007). 2 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA 2 Estimating demand This chapter presents the methodology used to estimate demand for electricity in 2015. Demand consists of (a) market demand, associated with different levels of economic growth, structural change and population growth; (b) suppressed demand, created by frequent blackouts and the ubiquitous practice of power rationing; and (c) social demand, as expressed in political targets for increasing access to electricity. Together they form our country-specific demand projection: Demand = market demand + suppressed demand + social demand These projections are used in the least-cost expansion model. Market demand This section estimates market demand in Sub-Saharan African countries in 2015. We define market demand growth as the “organic” growth resulting from income growth and structural changes in the economy. First, we present different specifications of the econometric model for electricity demand, discuss the robustness of the model, and explain our choice of the model for projections of future demand. We then use the model to project future electricity demand. The estimation model is used for each country analyzed. Econometric model of demand We model the market demand, measured as annual per capita electricity consumption, as a log-linear function of real gross domestic product (GDP) per capita (a proxy for income), and of the agricultural, manufacturing, and construction shares of the economy. In addition, we follow Fay and Yepes (2003) and include country-specific fixed effects and time trends to proxy for prices and technological innovation. We implement the econometric model Table 2.1 Data sources for econometric model on a panel covering 11 Sub-Saharan Variable Source countries. For the remaining countries, data Electricity consumption International Energy Agency Database are unreliable or missing. While there is Population Penn World Tables usually a benefit to basing quantitative Real GDP per capita Penn World Tables analysis on large data sets, adding more Structural shares of United Nations Statistics Division, National data can lead to biased estimates from the economy Accounts econometric model if the additional data Population growth United Nations World Population Prospects are plagued by measurement error.1 The data sources are presented in table 2.1. The estimation is based on 30 time periods that together with data for the 11 countries yield 330 observations. 1 In a multiple linear regression framework it is well known that measurement error in the explanatory variables can cause them to be biased toward zero. 3 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA We allow for first-order autocorrelation and panel correlation in the residuals to control for unobserved factors that may affect electricity demand and may be correlated over time or across countries. All variables are in natural logarithms; in this way, the estimated coefficients can be interpreted as elasticities. We estimate the model with feasible generalized least squares (FGLS), using the econometric software Stata 9. Our prior assumption is that the coefficient on GDP per capita (that is, the income elasticity) is positive and not too far from unity. One could also posit that the sectorial shares coefficients are negative for agriculture and positive for manufacturing (as in Bogetic and Fedderke, 2005). The estimation results are reported in table 2.2. Table 2.2 Econometric model for market demand: estimation results Regressor Coefficient Standard error Z P>z Income 1.071 0.056 19.08 0.000 Agriculture –0.255 0.047 –5.42 0.000 Manufacturing 0.127 0.053 2.38 0.017 Construction –0.074 0.032 –2.34 0.019 Constant –9.415 0.488 –19.28 0.000 Source: Authors’ original research. Note: Variables in logarithms. Manufacturing includes mining sector. Residuals assumed autoregressive and panel correlated. The results from the model agree with our prior assumptions and earlier results. Fay (2001) estimates an income elasticity of 1.1, while Bogetic and Fedderke’s (2005) estimate is somewhat higher. Turning first to the coefficient on real GDP per capita—our estimate of the income elasticity—we see that it is close to unity, and similar to the elasticity reported in Fay and Yepes (2003), while somewhat lower than the elasticity reported in Bogetic and Fedderke (2005). An F-test that the quotient of the coefficient divided by elasticity is equal to unity cannot be rejected. The coefficients on the sectorial shares for agriculture and manufacturing are consistent with the results in Bogetic and Fedderke (2005). The results are also robust to alternative specifications and sample sizes, with two major exceptions: • The coefficient on construction is sensitive to changes in specification, the use of alternative estimators, and changes in sample size. It is close to zero in many specifications and frequently not robust. We include it in the above specification, but exclude it from the projections of future demand. • The coefficient on income, representing income elasticity, drops from about 1.1 to 0.5 if one includes urbanization in the regression. The coefficient on urbanization is highly statistically significant and slightly below unity (0.9). The reason for this effect is the omitted variables bias. Since urbanization is positively correlated with both real income per capita and electricity consumption (because people in urban areas use more electricity than in rural areas), omitting it from the regression raises the 4 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA coefficient on income.2 We chose to proceed with the model with higher income elasticity, since projections of future urbanization are highly uncertain. In order to increase the robustness of the demand projections, we chose to omit construction and urbanization from the forecasting model. The resulting forecasting model is based on the results in table 2.2. %Electricity consumption per capita = –9.415 + 1.07%Income + 0.13%Manufacturing – 0.25%Agriculture Projections of explanatory variables In order to complete a forecast of electricity demand, we need to project the explanatory variables— more specifically, the percentage change in the explanatory variables—from 2006 to 2015. Projected values of real GDP per capita are taken from the international macroeconomic data set of the United States Department of Agriculture (USDA) Economic Research Service and World Development Indicators database. We also need projections for the agricultural and manufacturing shares of the economy. Such data are not available to our knowledge, but can be estimated from time-series data. One way is to simply extrapolate the time-series data on agriculture and manufacturing for each country. Alternatively, one can base a projection of future agricultural and manufacturing shares of the economy on the strong correlation between economic growth and structural change. The reallocation of labor from agriculture and into service and manufacturing that follows economic growth has been documented for both developed and developing economies (see, for example, Kongsamut, Rebelo, and Xie, 1997). Given the correlation present in our sample of SAPP countries, we estimate future sectorial shares based on the coefficients from two linear fixed-effects regressions of agricultural and manufacturing on (a) real GDP per capita and (b) projected future real GDP per capita.3 Projections of market demand Using the econometric panel data models and projections for the future development of the explanatory variables, we estimate the increase in market demand for electricity in the four regions of Sub-Saharan Africa and in the island states. 2 The decrease in income elasticity from the inclusion of urbanization can be understood in the light of omitted variable bias. Formally, suppose the correctly specified equation is . Omitting x2 from the regression leads to the following expectation of b1, the least squares estimate of : E[b1]= +( ’ )-1( ’ ) . Hence the bias in the coefficient on from omitting depends on the correlation between and and the correlation of and y (which determines the sign of ). 3 These equations are: Manufacturing = 0.88+0.28% GDP/capita; Agriculture=8.8-0.82% GDP/capita. In other words, the manufacturing share increases 2.8 percentage points when GDP per capita increases 1 percent. (For EAPP the figure is 3.2 percentage points.) The agricultural share decreases 8.2 percentage points from GDP per capita growth, but the base is positive and the net effect depends on the economic growth assumed. In all normal growth scenarios, when growth is sufficiently high, the agricultural share declines. The equations forecasting sectorial shares should be revisited in future work. 5 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The assumptions for population growth and GDP growth and the resulting projections for demand growth are shown in table 2.3. Electricity demand per capita, as estimated by the model, has been multiplied by 2005 and 2015 populations. Hence the figures for electricity demand growth show an increase in total demand, not per capita demand. In addition to the normal growth scenario, we also report a scenario with lower economic growth (GDP/capita growth reduced by 50 percent). The lower growth in electricity demand that results from slower economic growth is used in the low-growth scenario presented in chapter 6. Table 2.3 Increase in population, income, and market demand for electricity in study countries, 2006–15 (%) Base growth scenario Low growth scenario Annual population Total (annual) Total (annual) Country growth Electricity Electricity GDP/capita GDP/capita demand demand Southern Africa Power Pool Angola 2.8 84 (6.3) 130 (9.7) 42 (3.6) 75 (6.4) Botswana -0.4 56 (4.6) 58 (5.2) 28 (2.5) 21 (2.1) Congo, Dem. Rep. 2.5 16 (1.5) 45 (4.2) 8 (0.8) 35 (3.4) Lesotho -0.3 53 (4.3) 55 (5.0) 26 (2.4) 20 (2.1) Malawi 2.2 19 (1.7) 38 (3.6) 9 (0.9) 25 (2.5) Mozambique 1.7 21 (1.9) 34 (3.3) 10 (1.0) 21 (2.1) Namibia 1.0 48 (4.0) 62 (5.5) 24 (2.2) 30 (3.0) South Africa 0.1 43 (3.7) 47 (4.3) 22 (2.0) 18 (1.8) Zambia 1.7 24 (2.2) 38 (3.6) 12 (1.1) 22 (2.2) Zimbabwe 0.6 30 (2.7) 34 (3.3) 15 (1.4) 14 (1.5) Nile Basin–East Africa Power Pool Burundi 3.2 14 (1.3) 45 (4.2) 7 (0.7) 35 (3.4) Djibouti 1.9 15 (1.4) 29 (2.9) 8 (0.7) 19 (2.0) Egypt 1.8 26 (2.3) 41 (3.9) 13 (1.2) 24 (2.5) Ethiopia 2.3 29 (2.6) 52 (4.8) 14 (1.4) 33 (3.2) Kenya 2.6 16 (1.5) 39 (3.7) 8 (0.8) 28 (2.8) Rwanda 2.2 31 (2.7) 54 (4.9) 6 (0.6) 22 (2.2) Sudan 2.0 32 (2.8) 53 (4.8) 16 (1.5) 31 (3.1) Tanzania 1.8 26 (2.3) 41 (3.9) 13 (1.2) 24 (2.4) Uganda 3.8 13 (1.2) 51 (4.7) 7 (0.6) 43 (4.0) Western Africa Power Pool Benin 2.5 19 (1.8) 40 (4.0) 10 (0.9) 29 (2.9) Burkina Faso 3.0 4 (0.4) 28 (2.8) 2 (0.2) 25 (2.6) Cote d’Ivoire 1.9 11 (1.0) 23 (2.3) 5 (0.5) 16 (1.7) Gambia 2.7 13 (1.3) 36 (3.5) 7 (0.6) 28 (2.7) Ghana 1.9 37 (3.2) 58 (5.2) 19 (1.6) 33 (3.2) Guinea 2.6 17 (1.6) 41 (3.9) 9 (0.8) 29 (2.9) Guinea-Bissau 2.0 2 (0.2) 14 (1.4) 1 (0.1) 12 (1.3) Liberia 3.1 17 (1.6) 46 (4.3) 9 (0.8) 35 (3.4) Mali 2.8 21 (1.9) 49 (4.5) 11 (1.0) 35 (3.4) 6 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Base growth scenario Low growth scenario Annual population Total (annual) Total (annual) Country growth Electricity Electricity GDP/capita GDP/capita demand demand Mauritania 2.9 62 (4.9) 102 (8.1) 31(2.5) 61 (5.5) Niger 2.9 3 (0.3) 25 (2.5) 1 (0.1) 23 (2.3) Nigeria 2.4 38 (3.2) 65 (5.7) 19 (1.6) 40 (3.8) Senegal 2.5 22 (2.0) 46 (4.3) 11 (1.0) 31 (3.1) Sierra Leone 2.3 5 (0.5) 21 (2.1) 3 (0.3) 17 (1.8) Togo 2.7 2 (0.2) 22 (2.2) 1(0.1) 21 (2.1) Central Africa Power Pool Cameroon 2.2 22 (2.0) 41 (3.9) 11 (1.0) 27 (2.7) Central African Rep. 1.5 16 (1.5) 26 (2.6) 8 (0.8) 15 (1.5) Chad 2.9 8 (0.8) 33 (3.2) 4 (0.4) 27 (2.7) Congo, Rep. 2.8 24 (2.2) 52 (4.8) 12 (1.1) 36 (3.5) Equatorial Guinea 2.0 36 (3.1) 57 (5.1) 18 (1.6) 34 (3.3) Gabon 2.0 13 (1.2) 27 (2.7) 6 (0.6) 19 (1.9) African Island states Cape Verde 0.5 58 (4.7) 70 (6.0) 29 (2.3) 32 (3.1) Madagascar 2.5 12 (1.2) 33 (3.2) 6 (0.6) 25 (2.5) Mauritius 0.8 34 (2.9) 40 (3.8) 17 (1.5) 18 (1.9) Source: Authors’ original research. Note: Market demand for electricity is one of three categories of demand for electricity, the others being social demand/access, and suppressed demand. Suppressed demand In most low-income countries, some people who want to consume more energy at the going price are unable to. The difference between so-called notional demand and availability at the going price is called suppressed demand. Suppressed demand arises for two primary reasons. First, there are people on a waiting list to get connected to the grid who are not captured in our market demand estimate. Second, due to old and run-down infrastructure, blackouts and brownouts are relatively common; therefore, consumption is reduced, while notional demand remains unchanged. All of this implies that suppressed demand will immediately absorb a certain amount of new production before income growth or structural economic changes does so. Note that suppressed demand as defined here does not include the possible positive effect of increased supply on economic growth. In our model that effect is categorized as market demand. In a country where economic activity is limited by electricity supply, the market-demand effect can be strong. In our simulations, we account for the additional consumption of suppressed demand as follows: Waiting list. Since this is a direct result of slow expansion of new connections, we assume that this source of suppressed demand is picked up by social demand in each scenario. In fact, the pace of expansion is what determines our three expansion scenarios. Blackouts. In order to estimate the amount of suppressed demand from blackouts, we rely on the results of the World Bank’s enterprise surveys, which include estimates for blackout duration and 7 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA frequency. (For most countries, one to two data points are provided for the period 2003–6. For countries where no data is available, the regional average of the sample is applied.) This provides an estimate of the number of hours that an average household/business in a given country spends in blackout. The estimate for number of hours is then divided by 8,760 (the number of hours in a year) to convert the estimate into a percentage of the year. Then, electricity demand in energy terms (gigawatt-hours, GWh) is adjusted accordingly (table 2.4). Table 2.4 Summary of blackout data and suppressed demand in study countries Average duration Outages Down time Suppressed demand Country (hours) (hours per year) (% of a year) in 2005 (GWh) SAPP Angola 19.31 1780.8 20.3 435 Botswana 1.86 38.9 0.4 11 Congo, Dem. Rep. 3.63 659.2 7.5 351 Lesotho 7.65 177.9 2 8 Malawi 4.27 328.1 3.7 49 Mozambique 6.08 350.4 4 450 Namibia 2.32 46.1 0.5 13 South Africa 4.15 24.5 0.3 602 Zambia 5.48 219.9 2.5 157 Zimbabwe 6.08 350.4 4 512 Average for available sample 6.1 350.4 4 EAPP Burundi 10.34 1461.5 16.7 25 Djibouti 5.88 456.4 5.2 12 Egypt 2.48 43.3 0.5 417 Ethiopia 5.88 456.4 5.2 109 Kenya 8.2 702.6 8 366 Rwanda 4.47 346.9 4 5 Sudan 5.88 456.4 5.2 168 Tanzania 6.46 435.9 5 208 Uganda 6.55 463.8 5.3 84 Average for available sample 5.88 456.4 5.2 WAPP Benin 2.72 505 6 34 Burkina Faso 1.61 196 2 11 Cote d’Ivoire 5.94 1101 13 365 Gambia 6.86 1961 22 29 Ghana 12.59 1465 17 979 Guinea 6.78 2759 31 224 Guinea-Bissau 17.94 1978 23 14 Liberia 5.94 1101 41* 123 Mali 2.44 453 5 21 Mauritania 2.89 129 1 3 Niger 0.5 124 1 6 8 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Average duration Outages Down time Suppressed demand Country (hours) (hours per year) (% of a year) in 2005 (GWh) Nigeria 5.94 1101 64* 10803 Senegal 5.67 1052 17* 250 Sierra Leone 5.94 1101 82* 189 Togo 5.94 1101 13 73 Average for available sample 5.94 1176 22 CAPP Cameroon 4.03 613 7 241 Central African Rep. 5.2 950 10.8 11 Chad 5.2 950 10.8 10 Congo, Rep. 4.33 924 10.6 616 Equatorial Guinea 5.2 950 10.8 3 Gabon 5.2 950 10.8 134 Average for available sample 5.2 889 10.2 Island states Cape Verde 5.3 797 9 4 Madagascar 2.67 221.1 2.5 21 Mauritius 7.23 1321 15.1 35 Source: World Bank Enterprise Surveys. * WAPP Emergency Power Supply Report (2007). Note: Due to a lack of data, regional sample averages are applied to the following countries: Mozambique, Zimbabwe (SAPP); Ethiopia, Sudan and Djibouti (EAPP); Benin, Cote d’Ivoire, Liberia, Mali, Nigeria, Senegal, Sierra Leone, and Togo (WAPP). For CAPP, data is available for Cameroon and Republic of Congo only; for the other countries, regional average is applied. Social demand Social demand for electricity is composed of the expected demand of all new connections in the household sector in 2015. We develop three scenarios for electricity access: • The 2005 access scenario assumes that access rates are kept constant at their 2005 level. Because of population growth, even the current access scenario implies that a number of new connections are added and, hence, a higher demand level (measured in kilowatt-hours, kWh). • The regional target access scenario aims to increase access by roughly 1 to 2 percentage points per year in each region. Based on stakeholder discussions, the scenario appears to be a moderately ambitious, yet realistic, access target. • The national targets scenario reflects ambitions set by national governments for urban and rural electricity access. The relevance of this scenario lies in the fact that it is based on national aspirations. These scenarios provide for a reasonable lower-bound estimate (2005 access rates) and particularly ambitious estimate (national targets), together with a realistic yet ambitious target (regional targets) used by international organizations such as the World Bank in planning and strategy documents. We then arrive at our estimate for social demand in two steps: (a) we estimate the weighted average consumption per new connection, and (b) we estimate the number of new connections in the three scenarios. Let us describe the steps in turn. 9 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Average consumption per connection To estimate the total additional generation capacity required to satisfy the demand of new connections, we must first estimate the average demand per connection. We assume that all commercial demand is captured in market demand, and that social demand is thus made up of household connections.4 In arriving at an estimate of average consumption per connection, we use an extensive data set collected in Tanzania,5 and cross-check our estimates with those of the global review of rural electrification programs carried out by Zomers (2001). The resulting weighted average monthly consumption for rural and urban areas is 50.4 kWh and 114.7 kWh, respectively. The corresponding average daily maximum demand is 0.29 kW for rural households and 0.64 kW for urban households. These estimates correspond well with other international estimates, including those of Zomers’s global review of rural electrification projects (2001), which arrived at an average monthly consumption of 81.75 kWh and an average maximum demand of 0.4 kW. The fact that Zomers’s estimate is slightly higher corresponds with the fact that it is global, with only a few observations from Sub-Saharan Africa. Estimating the number of new connections We build the estimation of the number of connections in rural and urban areas around three variables: (a) the 2005 electricity access rates, (b) future population growth, and (c) targets for future electricity access rates. Population statistics and projections for rural and urban areas in 2005 are taken from the United Nations’s World Population Prospects and reported in table 2.5 (assuming that there are 5 persons per household). 4 Some new connections, especially in urban areas, may come ”organically” from market demand growth. Thus, by adding our estimates for social and market demand, there is a risk that the consumption of some new connections is counted double. But it is expected that this double counting is rather insignificant for two reasons. First, the historical data on which we base our market demand projections demonstrate very little access expansion— connection rates remain low in most African countries. This implies that our market demand estimate does not pick up significant new connections. Second, we do not count any commercial connections in our measure of social demand, although other social demand forecasts include those. Thus, it is expected that the neglect of commercial connections in social demand counteracts the potential double counting of residential connections, and the overall bias in the estimate, whether over or under, is limited. 5 These consumption figures are calculated as follows. Our firm’s survey of a large number of potential grid extension programs found that about 90 percent of all connections are households. We combine this estimate with DECON/SWECO (2005) estimates for income-class prevalence rates and expected-income class take-up rates. Using these distributions we calculate the distribution of connections across household income levels and commercial connections (for example, for a given 100 connections). We then combine this with DECON/SWECO’s (2005) specific consumption data to arrive at a weighted average consumption per connection for urban and rural, which was used in the model. 10 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 2.5 2005 and projected population in study countries No. of households 2005 No. of households 2015 Total Urban Rural Total Urban Rural SAPP Angola 3,188,200 1,186,010 2,002,190 4,189,400 1,881,041 2,308,359 Botswana 353,000 185,325 167,675 338,000 194,350 143,650 Congo, Dem. Rep. 11,509,800 3,769,000 7,740,800 14,770,600 6,201,000 8,569,600 Lesotho 359,000 65,338 293,662 348,800 73,248 275,552 Malawi 2,568,800 441,834 2,126,966 3,197,600 709,867 2,487,733 Mozambique 3,958,400 1,504,192 2,454,208 4,702,600 2,280,761 2,421,839 Namibia 406,200 136,077 270,123 449,600 178,941 270,659 South Africa 9,486,400 5,492,626 3,993,774 9,580,400 6,006,911 3,573,489 Zambia 2,333,600 858,765 1,474,835 2,768,200 1,129,426 1,638,774 Zimbabwe 2,602,000 934,118 1,667,882 2,760,800 1,142,971 1,617,829 Total SAPP 36,765,400 14,573,284 22,192,116 43,106,000 19,798,515 23,307,485 EAPP Burundi 1,571,800 157,180 1,414,620 2,244,600 303,021 1,941,579 Djibouti 160,800 138,449 22,351 190,200 170,419 19,781 Egypt 14,986,400 6,438,200 8,548,200 17,635,000 7,916,800 9,718,200 Ethiopia 15,486,000 2,502,200 12,983,800 19,431,000 3,844,400 15,586,600 Kenya 6,851,200 2,852,600 3,998,600 8,839,000 4,576,800 4,262,200 Rwanda 1,807,600 393,600 1,414,000 2,252,200 912,600 1,339,600 Sudan 7,246,600 2,955,000 4,291,600 8,807,000 4,340,000 4,467,000 Tanzania 7,665,800 2,874,600 4,791,200 9,130,200 4,272,000 4,858,200 Uganda 5,763,200 716,000 5,047,200 8,383,800 1,191,800 7,192,000 Total EAPP 61,539,400 19,027,829 42,511,571 76,913,000 27,527,840 49,385,160 WAPP Benin 1,529,872 613,479 916,393 1,964,237 876,050 1,088,187 Burkina Faso 2,698,347 493,798 2,204,550 3,628,240 827,239 2,801,001 Cote d' Ivoire 3,459,608 1,556,824 1,902,784 4,176,910 2,080,101 2,096,809 Gambia 319,017 171,950 147,067 415,923 257,040 158,883 Ghana 4,405,136 2,105,655 2,299,481 5,295,010 2,917,551 2,377,460 Guinea 1,890,534 623,876 1,266,658 2,448,382 932,833 1,515,548 Guinea-Bissau 282,689 83,676 199,013 345,071 107,317 237,754 Liberia 580,054 337,011 243,043 783,743 507,866 275,878 Mali 2,275,877 694,143 1,581,735 3,010,905 1,098,980 1,911,925 Mauritania 617,372 249,418 367,954 818,215 352,651 465,564 Niger 2,432,571 408,672 2,023,899 3,230,390 623,465 2,606,925 Nigeria 25,753,154 12,413,020 13,340,134 32,669,214 18,262,091 14,407,123 Senegal 2,372,086 986,788 1,385,298 3,041,419 1,359,514 1,681,905 Sierra Leone 1,173,485 477,608 695,877 1,473,488 710,221 763,267 Togo 1,079,998 433,079 646,919 1,412,145 669,357 742,788 Total WAPP 50,869,800 21,648,997 29,220,804 64,713,292 31,582,276 33,131,016 11 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA No. of households 2005 No. of households 2015 Total Urban Rural Total Urban Rural CAPP Cameroon 3,452,293 1,884,952 1,567,341 4,277,460 2,681,968 1,595,493 CAR 847,541 322,065 525,475 979,141 395,573 583,568 Chad 1,931,414 488,648 1,442,766 2,579,209 786,659 1,792,550 Congo, Rep. 720,454 433,713 286,741 950,719 600,855 349,865 Eq. Guinea 105,807 41,159 64,648 128,759 52,920 75,839 Gabon 278,861 233,128 45,733 340,578 298,687 41,891 Total CAPP 7,336,370 3,403,665 3,932,704 9,255,866 4,816,660 4,439,206 Island states Cape Verde 83,645 47,928 35,716 88,103 56,650 31,453 Madagascar 3,721,200 1,005,200 2,716,000 4,762,600 1,463,200 3,299,400 Mauritius 246,120 104,355 141,765 265,688 117,168 148,519 Total islands 4,050,965 1,157,484 2,893,482 5,116,390 1,637,018 3,479,372 Total SSA 160,561,935 59,811,259 100,750,676 199,104,549 85,362,310 113,742,239 Source: United Nations, World Population Prospects Three targets for electricity access Tables 2.6 (electricity access rates) and 2.7 (new connections) summarize the assumptions related to social demand in the three access scenarios. 2005 access 2005 access rates were obtained from the World Bank (World Development Indicators), based on the United Nations Statistical Office Annual Energy Questionnaire. In our opinion, this is the most comprehensive compilation of electricity access rates in Sub-Saharan Africa, and we have therefore relied on it to the largest extent possible. Where data are missing, we have turned to other sources, listed in the notes at the bottom of the tables showing current and projected population and access rates. Then, the number of new connections needed to maintain 2005 urban and rural electricity access rates is estimated. This represents new connections needed to keep up with population growth in rural and urban areas. It should be noted that the higher access rates in urban areas—along with urbanization trends—generally imply a slight increase in countrywide access rates. In addition, in those cases where population is projected to fall in rural areas, we assume zero expansion. Regional targets for access This ambitious yet realistic target lies between that of the other scenarios and is based on stakeholder discussions. In SAPP and EAPP countries, this involves increasing the regional average access to 35 percent. In SAPP countries, the target implies an average annual increase of 1.2 percentage points. In the EAPP it implies an annual increase of 2 percentage points, excluding Egypt. Given Egypt’s current access rates of 100 percent in urban areas and 87.5 percent in rural areas, it is assumed that the country will achieve 100 percent access, even in rural areas. In CAPP, the regional target is 44 percent and in WAPP, 12 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA the regional target is 53 percent access (implying an annual increase of 1.2 percentage points in both regions). National targets This scenario reflects the electricity access targets as defined by the national authorities. For SAPP, the number of new connections required to meet national governments’ electrification targets is estimated. Various adjustments or approximations have been made for those countries with a target set for a year other than 2015. For the Democratic Republic of Congo and Zimbabwe, no national targets could be identified; we thus approximated these by adding the average increase in access rates implied by the national targets of the remaining countries. For the Nile Basin–East African countries, the new connections needed to attain national urban and rural access targets were estimated. Stated national targets were applied to the greatest extent possible. It is assumed that Egypt will achieve 100 percent rural access by 2015 (as noted, urban access is already universal). For Ethiopia, the national target of 50 percent access by 2012 implies an expansion rate that would lead to 60 percent access by 2015. For Kenya, the national rural target of 20 percent access by 2010 is applied in the same way, while the regional urban target is 100 percent access. For Rwanda, the growth rate implied by the national target of 20 percent access by 2020 is employed, with urban access just keeping up with very rapid urban population growth. Sudan’s stated national target of 90 percent access “in coming years” is used. For Uganda, the growth rate implied by the rural target of 10 percent by 2012 is used, while the regional target of 100 percent urban access is applied. For Djibouti and Burundi, no national targets have been set. Accordingly, the regional average increase in access in percentage points during the period is applied to these two countries. For WAPP countries, ECOWAS targets are used (ECOWAS, 2006). Table 2.6 Targets for electricity access under various scenarios (percent of population) 2005 access Regional targets National targets Total Urban Rural Total Urban Rural Total Urban Rural SAPP Angolaa 14 26 4 24 42 9 46 84 15 Botswana 30 45 9 40 59 14 100 100 100 Congo, Dem. Rep. 8 16 2 20 37 8 39 76 12 Lesothob 6 23 1 17 68 4 35 121 13 Malawi 7 29 1 18 76 1 15 56 3 Mozambique 14 26 2 23 41 5 20 37 5 Namibiac 37 75 12 45 86 19 53 95 25 South Africac 71 80 50 79 87 66 100 100 100 Zambia 20 45 3 29 57 11 29 50 15 Zimbabwe 41 87 8 49 99 14 67 100 44 Total SAPP 26 45 11 37 60 17 51 79 27 EAPP Burundi 6 45 0 25 100 13 31 67 25 Djibouti 31 34 5 50 54 11 53 56 29 13 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA 2005 access Regional targets National targets Total Urban Rural Total Urban Rural Total Urban Rural Egypt 93 100 87 100 100 100 100 100 100 Ethiopia 15 76 0 32 100 16 60 100 50 Kenya 27 48 4 42 72 10 67 100 32 Rwanda 16 39 1 29 66 4 18 39 4 Sudan 34 56 12 50 77 23 60 100 21 Tanzania 13 27 1 30 63 2 29 38 22 Uganda 8 44 2 27 100 15 25 100 13 Total EAPP 35 64 19 50 84 31 60 88 45 WAPPd Benin 25 50 6 35 68 9 50 100 11 Burkina Faso 9 40 0 20 84 1 23 100 6 Cote d'Ivoire 54 86 23 63 96 30 73 100 46 Gambia 59 82 21 66 88 31 79 100 37 Ghana 55 82 21 62 89 30 76 100 37 Guinea 21 54 2 31 77 3 40 100 3 Guinea-Bissau 14 41 2 25 74 4 33 100 6 Liberia 1 0 2 13 1 4 66 100 6 Mali 15 37 2 25 60 5 39 100 7 Mauritania 23 50 3 34 73 4 46 100 6 Niger 7 37 0 18 93 1 20 100 1 Nigeria 59 84 28 67 89 40 82 100 49 Senegal 34 69 6 44 88 9 51 100 10 Sierra Leone 41 82 2 46 91 4 51 100 6 Togo 21 41 2 30 58 5 50 100 8 Total WAPP 45 75 16 53 85 23 66 100 34 CAPP Cameroon 61 85 21 68 90 31 71 84 49 Central Afr Rep 3 8 0 15 36 1 34 84 1 Chad 3 9 0 15 46 1 26 84 0 Congo, Rep. 22 35 0 33 52 0 53 84 0 Equatorial Guinea 11 26 0 22 54 0 35 84 0 Gabon 82 85 54 92 93 83 91 84 82 Total CAPP 35 59 8 44 73 12 53 84 19 Island states Cape Verde 55 72 24 62 80 31 84 100 54 Madagascar 18 48 5 36 91 12 36 89 13 Mauritius 100 100 100 100 100 100 100 100 100 Total islands 23 52 9 40 92 16 40 90 17 Source: World Bank, World Development Indicators; IEA, World Energy Outlook 2006 – Electricity Access; UNEP (2002). a. Data are for total-access rates only; the rural rate is based on average spread between the total and rural rates in Sub-Saharan Africa, which also yields an implicit estimate for urban electrification. b. Private Solutions for Infrastructure in Lesotho (Country Framework Report 34354, World Bank). c. Power Sector Reform in Africa: Assessing Impact on Poor People (ESMAP report 306/05). d. ECOWAS (2006). 14 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 2.7 New connections under various scenarios 2005 access Regional targets National targets Total Urban Rural Total Urban Rural Total Urban Rural SAPP Angola 189,517 177,270 12,247 626,062 495,006 131,056 1,544,540 1,270,679 273,861 Botswana 4,082 4,082 0 36,618 31,052 5,566 239,160 110,534 128,626 Congo, 405,946 387,713 18,234 2,158,523 1,681,848 476,675 4,986,424 4,118,393 868,031 Dem. Rep. Lesotho 1,834 1,834 0 41,612 34,804 6,808 104,828 73,248 31,580 Malawi 80,533 76,926 3,608 417,587 412,275 5,312 319,224 270,837 48,387 Mozambique 200,355 200,355 0 621,455 546,922 74,533 514,437 445,893 68,544 Namibia 32,212 32,148 64 69,050 51,026 18,024 103,186 67,936 35,250 South Africa 411,428 411,428 0 1,219,077 854,276 364,801 3,189,412 1,612,810 1,576,602 Zambia 126,822 122,068 4,754 386,443 255,967 130,476 380,456 177,410 203,046 Zimbabwe 182,538 182,538 0 396,442 313,325 83,117 896,805 326,552 570,253 Total SAPP 1,635,267 1,596,360 38,907 5,972,869 4,676,501 1,296,368 12,278,472 8,474,292 3,804,180 EAPP Burundi 68,320 66,212 2,108 473,983 231,661 242,322 610,986 132,306 478,680 Djibouti 10,887 10,887 0 46,190 45,142 1,049 52,787 48,058 4,729 Egypt 2,498,342 1,478,600 1,019,742 3,746,409 1,478,600 2,267,809 3,746,409 1,478,600 2,267,809 Ethiopia 1,033,168 1,022,756 10,411 4,316,696 1,937,724 2,378,972 9,699,988 1,937,724 7,762,265 Kenya 830,330 818,995 11,335 2,179,612 1,934,176 245,436 4,426,566 3,221,815 1,204,751 Rwanda 201,891 201,891 0 483,356 446,265 37,092 239,596 201,891 37,705 Sudan 798,611 777,563 21,048 2,198,508 1,677,706 520,802 3,110,220 2,681,012 429,208 Tanzania 382,227 381,490 737 1,944,457 1,895,796 48,661 1,839,775 823,674 1,016,101 Uganda 260,351 208,876 51,475 1,845,192 877,476 967,716 1,680,515 877,476 803,039 Total EAPP 6,084,126 4,967,270 1,116,856 17,234,404 10,524,545 6,709,859 25,406,842 11,402,556 14,004,287 15 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA WAPP Benin 139,684 130,235 9,449 338,376 290,292 48,084 635,693 571,764 63,929 Burkina 133,569 132,376 1,193 508,094 496,918 11,176 636,793 631,201 5,592 Faso Cote d' 493,151 449,496 43,656 871,445 660,116 211,329 1,285,820 742,790 543,030 Ivoire Gambia 72,255 69,774 2,481 102,539 84,114 18,424 155,214 116,041 39,173 Ghana 685,300 669,002 16,298 1,088,276 852,221 236,055 1,792,067 1,182,491 609,576 Guinea 169,952 166,219 3,733 400,901 379,423 21,478 621,287 597,188 24,099 Guinea- 10,156 9,575 581 49,891 45,852 4,039 77,215 73,428 3,786 Bissau Liberia 621 128 493 96,202 6,238 89,964 512,237 507,613 4,624 Mali 157,054 149,790 7,264 458,506 403,796 54,710 886,285 842,148 44,138 Mauritania 53,747 51,306 2,440 142,323 132,492 9,832 240,357 228,690 11,668 Niger 79,566 78,400 1,166 441,127 429,473 11,654 479,434 474,300 5,134 Nigeria 5,228,457 4,930,767 297,690 7,762,673 5,726,037 2,036,636 12,518,720 7,797,915 4,720,806 Senegal 274,605 256,809 17,796 583,312 519,770 63,542 785,043 679,618 105,426 Sierra Leone 205,924 204,913 1,011 277,461 269,994 7,466 345,993 332,753 13,240 Togo 99,647 97,346 2,301 230,519 212,066 18,453 510,621 490,928 19,693 Total WAPP 7,803,687 7,396,135 407,552 13,351,645 10,508,801 2,842,844 21,482,781 15,268,867 6,213,913 CAPP Cameroon 680,187 674,275 5,912 977,402 810,180 167,222 1,112,508 658,183 454,325 CAR 6,055 5,881 174 122,722 115,646 7,076 308,692 306,516 2,176 Chad 28,363 28,013 350 327,974 317,986 9,988 616,852 614,860 1,991 Congo, Rep. 57,978 57,911 66 163,163 162,836 326 354,860 354,444 416 Eq. Guinea 3,021 3,017 4 17,875 17,839 36 33,922 33,893 29 Gabon 55,792 55,791 1 90,583 80,555 10,028 86,592 52,504 34,089 Total CAPP 831,395 824,888 6,507 1,699,718 1,505,042 194,675 2,513,427 2,020,401 493,026 Island states Cape Verde 6,290 6,280 10 12,899 10,596 2,303 31,655 22,142 9,513 Madagascar 249,261 218,924 30,337 1,107,461 855,885 251,576 1,099,554 821,762 277,792 Mauritius 19,567 12,813 6,754 19,567 12,813 6,754 19,567 12,813 6,754 Total 275,118 238,017 37,101 1,139,928 879,295 260,633 1,150,776 856,717 294,059 islands Total SSA 16,629,592 15,022,670 1,606,922 39,398,563 28,094,184 11,304,379 62,832,299 38,022,833 24,809,466 Source: Authors’ original research. 16 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Total demand in 2015 Our demand projections for 2015 are summarized in table 2.8. The social demand figures refer to the national targets scenario, which will form the basis for the trade versus no-trade scenarios that are discussed below. In some countries, growth in market demand overwhelms social demand (Mozambique, South Africa, Egypt), whereas the situation is the opposite in other countries (Burundi, Ethiopia, Mali, Central African Republic, Chad). Table 2.8 Demand in study countries (TWh) Total net demand in Social demand with Total net demand Market demand 2015* 2005 national targets 2015 2015 Southern Africa Power Pool Angola 2.1 6.0 1.9 7.9 Botswana 2.4 4.0 0.2 4.2 Congo, Dem. Rep. 4.7 7.4 6.2 13.6 Lesotho 0.4 0.8 0.1 0.9 Malawi 1.3 1.9 0.4 2.3 Mozambique 11.2 15.7 0.7 16.4 Namibia 2.6 4.2 0.1 4.3 South Africa 215.0 316.0 3.2 319.2 Zambia 6.3 9.0 0.4 9.3 Zimbabwe 12.8 18.0 0.8 18.7 Total 258.8 383.0 14.0 396.9 Nile Basin–East Africa Power Pool Burundi 0.2 0.3 0.5 0.7 Djibouti 0.2 0.3 0.1 0.4 Egypt 84.4 119.9 3.4 123.3 Ethiopia 2.1 3.4 7.4 10.7 Kenya 4.6 6.8 5.2 12.0 Rwanda 0.1 0.2 0.3 0.5 Sudan 3.2 5.2 3.9 9.2 Tanzania 4.2 6.2 1.7 7.9 Uganda 1.6 2.5 1.7 4.2 Total 100.6 144.8 24.2 169.0 Western Africa Power Pool Benin 0.6 0.9 0.8 1.7 Burkina Faso 0.5 0.6 0.9 1.5 Cote d’Ivoire 2.9 4 1.4 5.4 Gambia 0.1 0.2 0.2 0.4 Ghana 5.9 10.8 2 12.8 Guinea 0.7 1.3 0.8 2.2 Guinea-Bissau 0.1 0.1 0.1 0.2 Liberia 0.3 0.6 0.7 1.3 Mali 0.4 0.6 1.2 1.8 17 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Total net demand in Social demand with Total net demand Market demand 2015* 2005 national targets 2015 2015 Mauritania 0.2 0.5 0.3 0.8 Niger 0.4 0.6 0.7 1.2 Nigeria 16.9 45.6 13.6 59.2 Senegal 1.5 2.5 1 3.5 Sierra Leone 0.2 0.5 0.5 1 Togo 0.6 0.8 0.7 1.5 Total 31.3 69.6 24.8 94.3 Central Africa Power Pool Cameroon 3.4 5.2 1.2 6.4 Central African Rep. 0.1 0.1 0.4 0.6 Chad 0.1 0.1 0.8 1.0 Congo, Rep. 5.84 9.8 0.5 10.3 Equatorial Guinea 0.03 0.1 0.05 0.1 Gabon 1.2 1.7 0.1 1.8 Total 10.7 17.1 3.1 20.2 Island states Cape Verde 0.04 0.1 0.04 0.1 Madagascar 0.8 1.1 1.4 2.6 Mauritius 0.2 0.4 0.02 0.4 Total 1.1 1.6 1.5 3.0 Source: Authors’ original research. Note: * Assuming all suppressed demand is met. 18 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA 3 Estimating supply We now estimate electricity supply, which consists of power generation plants, a transmission grid and a distribution grid. In addition to a grid-based supply of electricity, rural areas may have off-grid supply. The available technologies for power generation are hydropower (large-scale, small-scale and pumped storage), thermal (fueled by coal, natural gas, diesel, heavy fuel oil as well as nuclear power), solar PV and geothermal power. We describe each technology in terms of its capacity and cost. The starting point is the existing generation capacity as of 2005. Capacity expansion is possible through investments. In addition, refurbishment can extend the life span of existing capacity. Costs consist of capital and variable operating costs (that is, the cost of fuel and operations and maintenance, O&M). Investments are also necessary to extend the transmission and distribution grid, and refurbishments are necessary to keep the existing assets in working condition. The following sections outline assumptions for (a) installed capacity, refurbishment requirements and costs, the investment potential of different generation technologies, as well as costs of generation, based on fuel prices; (b) transmission and distribution assets; and (c) the supply side of social demand. Generation capacity Installed capacity, refurbishment requirements and costs The stock of operating generation assets in 2015 is assumed to equal installed (design) capacity in 2005, minus the assets that must be refurbished by 2015 to continue effective production. Refurbishment refers to efforts to either prolong the life of an outdated plant whose operating life is coming to an end by restoring it to full operational status, or to repair generation assets that have been seriously damaged, for example due to war. Refurbishments do not include costs for ordinary maintenance and repair work. Table 3.1 reports the existing generation capacity (as of 2005) and refurbishment requirements in 2015. (Note that the column for refurbishment in table 3.1 also includes capacity that was installed but not operational in 2005, but can be brought online before 2015.) For further documentation, see appendixes 1 and 2. The amount that is refurbished will be determined within the least-cost expansion model. For coal-fired power plants we base the assumptions for refurbishment costs on estimated costs of planned refurbishments of South African coal plants. The estimated average refurbishment cost of these plants is $272/kW. This figure will be used for South Africa. We cross-checked the estimate against international sources of refurbishment costs for thermal plants of similar age. The figure fits well with 19 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA other international experiences from both Europe and Asia.6 Therefore, as an estimate for refurbishment costs of coal-fired plants in the remaining countries we will use $300/kW. For gas-fired power plants, which carry a significantly lower overnight investment cost, a figure of $300/kW would be too high for refurbishment, as it would constitute almost half of the costs of new investments. We estimate the overnight refurbishment costs of gas-fired power plants to be 15 percent of the investment costs, or $100.5/kW. Hydropower refurbishment costs are to a large extent site-specific; see appendix 2 for further documentation. Table 3.1 Asset stock and refurbishment requirements for electricity generation in study countries Country Installed capacity, 2005 (MW) Refurbishment requirements (MW) Hydro Thermal Total Hydro Thermal Total SAPP Angola 830 320 1,150 306 103 409 Botswana 0 132 132 0 12 12 Congo, Dem. Rep. 2,451 0 2,451 2,455 0 2,455 Lesotho 72 0 72 0 0 0 Malawi 252 21 273 100 0 100 Mozambique 2,157 204 2,361 100 80 180 Namibia 240 144 384 0 120 120 South Africa* 2,242 33,232 35,474 600 21,432 22,032 Zambia 1,838 90 1,928 1,670 0 1,670 Zimbabwe 750 1,295 2,045 760 1,075 1,835 EAPP Burundi 26 5 31 18 0 18 Djibouti 0 85 85 0 0 0 Egypt 2,804 16,475 19,279 0 2,093 2,093 Ethiopia 1,115 76 1,191 335 0 335 Kenya 641 520 1,161 383 77 460 Rwanda 32 20 52 24 0 24 Sudan 1,623 1,129 2,752 0 133 133 Tanzania 553 332 885 305 0 305 Uganda 441 106 547 191 0 191 WAPP Benin 31 165 196 31 90 121 6 The International Atomic Energy Agency (2002) reports on energy supply options in Poland and estimates the rehabilitation costs of Polish power plants. Poland has an aging fleet of coal power plants (in 1996 coal-fired power plants constituted more than 95 percent of installed capacity). The age structure is similar to South Africa in that more than half of the installed capacity was constructed in the 1970s. The Polish Power Grid Company estimated rehabilitation requirements to be in the range of $50– $350 per kW of installed capacity. The Indian Ministry of Power issued guidelines in 2004 for undertaking life extension (LE) works on its thermal units, where the cost guidelines were to restrict the cost of LE works between 0.8 and 1.25 crore rupees/MW, which is, at current exchange rates, approximately $178–$278/kW. The Ministry of Power’s Performance Review of Thermal Power Stations (2006) reported a life extension/rehabilitation of 106 thermal units with combined generating capacity of 10,413 MW at a total cost of 9,200 crore rupees, or approximately $200/kW. 20 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Country Installed capacity, 2005 (MW) Refurbishment requirements (MW) Hydro Thermal Total Hydro Thermal Total Burkina Faso 33 183 216 63 117 180 Cote d’Ivoire 601 612 1,213 601 114 715 Gambia 0 103 103 0 81 81 Ghana 1,198 705 1,903 160 30 190 Guinea 103 108 211 28 64 92 Guinea-Bissau 0 22 22 0 22 22 Liberia 64 9 73 64 9 73 Mali 150 108 258 46 20 66 Mauritania 30 110 140 0 73 73 Niger 0 105 105 0 105 105 Nigeria 1,900 6,014 7,914 2,850 5,003 7,853 Senegal 66 496 562 0 291 291 Sierra Leone 50 66 116 0 55 55 Togo 31 60 91 31 60 91 CAPP Cameroon 722 222 944 596 98 694 Central African Republic 18 24 42 18 24 42 Chad 0 29 29 0 29 29 Congo, Rep. 99 28 127 129 28 157 Equatorial Guinea 0 44 44 0 34 34 Gabon 163 244 407 163 244 407 Island states Cape Verde 0 78 78 0 78 78 Madagascar 82 187 269 82 0 82 Mauritius 59 618 677 0 483 483 Source: Authors’ original research. See appendixes 1 and 2 for further documentation. Note: Refurbishment figures may exceed the original installed capacity, since capacity is often increased in the course of refurbishment. * Thermal capacity includes nuclear (1,800 MW) in South Africa. Generation investment: Unit cost and potential Investments in conventional thermal power plants (fueled by coal, natural gas, heavy fuel oil or diesel) and hydropower are determined endogenously in the model.7 In addition, it is possible to invest in solar PV, diesel and minihydro plants in rural areas. These will be discussed further under supply side assumptions related to social demand. Hydropower unit cost and potential Hydropower potential and investment costs are largely site specific. 7 Investments in nuclear power are not included in the model. In general, nuclear power investments are more expensive than alternatives. In some countries nuclear power is chosen for political reasons (for example, energy security). In other countries, nuclear power is not an option for political reasons. 21 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Our unit investment costs for each country are based on estimated investment costs for actual planned hydropower projects in the country. Where there are several planned projects, the unit investment costs used in the least-cost expansion model is the weighted, average-unit investment cost of planned projects in the country, where weights reflect plants’ planned capacity. We want to emphasize that this is not an exhaustive list of planned hydropower projects in any region, but a subset for which good estimates of investment costs could be obtained. Some countries have no potential for larger hydropower plants, and thus no planned investments. The investment cost estimates are documented in appendix 2. Thermal power unit cost and potential Thermal power plant technology is generic. Table 3.2 Unit investment costs for thermal power The unit costs are therefore similar across generation countries. For thermal power unit investment Source Unit cost, Unit cost, costs, we rely on values used in the power market gas coal models developed by our firm, Econ Pöyry. These ($/kW) ($/kW) values agree with accepted international values International Energy Agency— 400–800 1,000– Projected Costs of Generating 1,500 from several sources (table 3.2). Electricity (2005) U.S. Energy Information 580 975 The capital investment costs for small diesel Administration (2005) units are derived from earlier project work carried Royal Academy of Engineering—The 570 1,400 out by Econ Pöyry and Power Planning Associates Costs of Generating Electricity (2004) from Sub-Saharan Africa, and cross-checked Actual value used 670 1,100 against producers of mini-diesel plants such as Note: At current exchange rates. Jacobsen and Caterpillar. The unit capital cost used in the model is $335/kW. Nevertheless, fuel availability limits the feasibility of certain investments. Therefore, we allow for investments in natural gas-fired power plants only in countries that have domestic gas reserves or that are (or will be) connected to a gas pipeline (such as the West Africa Gas Pipeline). Similarly, investments in coal-fired power plants are limited to countries that have domestic coal resources. Oil products (diesel and heavy fuel oil) are assumed to be available in most countries. Fuel price assumptions For coal, the world market price of $52/ton is the benchmark for fuel costs, based on the projected price for 2015 from the U.S. Energy Information Administration’s Annual Energy Outlook 2006 (EIA, 2006). The world market price indicates the alternative value of domestic coal resources, that is, what the country could earn by selling the coal in the world market instead of using it for domestic electricity production. The price of oil products (diesel and HFO) is related to the world market price of crude oil. Our base case scenario is based on the oil price of $47/barrel.8 The world market price must, however, be adjusted for transport costs. For EAPP and SAPP, we use country-specific data for all countries, as provided by 8 Sensitivities that analyze the impact of different oil prices are presented in chapter 6. 22 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Meschies (2005). For CAPP and WAPP, the transport cost assumptions are based on Arthur Energy Advisors (2007). Estimating the price of gas is particularly challenging, since the price and availability of gas depends heavily on infrastructure, especially pipelines. First, we assume that only countries with domestic gas reserves or countries that are connected to a gas pipeline can invest in gas-fired power plants. Second, we assume that for countries that have potential to export gas (for example, Nigeria), the gas price is related to the world market price of oil (that is, gas has an alternative value); for countries that have no infrastructure to export gas (for example, Mauritania), the gas price is related to domestic production costs. Third, we adjust the final price of gas (delivered to the power plant) to reflect the transportation costs: countries with domestic resources are assumed to have lower gas price than countries that import gas through a pipeline (such as Benin, Togo, and Ghana). Transmission and distribution Existing assets, refurbishment requirements, and costs The existing stock of transmission and distribution (T&D) lines in each country are shown in table 3.3. Transmission lines are those with voltages of 66 kilovolts (kV) or greater, while the rest are defined as distribution lines. The refurbishment requirements are based on the asset age: we assume that lines older than 30 years need refurbishment at a unit cost of 60 percent of replacement value. In some countries, transmission systems are quite old. For example, in Niger, 96 percent of the system is more than 30 years old; in Zimbabwe, 85 percent; in Zambia, 80 percent, and in the Democratic Republic of Congo, 61 percent. Distribution systems are younger on average. For documentation of data sources for transmission and distribution, please see appendix 3. Table 3.3 Asset stock and refurbishment requirements related to transmission and distribution in study countries Refurbishment Percentage of cost Asset value assets older than ($ thousands) Country Type Line (km) ($ thousands) 30 years (Asset value x (weighted average) Assets older than 30 years x 60%) Southern Africa Power Pool Angola Transmission — — — 153,998 Distribution — — — 153,794 Botswana Transmission 2,664 280,258 43 72,516 Distribution 3,560 184,172 35 38,428 Congo, Dem. Rep. Transmission 3,050 421,477 61 154,279 Distribution 5,295 367,875 69 152,942 Lesotho Transmission 550 48,527 14 4,203 Distribution 3,980 57,115 37 12,578 Malawi Transmission 1,944 171,219 29 30,214 Distribution 810 46,453 44 12,240 Mozambique Transmission 3,116 416,713 35 87,187 Distribution 4,680 272,221 46 74,733 23 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Refurbishment Percentage of cost Asset value assets older than ($ thousands) Country Type Line (km) ($ thousands) 30 years (Asset value x (weighted average) Assets older than 30 years x 60%) Namibia Transmission 7,813 1,145,190 51 351,935 Distribution 10,457 503,848 29 88,043 South Africa Transmission 68,704 13,337,276 30 2,411,721 Distribution 404,861 26,858,652 25 4,032,881 Zambia Transmission 5,213 783,970 80 374,012 Distribution 20,645 810,452 58 280,102 Zimbabwe Transmission 6,717 1,401,851 85 717,069 Distribution 36,900 1,586,107 60 570,999 Nile Basin–East Africa Power Pool Transmission 270 31,962 20 3,835 Burundi Distribution 860 31,152 30 5,607 Transmission 5 354 0 0 Djibouti Distribution 579 15,534 20 2,026 Transmission 34,897 3,681,478 52 1,149,869 Egypt Distribution 19,725 1,309,630 30 235,733 Transmission 6,488 595,482 91 326,145 Ethiopia Distribution 17,480 1,273,742 46 353,166 Transmission 2,615 362,005 39 83,834 Kenya Distribution 34,927 1,727,075 29 303,628 Transmission 349 38,839 42 9,875 Rwanda Distribution 657 37,253 26 5,832 Transmission 3,400 416,960 53 131,913 Sudan Distribution 41,200 1,152,698 31 215,774 Transmission 5,511 643,218 52 201,698 Tanzania Distribution 14,550 328,675 44 87,040 Transmission 1,169 82,825 88 43,777 Uganda Distribution 13,412 540,190 56 182,514 Western African Power Pool Transmission 532 37,136 60 13,369 Benin Distribution 6,899 540,641 32 103,803 Transmission 1,161 201,774 50 60,532 Burkina Faso Distribution 6,396 321,447 32 61,718 Transmission 4,396 701,106 41 172,472 Cote d'Ivoire Distribution 32,177 895,940 36 193,523 Transmission – – – – Gambia Distribution 704 18,382 50 5,515 Transmission 5,374 562,819 53 178,976 Ghana Distribution 10,390 949,872 17 96,887 Transmission 415 64,611 10 3,877 Guinea Distribution 1,817 43,967 50 13,190 24 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Refurbishment Percentage of cost Asset value assets older than ($ thousands) Country Type Line (km) ($ thousands) 30 years (Asset value x (weighted average) Assets older than 30 years x 60%) Transmission – – – – Guinea-Bissau Distribution 348 8,496 50 2,549 Transmission 420 29,719 – – Liberia Distribution 1,763 41,959 – – Transmission 956 114,123 45 30,813 Mali Distribution 6,135 574,830 36 124,163 Transmission 300 35,937 – – Mauritania Distribution 1,780 135,065 50 40,520 Transmission 910 213,096 96 122,743 Niger Distribution 325 18,284 41 4,498 Transmission 10,919 2,236,773 51 684,453 Nigeria Distribution 184,880 7,370,061 36 1,591,933 Transmission 1,530 168,661 14 14,168 Senegal Distribution 448,928 5,794,945 30 1,043,090 Transmission 308 36,534 - - Sierra Leone Distribution 1,435 36,171 50 10,851 Transmission 820 86,737 37 19,256 Togo Distribution 555 33,654 20 4,038 Central African Power Pool Transmission 1,860 274,834 48 79,152 Cameroon Distribution 14,635 412,998 31 76,818 Transmission 195 30,695 50 9,209 Central African Rep. Distribution 624 9,205 50 2,762 Transmission – – – – Chad Distribution 355 14,156 37 3,143 Transmission 1,128 131,398 – – Congo, Rep. Distribution 4,148 73,367 – – Transmission – – – – Equatorial Guinea Distribution 218 7,140 50 2,142 Transmission 556 58,454 50 17,536 Gabon Distribution 5,748 67,836 50 20,351 Island states Transmission 80 14,154 50 4,246 Cape Verde Distribution 146 4,760 50 1,428 Transmission 138 7,639 50 2,292 Madagascar Distribution 1,783 82,321 28 13,830 Transmission 252 21,759 20 2,611 Mauritius Distribution 5,060 322,019 20 38,642 Source: Authors’ original research. See appendix 3 for details and data sources. 25 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Note: The weighted average asset age is the average percentage, across voltage levels, of assets older than 30 years. Weights correspond to the fraction of a voltage level’s asset value to total asset value. Investments in transmission and distribution Here we present assumptions for transmission and distribution costs related to market demand. Subtransmission and distribution costs related to social demand are described later. Domestic transmission and distribution For domestic T&D infrastructure, we have assumed that T&D stock will grow at half the rate of market demand, an assumption used by the regulatory authorities in both Norway and Great Britain. In addition, we have assumed that the original asset stock will depreciate at 3 percent per year before 2015.9 Combined with annual O&M costs of 2 percent, we assume that the relevant annual investment to maintain capacity over the next ten years will be 5 percent of original T&D stock. Cross-border transmission lines Investment volume in cross-border transmission is an endogenous decision variable determined by the model. The set of potential cross-border transmission lines is exogenous. For example, we exogenously determine the feasibility of building a line from Angola to Zambia. Cross-border transmission can be built in two high-voltage levels—namely, the two highest voltage levels historically used in a given country. These have different unit costs and different average line ratings in MW. The average line ratings have been used to set upper limits for the capacity that may be built between countries. Finally, since the maximum capacity varies among countries (since they are using different voltage levels for transmission), the maximum capacity (in MW) for a line between the two countries is assumed to be determined by the highest capacity of the two countries. The unit costs—per km and per MW—of investing in these lines have been provided for each country. The unit costs are country specific (refer to appendix 3 for documentation). For lines between countries, we have used weighted unit costs, in which the weights are based on the approximate share of the line built in each country. The total cost of a cross-border line depends on the unit cost per km per MW and the line length (in km). The line length depends on the distance between load centers in the two countries, or between a load center and a possible power-plant location (where plans are known). For instance, the relevant distance for a line between Mozambique and South Africa is the distance from Cahora Bassa North to Johannesburg. Similarly, the Mmamabula power plant in Botswana will most likely be connected to the South African transmission grid in the vicinity of the Matimba power plant in South Africa; therefore, the distance between the two is the relevant distance. Based on the total cost of a line (that is, unit cost per km per MW, multiplied by the line length in km), the model decides how much capacity to build between the countries. 9 This is based on a 30-year lifetime, assuming that technological improvements will allow for slightly cheaper alternatives. 26 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Supply-side assumptions related to social demand Some aspects of the model are particular to social demand, such as the connection cost for grid- connected and off-grid technologies, the distribution of social demand connections across grid-connected and off-grid technologies, and distribution among off-grid technologies. This section presents the assumptions and data related to the supply side of social demand. Technology mix for social demand: New connections New urban connections are assumed to be connected to the grid. New rural connections are covered either by grid expansion (70 percent), minihydro (20 percent), diesel (5 percent), or solar PV (5 percent). The basis for these assumptions is professional judgment; allowing the model to determine the technology mix for new connections endogenously would rest too much on insignificant differences in unit costs and could result in coincidental outcomes. To compare the associated costs, we assume the same shares in all simulations. Urban connection costs The urban population in Sub-Saharan Africa is growing rapidly; in fact the region has the highest urban growth rate in the world (Karezeki and Majoro, 2002). For urban electrification, the majority of costs are the result of connection to existing low-voltage distribution networks and the installation of credit meters. Econ Pöyry (2005) and Modi (2004) both estimate connection costs at $100 per connection. The additional costs of low- and medium-voltage distribution lines are substantially lower per customer (given the higher population density and electrification rates in urban areas). USAID (2004) and Gaunt (2005) provide estimates of unit costs of urban electrification programs in South Africa of approximately $420 and $350, respectively. We use a similar but slightly higher figure of $500 per connection for the remaining countries, since the cost of urban electrification is to a large extent generic. Rural connection and generation costs We allow for four technical solutions for rural electrification (grid extension, mini- and microhydro, diesel generation, and solar PV) and estimate the connection costs for each (table 3.4). Grid extension and mini- and microhydro These technical solutions both require the installation of low- and medium-voltage subtransmission lines (for example, 11 kV or 33 kV) to deliver the electricity to the local grid. Thus, the cost per connection in both cases includes subtransmission lines, a distribution network, and actual connection costs. Note that the generation costs of mini- and microhydro stations are included in the connection cost estimate. It is difficult to estimate the cost per connection for a large-scale and relatively effective rural electrification program. First, this cost depends on area-specific characteristics such as population density, distance from the grid, and terrain. Additionally, there are relatively few successful large-scale rural electrification programs in Africa from which we can extract data. 27 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Probably the most successful program is in South Africa, where Gaunt (2005) estimated that between 1995 and 2001, approximately 500,000 new connections were made annually, and costs fell from ZAR 3,600 in 1995 to ZAR 2,600 in 2001 (less than $350 at current exchange rates). But recognizing country- specific challenges—whether geographic or institutional—this is probably a low estimate. Our estimates for cost per connection are based on the extensive global review of rural electrification projects and studies by Zomers (2001), who arrives at an estimated grid extension cost of nearly $1,500 per connection (distribution cost of $1,200, transmission cost of $125, and service connection of $100). This corresponds with the estimates produced in a particularly ambitious study on the costs of grid-based rural electrification in Tanzania by Econ Pöyry (2006), which provides detailed estimates of costs per connection for several districts in 19 different regions in Tanzania, all with varying access rates and population densities. The resulting average unit cost per connection in the 19 different regions lies between $544 and $4,336. The average unit cost across all regions was $1,547. Studies from Zambia (Sanghvi, 2005) and Mozambique (Bergman and Davies, 2005) estimate a cost per connection of around $2,000, while a World Bank Project Appraisal Document from Nigeria arrives at an estimate of $1,000. Based on Econ Pöyry’s work, the findings of Zomers, and discussions with World Bank staff, we assume that the cost for large-scale electrification programs converges to approximately $1,550 in most countries (see table 3.4 for exceptions). Diesel generation and isolated grids This technical solution involves a diesel generator accompanied by a distribution network, generally with no need for transmission or subtransmission lines. Thus, the relevant connection cost is simply the unit distribution and service connection costs. Here, Zomers’s (2001) cost of $1,325 per connection is used. Solar PV The solar PV markets in most Sub-Saharan African countries have only recently begun to be developed. As a result, prices have fallen considerably in most countries over recent years. CORE (2003) reports the results of Afrepren, which arrives at an estimated average price for a 50 Watt-peak (Wp) system of $1,100 for SAPP countries. But work carried out by Econ Pöyry in Tanzania and Zambia, where the markets are undeveloped, reveal prices of $650–$800. Other more recent sources from South Africa, Ethiopia, Sudan, Kenya, and Uganda reveal similar prices.10 Thus, country-specific cost data are applied for those countries where it is available, while a conservative estimate of $800 is applied to the remaining countries (see table 3.4). 10 For South Africa, see http://roo.undp.org/gef/solarpv/forum/fileattachments/Discussion_Solar_PV_Forum_- _Disgest_1_Jan_2004.doc. For Ethiopia, Sudan, Kenya and Uganda, http://roo.undp.org/gef/solarpv/forum/fileattachments/ACFE91.doc. 28 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 3.4 Summary of costs per connection in study countries ($) Urban Rural Grid extension Grid extension or Minigrid (diesel) PV (50 Wp) minihydro General assumption 500 1,550 1,325 800 Exceptions: EAPP Burundi 500 1,000 1,325 670 Djibouti 500 1,550 1,325 750 Egypt 500 1,550 1,325 670 Ethiopia 500 1,550 1,325 750 Kenya 500 1,000 1,325 550 Rwanda 500 1,000 1,325 670 Sudan 500 1,550 1,325 650 Tanzania 500 1,550 1,325 650 Uganda 500 1,000 1,325 730 SAPP South Africa 350 350 1,325 500 Islands Cape Verde 500 2,500 1,325 1,100 Madagascar 500 2,500 1,325 1,100 Mauritius 500 2,500 1,325 1,100 Source: Authors’ original research. 29 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA 4 Least-cost expansion model In estimating the investment needs in each country, we used a least-cost expansion model that simulates optimal electricity generation and investment strategies in response to demand increase in African countries.11 The endogenous decision variables in the model are: • Electricity generation (gigawatt-hours, GWh) • Electricity transmission (GWh) • Refurbishment of existing capacity for electricity generation (MW) • Investments in new capacity for electricity generation (MW) • Investments in new capacity for cross-border electricity transmission (MW) Model assumptions and simplifications Time assumptions The model simulates one year. The year consists of two periods, one peak and one off-peak. The peak period represents all days in the year between the hours of 8 A.M. and 8 P.M. Demand assumptions Demand is modeled by using demand centers, hereafter referred to as nodes. Each country consists of one urban and one rural node. Demand is market driven or social, and is located in one of the two nodes per country. Market demand is located in the urban node, whereas social demand is distributed between the urban and rural nodes. Total demand is the annual total sum of market driven and social demand. Input data for market demand are provided by the econometric model described in chapter 2. Data for suppressed demand are added to the original 2005 market demand figures. Social demand determined by electricity access targets creates additional demand for electricity. This social demand entails extra costs, as new households must be connected to the electricity grid in order to increase the access rate. For the urban nodes, it is assumed that 70 percent of electricity demand occurs during peak periods. The rationale for this is that most of the industry is located in densely populated areas. For rural nodes, there is a 50–50 split in demand between peak and off-peak periods. Power generation assumptions Existing and potential generation capacity can be located in all nodes in the model. 11 AICD intends to make the model available on its website. 30 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Generation capacity is divided into an exogenous set of technology types. Hydropower capacity is split between large-scale hydro (including pumped storage) and minihydro. Thermal capacity comprises power stations or generators based on coal, gas, diesel, HFO, nuclear, and geothermal. In addition, photovoltaic production equipment (solar power) can be built off-grid. Production in each technology is limited by capacity and maximum number of production hours during a year. For thermal technologies, the maximum number of hours per year that a power plant can run is limited by the need for maintenance and other planned outages. For hydropower and solar, however, the number of hours per year will be limited by factors such as water inflow and sunlight. Consequently, the number of available production hours per year will be considerably lower than for the thermal technologies. Large-scale hydropower differs from minihydro in terms of production flexibility and investment potential. It is assumed that all minihydro is run-of-river without the possibility of storing water in reservoirs. This means that minihydro production remains constant day and night. For large-scale hydro, the model allows for flexibility so that it is possible to shift production from the low-demand (off-peak) period to the high-demand (peak) period. This flexibility varies from country to country. Investments in new capacity in the different technologies are determined by the model, but upper and lower limits for investments in each technology are set exogenously. This is particularly important for hydropower generation, as hydropower potential varies significantly from country to country. Similarly, investments in thermal technologies can be limited by availability of fuel (as described in chapter 3). Lower limits are used for thermal power projects that are being developed but are not yet online. Transmission and distribution assumptions Electricity can flow freely between urban nodes within the countries of the model, limited only by the maximum cross-border transmission capacity (measured in MW). This maximum capacity can be increased by building new lines between two urban nodes in two countries. The set of potential transmission connections between countries is set exogenously, as well as the upper bound (in MW) for new transmission investments. Only transmission lines (two highest voltage levels available in a country) can be built between countries. All export to and import from outside the model countries is assumed to be exogenous. Urban nodes are connected to the national grid and can export or import electricity, depending on transmission capacity. Rural nodes “import” a predetermined share (70 percent) of electricity from the urban node of the same country. The rest (30 percent) of electricity demand in rural nodes is met by off- grid technologies (minihydro, diesel and solar PV). The rural nodes cannot import from or export electricity to other countries, nor is it possible to build additional transmission lines between the urban and rural nodes in the model. Investments in the domestic distribution and subtransmission grid are not explicitly modeled as decision variables (in contrast to investments in transmission grid between countries). Instead, the distribution and subtransmission grid within a country is assumed to grow with increasing market and social demand according to a predefined rate. 31 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Transmission losses are assumed to occur when electricity flows from one node to another. This loss depends on the nodes and the direction of the flows. Distribution losses (in the low-voltage grid) are set as a percentage of net demand for each node. The mathematical formulation of the model The approach used is a linear programming model, in which all the decision variables are modeled as continuous positive variables. Objective function and model constraints The model minimizes the cost of producing and distributing electricity to meet a certain demand level. The costs include production as well as investment and refurbishment costs for generation; investment costs of new cross-border transmission lines; investment, refurbishment, and operation costs of domestic distribution lines; and costs of new connections to the power grid. Expression (1) is the objective function, which is to be minimized: (1) It consists of the following: • Variable costs of electricity generation • Costs of investing in new generation capacity and refurbishing the existing capacity • Costs of electricity transmission between endogenous nodes • Costs of investing in transmission grids between countries for all voltage levels • Costs of connecting new households to the low-voltage distribution grid (this includes the distribution and subtransmission cost of social demand) • Cost of expanding the distribution and subtransmission grid to accommodate increased marked demand • Costs of refurbishing the distribution and subtransmission grid The objective function is subject to several constraints that will be explained below (refer to the list of endogenous and exogenous variables below). Expression (2) is the power balance constraint. For each time period and node, the sum of electricity imports from other endogenous nodes, imports from exogenous countries, and internal electricity generation (left-hand side of the constraint) must be large enough to cover market demand, social demand, exports to other endogenous nodes, and exports to exogenous countries (right-hand side of the constraint). 32 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA (2) Expression (3) reveals that the electricity flow on a given line between two endogenous nodes (in either direction) must not exceed the transmission capacity (in MW) multiplied by the number of hours during time period t, . (3) Expression (4) states that investments in new generation capacity of technology k in node i must not exceed the potential maximum capacity (in MW). Expression (5) states that investments in new generation capacity of technology k in node i must be at least as high as a certain minimum capacity (MW). (4) (5) Refurbishment of existing generation capacity of technology k in node i must not exceed a certain maximum capacity (MW, expression 6). (6) Annual electricity generation (in GWh) by technology k in node i must be lower than the upper bound for generation (right side of expression 7), and higher than an exogenously set minimum generation level (right side of expression 8). (7) (8) Investments in new high-voltage grid capacity (in MW) of voltage type z between two countries cannot exceed the maximum potential for high-voltage grid capacity (expression 9). (9) Sets U represents all possible transmission connections between endogenous nodes; I is the set of endogenous nodes; K is the set of all electricity generation technologies; Y is the set of exogenous countries; T is the set of time periods. Endogenous (decision) variables is the generation (GWh) in endogenous node i using technology k during time period t. represents investment (MW) in generation technology k in endogenous node i. is refurbishment (MW) of generation capacity of technology k in endogenous node i. 33 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA represents the electricity transmission flow (GWh) from endogenous node i to endogenous node j during time period t. stands for investment (MW) in transmission capacity of voltage z between endogenous nodes i and j. All endogenous decision variables are assumed to be positive and continuous variables. Exogenous variables is the variable cost of producing one energy unit (GWh) with generation technology k in node i. is the annualized fixed cost per MW of investing in generation technology k in node i. is the annualized fixed cost per MW of refurbishing existing generation technology k in node i. is the variable transmission cost of sending one energy unit (GWh) from endogenous node i to endogenous node j. is the annualized fixed cost per MW of investing in transmission lines of voltage type z between endogenous urban nodes i and j. is the cost of expanding the domestic electricity grid in the model. This cost is assumed to be a function of market demand and social demand (new connections). is market demand (GWh) in node i in period t. is social demand (GWh) in node i in period t. Social demand equals the number of new households connected to the grid in node i with the expected annual consumption per household. The number of new connections varies between the scenarios. is the fixed annualized cost of refurbishing the internal electricity network in node i. is the amount of power imported (GWh) from exogenous country y to endogenous node i in time period t. The level of imports is set exogenously. is the amount of power exported (GWh) from endogenous node i to exogenous country y in time period t. The level of exports is set exogenously. is the original transmission capacity between endogenous nodes i and j. is the number of hours during time period t. is the upper bound for investments in new generation capacity of technology k in endogenous node i. is the lower bound for investments in new generation capacity of technology k in endogenous node i. 34 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA is the upper bound for refurbishment of existing generation capacity of technology k in endogenous node i. is the existing generation capacity of technology k in node i which does not need refurbishment to operate. represents the average numbers of full load hours for generation technology k in node i. is the minimum production level of technology k in node i and time period t (as a percentage of the maximum production level (GWh) in that period). is the maximum potential transmission capacity (MW) between nodes i and j for lines with voltage type z. The maximum potential capacity is exogenous but a function of the voltage or size. For instance, the upper bound for a 400 kV line is 1,440 MW (which is an average line rating for 400 kV lines). 35 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA 5 Results: Spending needs through 2015 In our case scenarios, we have analyzed the costs of achieving three levels of access to electricity 10 years from now. The cases are: • Keeping the access rates constant at 2005 level (2005 access rate scenario) • Increasing the access rates by 1 to 2 percentage points in each country to reach a regional average (regional target access rate scenario) • Increasing the access rates to the national targets for access rates of individual countries (national access targets scenario) A scenario with lower growth would help explain how lower GDP growth might affect outcomes. The low-growth scenario is discussed in the next chapter, but data are included in the tables presented here. The next chapter also presents several sensitivities that test how vulnerable the results are to assumptions about various parameter values. We wish to analyze the impacts of a ten-year expansion program for electricity. The starting point for all scenarios is 2005; therefore, the costs discussed in this chapter refer to the costs of expanding the power system from 2005 to 2015. The terms used are defined below. The overall cost of developing the power system appears high, but not unattainable given the GDP of each of the regional trading areas. But both GDP and power investment requirements are very unevenly distributed within the regional pools. As a result, under certain scenarios, some countries face power spending requirements that are very burdensome relative to the size of their economies (figure 5.1). In SAPP, depending on the electrification target and other variables, spending requirements may exceed 6 percent of GDP in the Democratic Republic of Congo, Mozambique, and Zimbabwe. In EAPP, countries such as Egypt, Burundi, and Ethiopia may reach similar levels. About half of the countries in WAPP have investment requirements of almost 10 percent of GDP, and Guinea and Liberia stand out at almost 30 percent. In CAPP, only the Republic of Congo requires investments of more than 5 percent of GDP. Some of these countries have the potential to become major exporters of power, provided they receive cross-border injections of capital to develop their power infrastructure. The necessary capital is not likely to materialize unless trade in power expands. The next sections survey the investment requirements and costs in the four African regions (SAPP, EAPP, WAPP, and CAPP). More detailed output tables for each country (including the island states) can be found in annex 1 (separately bound). The section on each region follows the same basic outline, which we present below. Included in the outline are definitions of key terms used throughout the chapter. 36 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.1 Overall power spending needs to reach national targets for electricity access under alternative trade scenarios by country % of GDP in 2015 (a) Southern Africa Power Pool (b) East Africa Power Pool (c) Western Africa Power Pool (d) Central Africa Power Pool Maintaining electricity access at 2005 rates We begin the discussion of each region by looking at the investment needs and costs of the least- ambitious target scenario—where access rates are kept at the 2005 level. The types of costs are as follows: Investment needs This includes both investment in new capacity and refurbishment of existing capacity. Refurbishment consists of (a) power plants that were operational in 2005 but will exceed their economic lifetime before 2015 and will need to be refurbished to stay operational and (b) power plants that were not operational in 2005 because they were in such a bad state. This estimate is based on information about facility ages and conditions that we have assembled for this project. Overnight investment costs Generation costs: The investment costs of expanding the generation system between 2005 and 2015. 37 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Transmission, distribution, and connection costs: The additional costs needed to bring power from power plants to consumers—the costs of transmission, distribution and new connections. Total overnight investment costs: The total cost of expanding the power system, including both new investment and refurbishment costs. This is the total investment needed to meet market demand and provide electricity in 2015 to the same share of households as in 2005. It does not include variable costs. Annualized costs Annualized investment (capital) cost: The capital amount that must be spent annually to reach the access target. The annualized capital cost does not, however, equal the total overnight investment cost divided by 10 years. The cost takes into account both the discount rate and the different economic lifetimes of the investments. For example, we assume that the economic lifetime of hydropower plants is 40 years, compared to 30 years for coal-fired power plants and 25 years for natural gas fired power plants. The formula used for annualizing investment cost is the following: annualized capital cost = OIC * r(1–(1+r)–T) where OIC is the overnight investment cost, r is the discount rate, and T is the economic lifetime of the investment. Annual variable costs: Includes fuel costs and variable costs of operation and maintenance of the system. The system includes both existing capacity as of 2005 that is still operating in 2015 and new capacity added over the ten year period. Total annualized cost of system expansion: Includes the annualized investment cost and the annual variable costs of operating the new capacity. The variable costs of operating the existing capacity are not included in this figure. Total annualized cost of system expansion and operation: Includes the total annualized cost of system expansion and the annual variable costs of the existing system. Cost effects of raising access rates This section discusses the increased costs needed to meet the more ambitious regional and national targets for access rates for each region. The effects of trade In this section we compare the effects of trade expansion and trade stagnation in each region. In the scenarios described above, we assume that trade will expand wherever it is worth the cost—that is, wherever the benefits of trade outweigh the costs of the additional infrastructure needed to support expanded trade. This is the trade-expansion scenario. Optimal trade expansion requires close regional cooperation in financing and institutional arrangements. Well-functioning power pools, mutual confidence, and stability are critical if trade is to yield its economic benefits. Exporters of power must be confident that they will have access to other countries’ markets. Without that confidence, they will not take the risks of investing in exportable capacity. Importers, on the other hand, must be sure that they will 38 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA receive the power they need. In the absence of such assurance, policy makers will rely on self-sufficiency, regardless of its economic and environmental costs. Therefore, achieving optimal power trade will require important political, legal, and economic commitments from the cooperating countries. In our trade-stagnation scenario, on the other hand, we assume that no new transmission lines will be built between countries. The existing cross-border lines can, of course, be used. In all scenarios, trade flows and volumes are determined endogenously, and the trade flows along the existing lines can change (increasing if there is spare capacity or even changing direction). We only consider trade stagnation in the context of national targets for access rates, but the impact of trade is more or less independent of the target access rate. We also discuss the costs and gains of trade expansion in each region. Trade allows for the development of the cheapest power resources in the region, taking into account the costs of investing in generation and necessary transmission lines. Trade-stagnation therefore leads to economic loss. While hydropower infrastructure is expensive to build (due to higher capital costs), it is cheap to operate once built. Thermal power plants (coal, gas, and HFO-fueled power plants), on the other hand, have higher variable costs from fuel. Additionally, hydropower has a longer economic lifetime than thermal power plants. Therefore, replacing thermal power with hydropower requires greater up-front investments but lower operating costs. This replacement takes place under trade expansion. Returns on investing in trade. The gains from trade can be regarded as returns on investment, which we estimate for each region. Long-run marginal cost. Differences in long-run marginal cost can also illustrate the gains from trade. The long-run marginal cost reflects the shadow cost of a marginal increase in net demand (consumption) in 2015,12 or in other words, the least-cost option for meeting increased demand: increased generation in existing plants (domestic or through imports), or generation capacity expansion, together with the necessary investments in distribution and cross-border transmission lines. The costs considered when determining the long-run marginal cost are marginal variable costs (fuel and O&M costs) of either existing or new capacity in 2015, marginal investment and refurbishment costs of generation, marginal cost of imports (through transmission loss), and marginal costs of investments in domestic T&D and cross-border transmission lines. Since the marginal costs of social demand (new connections) are driven by nonmarket considerations, they tend not to equalize with trade. Therefore, they are not considered in the long-run marginal cost calculation. The model seems to indicate that trade benefits the importing countries. But the financial flows and agreements (on prices, for example) that are necessary to implement the least-cost solution are outside the scope of the model. It is these flows and agreements that will determine the allocation of the benefit. The costs of T&D investment and refurbishment reflect the costs of a domestic grid and are not passed on to the importing countries. These costs are therefore the same under trade stagnation and trade expansion. 12 Mathematically they are multipliers of the “supply equals demand” equations of the linear programming model. 39 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA SAPP Maintaining electricity access at 2005 rates Investment needs The SAPP region needs almost 31,300 MW of new capacity by 2015 to keep up with the demand growth resulting from general economic development and population growth. In addition, 28,000 MW of existing capacity must be refurbished. Table 5.1 provides an overview of generation capacity in 2015, refurbishment and investment volumes over the ten years, and the resulting capacity mix for all scenarios. Table 5.1 Investment needs and generation mix in SAPP in 2015 Trade- Low-growth Trade-expansion scenario stagnation scenario scenario National targets Regional target National targets National targets for access 2005 access rate access rate for access rates for access rates rates, trade expansion Generation capacity (in MW) - Installed 17,136 17,136 17,136 17,136 17,136 - Refurbishment 28,029 28,035 28,046 28,148 28,046 - New investments 31,297 32,168 33,319 32,013 20,729 Generation capacity mix (%) - Hydro 33 33 34 25 40 - Coal 60 60 59 66 52 - Gas 0 0 0 2 0 - Other 7 7 7 7 8 Source: Authors’ original research. Note: Installed capacity refers to capacity in place in 2005 that is not refurbished before 2015. Existing capacity that is refurbished before 2015 is not included in the installed capacity figure, but in the refurbishment figure. South Africa accounts for about 80 percent of electricity demand in SAPP. Therefore, development in South Africa is an important driving force for the rest of the region. Figure 5.2 shows demand and production in the different countries in SAPP as a share of the region’s total. 40 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.2 Demand and production in SAPP in 2015 Current access scenario; percentage of regional total In the 2005 access rate scenario, power generation investments in South Africa amount to 18,700 MW (60 percent of total); in addition, 21,700 MW of capacity is refurbished in South Africa. The largest share of investments in South Africa is in coal-fired power plants; the rest include 3,000 MW in open- cycle gas turbine generators (that are already committed and thus included exogenously in the model) and 2,000 MW in hydropower and pumped storage. In addition to domestic investments in South Africa, large investments take place in countries that are rich in hydropower: the Democratic Republic of Congo (7,200 MW), Mozambique (3,200 MW), and Zimbabwe (2,200 MW). While development in Zimbabwe meets domestic demand (the country imported 14 percent of its electricity as of 2005), exports from the Democratic Republic of Congo and Mozambique supply power to the rest of the region, with net exports of 50 TWh and 6 TWh, respectively. Overnight investment costs Generation costs are almost $38 billion (table 5.2). The largest share of this total is investments in new capacity ($30.3 billion), while the cost of refurbishment is only $7.5 billion. Even though refurbishment contributes about the same amount (in MW) to the total capacity as new capacity, refurbishment of existing capacity is much cheaper. Coal power plants in South Africa are an exception here; the cost of refurbishing them is almost as expensive as investing in new plants. Transmission, distribution, and connection costs are lower than the costs of building new power plants: $26 billion is needed for investments in the expanded grid and refurbishment of the existing grid (table 5.2). Investments in new transmission and distribution lines account for more than half of this sum ($16 billion). 41 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The direct cost of connecting new customers to the grid is only a fraction of the total grid cost: $0.7 billion is needed to maintain the access rate at the 2005 level (aligned with population growth) (table 5.2). More than 90 percent of this would be spent in urban areas. Total overnight investment costs are slightly less than $64 billion in the 2005 access rate scenario. Table 5.2 Overnight investment costs in SAPP, 2005–15 $ millions Trade- Low-growth Trade-expansion scenario stagnation scenario scenario National targets 2005 access Regional target National targets National targets for access rates, rate access rate for access rates for access rates trade expansion Generation Investment cost 30,277 31,103 32,242 34,644 18,589 Refurbishment cost 7,572 7,574 7,577 7,587 7,577 T&D and connection Investment cost 16,384 19,422 23,711 20,653 16,606 - Cross-border transmission lines 3,009 2,991 3,058 0 3,082 - Distribution grid 12,674 12,674 12,674 12,674 5,544 - Connection cost (urban) 643 2,210 3,995 3,995 3,995 - Connection cost (rural) 58 1,547 3,985 3,985 3,985 Refurbishment cost 9,775 9,775 9,775 9,775 9,775 Total 64,008 67,874 73,304 72,659 52,546 Source: Authors’ original research. Annualized costs The annualized capital cost of meeting market demand and maintaining current access is almost $9 billion: $3.2 billion in transmission, distribution, and connection; and $5.6 billion in generation (table 5.3). In addition, the annual variable costs of operating the system amount to $8.3 billion. About $2 billion of this total is related to operating new power plants, while the rest is related to operating existing and refurbished power plants ($3.2 billion) and the grid ($3.1 billion). The total annualized cost of system expansion is equal to 2.2 percent of GDP of the region in 2015 in the current access rate scenario (table 5.4). Adding the variable operation costs of existing capacity, the annualized costs of system expansion and operation are equal to 3.4 percent of GDP. There are, however, large differences between costs in different countries. First, the costs of generation capacity expansion are high in countries with large hydropower development: 5.8 percent of GDP in the Democratic Republic of Congo, 6.2 percent in Mozambique, and 8.5 percent in Zimbabwe. Second, grid-related costs (investments, refurbishment, and operation) are significant in countries such as Zimbabwe, Zambia, Namibia, and the Democratic Republic of Congo. Third, even though generation- capacity expansion requires only 0.7 percent of GDP in 2015 in South Africa, the variable costs of the new coal-fired power plants amount to 0.6 percent of GDP. 42 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Detailed country tables can be found in annex 1 (separately bound). Table 5.3 Annualized costs of system expansion in SAPP $ millions Trade-stagnation Low-growth Trade-expansion scenario scenario scenario Regional National targets 2005 access National targets National targets target access for access rates, rate for access rates for access rates rate trade expansion Generation Investment cost 4,267 4,386 4,544 4,899 2,536 Refurbishment cost 1,340 1,340 1,341 1,342 1,341 Variable cost (fuel, O&M) 5,184 5,230 5,259 6,341 3,978 - New capacity 1,986 2,049 2,084 2,953 257 - Existing capacity 3,199 3,181 3,175 3,389 3,722 T&D and connection Investment cost 2,034 2,411 2,944 2,564 2,061 - Cross-border 374 371 380 0 383 - Distribution grid 1,573 1,573 1,573 1,573 688 - Urban connection 80 274 496 496 496 - Rural connection 7 192 495 495 495 Refurbishment cost 1,213 1,213 1,213 1,213 1,213 Variable cost (existing capacity) 3,100 3,100 3,100 3,100 3,100 Total Capital cost 8,855 9,351 10,042 10,018 7,152 - Investment cost 6,301 6,797 7,488 7,463 4,598 - Refurbishment cost 2,553 2,554 2,554 2,556 2,554 Variable cost 8,284 8,330 8,359 9,442 7,078 Total 17,139 17,681 18,401 19,460 14,230 Source: Authors’ original research. Note: We assume that there are no variable costs related to the new T&D assets. 43 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.4 Annualized costs as a share of GDP in SAPP in 2015 Percent Trade- Low-growth Trade-expansion scenario stagnation scenario scenario National National Regional National targets for Current targets for target targets for access rates, access rate access access rate access rates trade rates expansion Generation Investment cost 0.9 0.9 0.9 1.0 0.6 Refurbishment cost 0.3 0.3 0.3 0.3 0.3 Variable cost (fuel, O&M) 1.0 1.1 1.1 1.3 1.0 - New capacity 0.4 0.4 0.4 0.6 0.1 - Existing capacity 0.6 0.6 0.6 0.7 0.9 T&D and connection Investment cost 0.4 0.5 0.6 0.5 0.5 - Cross-border 0.1 0.1 0.1 0.0 0.1 - Distribution grid 0.3 0.3 0.3 0.3 0.2 - Urban connection 0.0 0.1 0.1 0.1 0.1 - Rural connection 0.0 0.0 0.1 0.1 0.1 Refurbishment cost 0.2 0.2 0.2 0.2 0.3 Variable cost (existing capacity) 0.6 0.6 0.6 0.6 0.7 Total Capital cost 1.8 1.9 2.0 2.0 1.7 - Investment cost 1.3 1.4 1.5 1.5 1.1 - Refurbishment cost 0.5 0.5 0.5 0.5 0.6 Variable cost 1.7 1.7 1.7 1.9 1.7 Total 3.4 3.6 3.7 3.9 3.4 Source: Authors’ original research. Note: We assume that there are no variable costs related to the new T&D assets. Cost effects of raising access rates Regional targets for access rates: Electricity access of 35 percent on average Compared with the 2005 access rate, meeting the regional target for electricity access—35 percent access on average— requires an additional investment of almost $3.9 billion. This figure corresponds to about $0.5 billion in annualized capital costs (see tables 5.1–5.4). The largest contributors to this increase are the costs of transmission, distribution and connection: connecting new households to the grid involves extra costs of about $3 billion ($380 million in annualized costs). About 40 percent of this amount is spent in rural areas, compared with only 10 percent in the current access scenario. 44 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Some additional power generation capacity is also needed to meet the increased demand: investment costs are $0.8 billion higher ($120 million in annualized costs) than in the current access rate scenario. The additional costs of operating the new capacity (variable costs) are much less—$50 million annually. The annualized costs of system expansion correspond to 2.3 percent of GDP in 2015 for the region. Including variable costs of existing capacity increases the total annualized cost of system expansion and operation to 3.6 percent of GDP. National targets for electricity access Meeting national targets requires $9.3 billion more than keeping the access rate constant at 2005 rates. This corresponds to almost $1.3 billion in annualized costs (see tables 5.1–5.4). The largest contributors to this increase are the costs of transmission, distribution and connection. For example, connecting new households to the grid involves an extra cost of about $7.3 billion ($0.9 billion in annualized costs). About half of this is spent in rural areas, compared with only 10 percent in the 2005 access rate scenario. The national targets scenario also requires additional power generation capacity to meet increased demand: investment costs are $2 billion higher ($280 million in annualized costs) than in the 2005 access rate scenario. The additional costs of operating the new capacity (variable costs) are much less—$75 million annually. The annualized costs of system expansion are 2.4 percent of GDP in 2015 for the region. Including variable costs of current capacity increases the total annualized costs of system expansion and operation to 3.7 percent of GDP. The effect of trade Detailed figures are reported in tables 5.1–5.4. Thermal power substitutes for hydropower The most visible effect of restricting investments in cross-border transmission lines is reduced hydropower development, especially in the Democratic Republic of Congo (7,640 MW of forgone investments). The trade-expansion scenario leads to a generation mix with a higher proportion of renewable energy and less coal: only 25 percent of total capacity is supplied by hydropower under trade stagnation, compared to 34 percent in the trade-expansion scenario (figure 5.3). 45 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.3 Capacity mix in SAPP in 2015 with different trade assumptions Power production in individual countries (as a share of total production in the area) would change with different assumptions about trade. For example, if trade were to expand, the Democratic Republic of Congo would produce and export more power (figure 5.4). With trade stagnation, the Democratic Republic of Congo’s exports are reduced from 41 TWh/year to 2.3 TWh/year. Reduced import possibilities in South Africa and Botswana are replaced by more coal- fired power plants. In Botswana the expansion of the Morupule plant becomes viable. (We assume the Mmamabula project will be built in both scenarios because of the short distance to the grid in South Africa). Gas-fired power plants and hydropower are built in Angola and Namibia to meet domestic demand in the trade-stagnation scenario. Zambia imports electricity in the trade-expansion scenario and does not develop the domestic hydropower resources that are developed in the trade-stagnation scenario. This is because it is cheaper to develop hydropower resources in neighboring countries, and these resources (particularly in the Democratic Republic of Congo) are developed first in the trade-expansion scenario. 46 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.4 Production in SAPP in 2015 with different trade assumptions Percentage of regional total Reduced hydropower development in Mozambique and Zimbabwe is partly offset by increased coal- power investments in these countries. This is because we have exogenously constrained the potential for hydropower development in these countries in the trade-stagnation scenario; it has been claimed that these investments are not viable without simultaneous development of regional transmission possibilities. According to unconstrained model results, however, at least some of these hydropower investments are necessary to meet the domestic demand. Despite the results of the unconstrained model simulations, it is possible that the exogenously constrained result—substituting hydropower with coal-fired power plants to meet domestic demand—is a reasonable outcome.13 For instance, hydropower resources are located in the north of Mozambique, while domestic demand is located in the south. Hence, large investments in the domestic grid would be necessary to provide the south with the hydropower from the north. Since we do not model the domestic grid in detail, the costs of this grid development would have been underestimated in the model. Coal-fired power plants, on the other hand, can easily be built in the vicinity of demand centers. Therefore, the results illustrate the trade-off between investments in grid and generation, and between different kinds of generation capacity. Under trade stagnation, power flows from the north to the south are replaced by more power exchange—that is, power flows in different directions in peak and off-peak hours. For instance, there is export during off-peak hours from South Africa to Mozambique, Namibia, and Zimbabwe. 13 On the other hand, it is also possible that the assumed growth rates (that form the basis for electricity demand after ten years) are overestimated (especially for Zimbabwe). 47 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The impact of trade stagnation on the individual Table 5.5 Net imports (+)/exports (-) as share of countries’ power balance is illustrated in figure 5.5 domestic demand and table 5.5. One of the most striking effects, Percent described above, involves the Democratic Republic of Trade- Trade- Congo, where net exports increase almost 20 times expansion stagnation scenario scenario with trade expansion as other countries (Botswana, Angola 65 0 Malawi, Zambia) become net importers instead of net Botswana 93 -117 exporters. Under the trade-expansion scenario, only Congo, Dem. Rep. -369 -16 the Democratic Republic of Congo and Mozambique Lesotho 68 68 remain power exporters. Malawi 58 -60 Trade from the north (especially from the Mozambique -33 -12 Democratic Republic of Congo and Mozambique) to Namibia 72 -17 the south flows along several corridors (figure 5.5). South Africa 10 4 The transit countries (Zambia, Zimbabwe, Botswana, Zambia 18 -62 Angola, and Namibia) are also net importers. This Zimbabwe 17 18 pronounced trade pattern requires sufficient Source: Authors’ original research. investments in transmission lines throughout the region. Figure 5.5 Trade flows with trade expansion (left) and trade stagnation (right) in SAPP in 2015, TWh Source: Authors’ original research. Costs and gains Overnight investment costs. Total overnight capital costs are $645 million higher in the trade- expansion scenario than in the trade-stagnation scenario. The costs related to transmission-grid expansion are about $3 billion higher with trade expansion. Yet, the costs related to investments in generation 48 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA capacity are $2.5 billion lower in the trade-expansion scenario despite the investment level being 1,300 MW higher. In other words, expensive capacity replaces cheap capacity when trade is restricted. Annualized costs. The total annualized costs of system expansion in the SAPP region are more than $1 billion lower in the trade-expansion scenario than in the trade-stagnation scenario. Annualized investments costs are $25 million higher in the trade-expansion scenario. This reflects offsetting contributions from generation and transmission. Considerable cost savings stem from the operation of the integrated system: variable costs (that is, fuel costs and variable O&M costs) are significantly lower in the trade-expansion scenario. These cost savings stem both from using existing generation capacity more efficiently in the integrated system ($214 million), and the different capacity composition of the trade- expansion scenario ($868 million). Returns on investing in trade. Recall that the capital costs in the trade-expansion scenario are $645 million higher than in the trade-stagnation scenario. On the other hand, the annual variable costs of the enlarged system in 2015 are more than $1 billion lower. This implies that the extra expenditure is recouped in a little more than half a year. This is equivalent to a return on investment of 168 percent. The savings from trade expansion are 8.2 percent of the annualized cost of expansion.14 Long-run marginal cost. Table 5.6 reports the long-run marginal cost of power in the SAPP countries under trade expansion and trade stagnation. The cost figures illustrate how the cost of power increases from production to consumption centers (see figure 5.4): The cost of power generation (column 2) is the lowest in the Democratic Republic of Congo; as the power is transmitted to the south, the need for international transmission lines creates additional costs for each country that is passed. Comparing this with the costs of generation in the trade-stagnation scenario (column 5) reveals that expanding domestic power generation capacity is much more expensive in countries that import during trade. In the case of the Democratic Republic of Congo, the difference in costs reflects the cost of international transmission lines that are necessary to export the power. The model seems to indicate that trade benefits the importing countries. But the financial flows and agreements (on prices, for example) that are necessary to implement the least-cost solution are outside the scope of the model. It is these flows and agreements that will determine the allocation of the benefit. The costs of T&D investment and refurbishment reflect the costs of a domestic grid and are not passed on to the importing countries. These costs are therefore the same in both scenarios. 14 Annual savings are $1,059 million. The cost of operation is $19,460 million, but that includes variable costs of current capacity ($6,489); $1,059 is 8.2 percent of the difference, that is, 8.2 percent of $12,971 million. Note that the discounted value of $1,059 into infinity at 12 percent discount is $8,825 million. 49 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.6 Long-run marginal costs in SAPP $ per MWh Trade-expansion scenario Trade-stagnation scenario Cost of Cost of Cost of Cost of generation and generation and domestic Total long-run domestic Total long-run international international transmission marginal costs transmission marginal costs transmission transmission and distribution and distribution lines lines Angola 18 46 64 59 46 105 Botswana 35 21 56 40 21 61 Congo, Dem. 14 24 39 11 24 36 Rep. Lesotho 36 22 58 48 22 70 Malawi 33 19 51 35 19 54 Mozambique 33 8 41 47 8 55 Namibia 36 73 109 47 73 120 South Africa 36 20 55 47 20 67 Zambia 29 46 75 33 46 78 Zimbabwe 32 46 78 39 46 85 Average 33 27 60 46 27 73 Note: Average is weighted by annualized cost. In some cases power exporting countries report higher LRMC under trade expansion. Even if the cost of meeting domestic power consumption may be higher with trade than without; the higher revenues earned from exports would more than compensate for that increment. EAPP Maintaining electricity access at the 2005 rate Investment needs About 23,000 MW of new capacity are needed in the EAPP region in 2015 to meet the demand growth resulting from general economic development and population growth. In addition, more than 1,000 MW of existing capacity must be refurbished. We may have underestimated the capacity potential for EAPP, which is much lower than that of SAPP. See appendix 2 for details. Table 5.7 provides an overview of generation capacity, refurbishment, and investment costs in 2015, as well as the resulting capacity mix in all scenarios. 50 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.7 Investment needs and generation mix in EAPP in 2015 Trade- Low-growth Trade-expansion scenario stagnation scenario scenario National targets 2005 Regional target National targets National targets for access access access rate for access rates for access rates rates, trade rate expansion Generation capacity (MW) - Installed 22,132 22,132 22,132 22,132 22,132 - Refurbishment 1,369 1,375 1,375 1,381 1,375 - New investments 23,045 24,639 25,637 17,972 23,540 Generation capacity mix (%) - Hydro 49 47 48 28 48 - Coal 2 2 2 3 2 - Gas 47 48 49 64 45 - Other 2 3 4 5 4 Source: Authors’ original research. Note: Installed capacity refers to capacity in place in 2005 that is not refurbished before 2015. Existing capacity that is refurbished before 2015 is not included in the installed capacity figure, but in the refurbishment figure. Egypt accounts for about 80 percent of total demand in EAPP with the 2005 access targets (figure 5.6). Since Egypt imports about 40 percent of its electricity (55 TWh), development in Egypt is of considerable importance for the EAPP region. Egypt requires investment of almost 7,000 MW in natural-gas-fired power plants before 2015 to meet growing domestic demand. In addition, countries that have hydropower resources make substantial investments in hydropower: 8,150 MW in Ethiopia, 3,700 MW in Sudan, 1,200 MW in Tanzania and Uganda each, and 300 MW in Rwanda. In addition, Kenya and Tanzania invest in coal-fired power plants. Both Ethiopia and Sudan are large net exporters. 51 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.6 Demand and production in EAPP countries in 2015 Current access scenario; percentage of regional total Overnight investment costs Generation costs. Generation capacity must be more than doubled from its 2005 level to keep up with electricity demand growth. The costs of expanding the generation system over ten years are more than $29 billion (table 5.8). Almost all of this is due to investments in new capacity; the cost of refurbishment is negligible. Transmission, distribution, and connection costs are about one-quarter of total costs ($11 billion). Investments in the grid amount to $7.5 billion. Investments in transmission and distribution lines account for most of this total. Costs directly related to connecting new customers amount to about 40 percent of the total grid costs; $3 billion is needed to maintain the access rate at the 2005 level—that is, merely keeping in line with the population growth. About 80 percent of the connection cost is spent in urban areas. Refurbishment of the existing grid requires $3.3 billion. Total overnight investment costs. The total overnight investment costs are $40.2 billion in the 2005 access rate scenario. 52 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.8 Overnight investment costs in EAPP, 2005–15 $ millions Trade-stagnation Low-growth Trade-expansion scenario scenario scenario National targets Regional target National targets National targets 2005 access rate for access rates, access rate for access rates for access rates trade expansion Generation Investment cost 28,913 30,802 32,667 18,621 31,275 Refurbishment cost 396 398 398 399 398 T&D and connection Investment cost 7,549 16,430 27,385 26,372 26,301 - Cross-border transmission lines 1,320 937 1,013 0 964 - Distribution grid 3,072 3,072 3,072 3,072 2,037 - Connection cost (urban) 2,484 5,263 5,702 5,702 5,702 - Connection cost (rural) 674 7,159 17,599 17,599 17,599 Refurbishment cost 3,342 3,342 3,342 3,342 3,342 Total 40,200 50,973 63,793 48,735 61,317 Annualized costs The annualized capital cost of keeping up with GDP and population growth for ten years is about $5.3 billion. The share of households with access to electricity would remain the same as in 2005. The annualized cost of transmission, distribution, and connection is $1.3 billion and the annualized cost of generation capacity is $4 billion (table 5.9). In addition, the annual variable costs of operating the system amount to $5.8 billion. Most of this ($4.4 billion) is related to operating new power plants (fuel costs and variable operation and maintenance costs), while the rest is related to existing and refurbished power plants ($0.7 billion) and the grid ($0.8 billion). The total annualized cost of system expansion is equal to 3.6 percent of GDP of the region in 2015 in the 2005 access rate scenario (table 5.10). Adding the variable costs of system operation, the total annualized costs of system expansion and operation are 4.2 percent of GDP. The costs of system expansion are relatively high in Egypt, the largest country in the region: 3.8 percent of GDP. The capital costs are only 0.9 percent, but since the new capacity is gas-fired, the fuel costs amount to as much as 3 percent of GDP. Ethiopia is the country with the highest outlay in terms of share of GDP—9.2 percent of GDP in total. But two-thirds of these costs are related to investments in generation capacity used for exports. In addition, some costs are related to investments in transmission and distribution lines and variable costs. The costs correspond to 1–2 percent of GDP in Burundi and Djibouti. In other countries, the costs are in the range of 2.5–3.5 percent of GDP. Detailed country tables can be found in annex 1 (separately bound). 53 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.9 Annualized costs of system expansion in EAPP, 2015 $ millions Trade- Low-growth Trade-expansion scenario stagnation scenario scenario National Regional National National targets for 2005 access target access targets for targets for access rates, rate rate access rates access rates trade expansion Generation Investment cost 3,856 4,123 4,379 2,583 4,202 Refurbishment cost 70 70 70 70 70 Variable cost (fuel, O&M) 5,082 5,778 6,045 8,965 4,858 - New capacity 4,390 5,112 5,378 6,912 4,179 - Existing capacity 693 666 667 2,053 679 T&D and connection Investment cost 933 2,010 3,334 3,208 3,199 - Cross-border 164 116 126 0 120 - Distribution grid 381 381 381 381 253 - Urban connection 308 653 708 708 708 - Rural connection 79 859 2,119 2,119 2,119 Refurbishment cost 415 415 415 415 415 Variable cost (existing capacity) 762 762 762 762 762 Total Capital cost 5,273 6,618 8,198 6,277 7,862 - Investment cost 4,789 6,133 7,712 5,791 7,377 - Refurbishment cost 485 485 485 485 485 Variable cost 5,844 6,540 6,807 9,727 5,620 Total 11,118 13,158 15,004 16,003 13,506 Source: Authors’ original research. Note: We assume no variable costs related to the new T&D assets. 54 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.10 Annualized costs as a share of GDP in EAPP, 2015 Percent Trade- Low-growth Trade-expansion scenario stagnation scenario scenario National Regional National National targets for 2005 access target access targets for targets for access rates, rate rate access rates access rates trade expansion Generation Investment cost 1.5 1.6 1.7 1.0 1.8 Refurbishment cost 0.0 0.0 0.0 0.0 0.0 Variable cost (fuel, O&M) 1.9 2.2 2.3 3.4 2.0 - New capacity 1.7 1.9 2.0 2.6 1.8 - Existing capacity 0.3 0.3 0.3 0.8 0.3 T&D and connection Investment cost 0.4 0.8 1.3 1.2 1.3 - Cross-border 0.1 0.0 0.0 0.0 0.1 - Distribution grid 0.1 0.1 0.1 0.1 0.1 - Urban connection 0.1 0.2 0.3 0.3 0.3 - Rural connection 0.0 0.3 0.8 0.8 0.9 Refurbishment cost 0.2 0.2 0.2 0.2 0.2 Variable cost (existing capacity) 0.3 0.3 0.3 0.3 0.4 Total Capital cost 2.0 2.5 3.1 2.4 3.6 - Investment cost 1.8 2.3 2.9 2.2 3.1 - Refurbishment cost 0.2 0.2 0.2 0.2 0.2 Variable cost 2.2 2.5 2.6 3.7 2.4 Total 4.2 5.0 5.7 6.0 5.6 Source: Authors’ original research. Note: We assume no variable costs related to the new T&D assets. Cost effects of raising access rates Regional target for access rate: electricity access of 35 percent on average Meeting the international target for electricity access (35 percent on average) requires an additional investment of almost $11 billion compared to maintaining the 2005 access rate. This figure corresponds to about $1.3 billion in annualized capital costs (see tables 5.7–5.10). The largest contributors to the increased costs are the costs of transmission, distribution, and connection: connecting new households to the grid involves extra costs of about $9 billion ($1.1 billion in annualized costs). About 60 percent of the connection costs occur in rural areas. 55 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The increase in demand resulting from increased access rates will also require additional power generation capacity: investment costs are $2 billion higher ($270 million in annualized costs) than in the current access-rate scenario. In addition, the variable costs of operating the system (fuel and O&M costs) are $700 million annually. The total annual costs of system expansion and operation increase to 5 percent of the region’s GDP in 2015. The costs of operating the existing system are only 0.5 percent of GDP. The costs of expanding the system amount to 4.4 percent of GDP. National targets for electricity access Meeting national targets requires $24 billion more than keeping access constant at the 2005 rate. This figure corresponds to about $3 billion in annualized capital costs (see tables 5.7–5.10). The largest contributors to the increased costs are the costs of transmission, distribution, and connection: connecting new households to the grid requires almost $20 billion ($2.4 billion in annualized costs). About 75 percent of the connection costs occur in rural areas. Additional power generation capacity is also needed to meet the resulting increase in demand: investment costs are $3.8 billion higher ($520 million in annualized costs) than in the 2005 access rate scenario. In addition, the variable costs of operating the system (fuel and O&M costs) are $1 billion higher in the national targets scenario than in the 2005 access scenario. The total annual costs of system expansion and operation increase to 5.7 percent in the national- targets scenario. Excluding the costs of operating the existing system, the costs of system expansion amount to 5.1 percent. The effect of trade Detailed figures are reported in tables 5.7–5.10. Less total investment and less hydropower The investment level is almost 7,700 MW lower in the trade-stagnation scenario. But investments in hydropower are 10,600 MW lower, while investments in gas-fired power plants are almost 3,000 MW higher. The share of hydropower in the region is reduced from 46 percent to 28 percent (figure 5.7). This difference in investment volume (in MW) is a result of gas having higher firm production capacity than hydropower. 56 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.7 Capacity mix in EAPP in 2015 with different trade assumptions Most notably under trade stagnation, gas-fired power capacity and production is much higher in Egypt and Kenya, while less hydropower is developed in Ethiopia, Sudan, Uganda, and Rwanda. Egypt eliminates net imports (42 TWh), while Ethiopia, Sudan, and Tanzania cease to be net exporters. Power production in individual countries (as a share of total production in the area) varies with different assumptions about trade. For example, under trade expansion, power production drops substantially in Egypt but soars in Ethiopia and Sudan (figure 5.8). Ethiopia and Sudan would become major power exporters, trading more than they produced for domestic consumption and sending their power northward into Egypt. With the existing transmission network in the trade-stagnation scenario, there is very little trade. Exploiting the region’s power resources in an optimal way results in huge trade flows from the south (primarily Ethiopia and Sudan) to the north (Egypt); see figure 5.9. 57 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.8 Production in EAPP in 2015 with different trade assumptions Percentage of regional total Figure 5.9 Trade flows with trade expansion (left) and trade stagnation (right) in EAPP in 2015, TWh Source: Authors’ original research. 58 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.11 shows the impact of trade expansion on the Table 5.11 Net imports (+)/exports (-) individual countries’ power balance: Ethiopia, Sudan, Rwanda, as share of domestic demand and Uganda become net exporters, while Egypt and Burundi Percent become net importers. Trade– Trade– expansion stagnation scenario scenario Costs and gains Burundi 78 0 In contrast to SAPP, in EAPP it is not possible to switch Djibouti 0 0 expensive hydropower for cheap hydropower Egypt 32 0 Overnight investment costs. Total overnight investment costs Ethiopia –227 1 are about $15 billion higher in the trade-expansion scenario than Kenya 22 3 Rwanda –191 0 in the trade-stagnation scenario. The costs related to Sudan –134 0 transmission-grid expansion are about $1 billion higher (since Tanzania –22 0 there are additional investments in cross-border transmission Uganda –61 –9 lines). The costs related to investments in generation capacity are Source: Authors’ original research. $14 billion higher in the trade-expansion scenario. Annualized costs. The total annualized costs of system expansion in the EAPP are about $1 billion lower in the trade-expansion scenario than in the trade-stagnation scenario (table 5.9). The annualized capital costs are $2 billion lower in the trade-stagnation scenario. There is a small savings ($126 million) from building less cross-border transmission capacity when trade is restricted, and a bigger savings ($1.8 billion) from building thermal instead of hydro. On the other hand, variable operating costs (including fuel costs) are $3 billion higher when trade is restricted. Trade expansion generates costs savings both from using the existing generating capacity ($1.4 billion), implying that an integrated system can use the existing system better, and from the different capacity composition ($1.5 billion). When import possibilities are limited, the new capacity has higher variable costs, and the existing (relatively old and expensive) capacity is used more. Returns on investing in trade. The capital costs in the case of trade expansion are $15 billion higher than in the trade-stagnation scenario. On the other hand, the annual costs of the enlarged system (variable costs) ten years from now are about $3 billion lower in the enlarged system. This implies that the extra expenditure is saved in five years. This is equivalent to a return of 20 percent. Savings are 7.6 percent of the annualized cost of expansion.15 Long-run marginal cost. Table 5.12 shows the long-run marginal cost in EAPP countries with trade expansion and trade stagnation. Comparing these figures with figure 5.8, we see that the cost of power increases from production to consumption centers, for instance from Ethiopia via Sudan to Egypt or from Tanzania to Uganda and Kenya. 15 Annual savings are $999 million. The cost of operation is $16,003 million, which includes the variable cost of current capacity ($2,815 million). $999 million is 7.6 percent of the difference between cost of operation and variable cost of current capacity, or $13,188 million. Note that the discounted value of $999 million into infinity at a 12 percent discount is $8,325 million. 59 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.12 Long-run marginal costs in EAPP $ per MWh Trade-expansion scenario Trade-stagnation scenario Cost of Cost of domestic Total LRMC Cost of Cost of domestic Total LRMC generation and transmission and generation and transmission and international distribution international distribution transmission transmission lines lines Burundi 68 46 114 103 46 149 Djibouti 66 6 72 66 6 72 Egypt 76 9 85 77 9 86 Ethiopia 69 122 190 40 122 161 Kenya 74 50 124 84 50 133 Rwanda 64 61 124 60 61 121 Sudan 75 52 127 74 52 126 Tanzania 65 32 97 45 32 78 Uganda 68 54 123 59 54 113 Average 74 47 121 75 47 122 Note: Average is weighted by annualized cost. In some cases power exporting countries report higher LRMC under trade expansion. Even if the cost of meeting domestic power consumption may be higher with trade than without; the higher revenues earned from exports would more than compensate for that increment. Source: Authors’ original research. The most striking result of the EAPP analysis, however, is the much higher cost figures compared to SAPP. This is primarily due to the long distances that the transmission lines must cover. It also reflects the scarcity of cheap resources in EAPP. Even in the trade-stagnation scenario, where no international transmission lines are built, the generation costs in Ethiopia (the cheapest in EAPP) are four times higher than in the Democratic Republic of Congo, the cheapest generator in SAPP. There is, of course, uncertainty in these figures and comparisons, which ultimately build on project cost estimates in the individual countries, which incorporate exchange rate considerations and other quantities not precisely known. Comparing the costs of generation in the cases of trade expansion and stagnation (columns 2 and 5 of table 5.12) reveals that relying on imports is much cheaper than expanding domestic power generation capacity in importing countries; in other words, the region benefits from gains from trade. In the case of the exporting countries Ethiopia and Tanzania, generation costs are greater in the trade-expansion scenario, reflecting the cost of international transmission lines necessary to export power. These countries may require compensation for the increased costs to trigger investments since the trade benefit accrues to the importing countries, not the exporters. 60 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA WAPP Maintaining electricity access at current rates Investment needs Almost 16,000 MW of new capacity is needed in the WAPP region over ten years to keep up with the demand growth resulting from general economic development and population growth. In addition, more than 5,500 MW of existing capacity must be refurbished. Table 5.13 provides an overview of generation capacity, refurbishment, and investment costs in 2015, as well as the resulting capacity mix in all scenarios. Table 5.13 Investment needs and generation mix in WAPP in 2015 Trade-expansion scenario Trade-stagnation Low-growth scenario scenario 2005 access rate Regional target National targets National targets National targets access rate for access rates for access rates for access rates, trade expansion Generation capacity (in MW) - Installed 4,096 4,096 4,096 4,096 4,096 - Refurbishment 5,530 6,162 6,972 6,842 5,535 - New investments 15,979 16,634 18,003 16,239 17,186 Generation capacity mix ( percent): - Hydro 82 79 77 73 80 - Coal 1 1 1 1 1 - Gas 13 14 16 19 12 - Other 4 5 6 7 7 Source: Authors’ original research. Note: Installed capacity refers to capacity in place in 2005 that is not refurbished before 2015. Existing capacity that is refurbished before 2015 is not included in the installed capacity figure, but in the refurbishment figure. Nigeria accounts for two-thirds of electricity consumption in the region. Hence, developments in Nigeria that influence electricity demand (such as economic development and the politically determined electricity access targets) have a large impact on the WAPP region’s total costs of electricity sector development. Nigeria does not, however, significantly affect the trade patterns and resource development in the rest of region. First of all, Nigeria is not centrally situated; therefore, any large exports would need large investments in transmission lines as well. Second, Nigeria meets domestic demand growth with large and relatively cheap hydropower resources. The ample gas resources that could be used to develop gas-fired power plants are more expensive than hydropower in other countries. Ghana accounts for 15 percent and Cote d’Ivoire for 6 percent of the region’s electricity demand. In contrast to Nigeria, these countries import about half of their electricity demand with 2005 access targets. 61 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Guinea produces almost 20 percent of the region’s power, exporting large amounts of hydropower, which has a competitive cost. Guinea exports more than eight times domestic demand. Figure 5.10 shows demand and production in the different countries in WAPP as a share of the region’s total. Figure 5.10 Demand and production in WAPP countries in 2015 Current access scenario; percentage of regional total In total, the WAPP region carries out 16,000 MW investments in power generation in the 2005 access scenario. Almost all of this is hydropower:16 10,290 MW in Nigeria, 4,290 MW in Guinea, 1,000 MW in Ghana, and 130 MW in Cote d’Ivoire. This means that Nigeria, Guinea, and Ghana fully exploit17 their available hydropower resources.18 In addition, Senegal builds one coal-fired power plant (250 MW) and some off-grid technologies in rural areas. In addition to investments in new generation capacity, 5,530 MW of existing capacity is refurbished: almost 4,000 MW of hydropower (2,850 MW in Nigeria), 1,200 MW of natural gas-fired power in Nigeria, and 410 MW of HFO-fueled thermal power plants in various countries. 16 In addition, there are minimal investments in off-grid technologies in rural areas. 17 Note that “fully exploited” refers to the assumed maximum potential for hydropower in the model. In most cases, this maximum potential has been set equal to identified projects and plans, even though the full hydropower potential of a country may be much larger. The identified projects serve as a proxy for developments that are realistic in the time frame in focus here (the ten years before 2015). 18 As we only use one (average) investment cost per technology per country, not individual costs per project, cheaper resources are often fully utilized in one country before the more expensive resources are developed in a neighboring country. The cost of building international transmission lines counteracts this to some extent. 62 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Overnight investment costs Generation costs. The investment costs of expanding the generation system between 2005 and 2015 are slightly more than $23.3 billion (table 5.14). The largest share of this by far is investment in new capacity ($22 billion), while the cost of refurbishment is only $1.4 billion. Transmission, distribution, and connection costs are almost equal to the costs of new generation capacity: $23.3 billion in investment is needed to expand and refurbish the grid (table 5.14). Investments in new transmission and distribution lines account for most of this total (over $17 billion). Only 6 percent of this last figure is related to international transmission lines. The direct cost of connecting new customers to the grid is less than 20 percent of the total grid cost: $4.3 billion is needed to maintain the access rate at the 2005 level (aligned with population growth) (table 5.14). Of this total, 86 percent would be spent in urban areas. Total overnight investment costs. These are $46.6 billion in the 2005 access rate scenario. Table 5.14 Overnight investment costs in WAPP, 2005–15 $ millions Trade-stagnation Low-growth Trade-expansion scenario scenario scenario National targets 2005 access Regional target National targets National targets for access rates, rate access rate for access rates for access rates trade expansion Generation Investment cost 21,955 23,632 26,992 25,822 25,128 Refurbishment cost 1,363 1,429 1,511 1,496 1,366 T&D and connection Investment cost 17,241 22,399 29,813 28,872 23,206 - Cross-border transmission lines 1,022 968 941 0 912 - Distribution grid 11,909 11,909 11,909 11,909 5,332 - Connection cost (urban) 3,698 5,254 7,634 7,634 7,634 - Connection cost (rural) 612 4,268 9,329 9,329 9,329 Refurbishment cost 6,057 6,057 6,057 6,057 6,057 Total 46,615 53,518 64,373 62,247 55,758 Source: Authors’ original research. Annualized costs The annualized capital cost of meeting market demand and maintaining current access over ten years is $6 billion, of which almost $3 billion is for transmission, distribution, and connection, and $3.1 billion for generation (table 5.15). In addition, the annual variable costs of operating the system (that is, fuel and variable operation and maintenance costs) amount to $3.2 billion. About half of this is related to operating new power plants, while the other half is related to operating existing and refurbished power plants ($0.3 billion) and the grid ($1.3 billion). 63 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The total annualized cost of system expansion is equal to 2.1 percent of GDP of the region in 2015 (table 5.16). Adding the variable operation costs of existing capacity, the annualized costs of system expansion and operation are equal to 3.2 percent of GDP. This regional figure conceals large differences between costs in different countries with very different investment patterns. Guinea invests in hydropower for export purposes, and the investment costs are as much as 20 percent of GDP. In Gambia, the fuel costs of existing HFO-fueled capacity contribute to variable costs, which equal up 4.5 percent of GDP. In addition, the grid cost accounts for another 1 percent of GDP. In Senegal, both the grid-related (investment and variable) and variable generation costs lift the total to 7 percent of GDP. Detailed country tables can be found in annex 1 (separately bound). Table 5.15 Annualized costs of system expansion in WAPP, 2015 $ millions Trade-stagnation Low-growth Trade-expansion scenario scenario scenario National targets Regional target National targets National targets 2005 access rate for access rates, access rate for access rates for access rates trade expansion Generation Investment cost 2,874 3,092 3,527 3,365 3,291 Refurbishment cost 237 247 258 258 237 Variable cost (fuel, O&M) 1,887 2,259 2,728 3,442 1,912 - New capacity 260 266 275 620 267 - Existing capacity 1,627 1,993 2,453 2,822 1,646 T&D and connection Investment cost 2,140 2,781 3,701 3,584 2,881 - Cross-border 127 120 117 0 113 - Distribution grid 1,478 1,478 1,478 1,478 662 - Urban connection 459 652 948 948 948 - Rural connection 76 530 1,158 1,158 1,158 Refurbishment cost 752 752 752 752 752 Variable cost (existing 1,320 1,320 1,320 1,320 342 capacity) Total Capital cost 6,003 6,871 8,238 7,959 7,161 - Investment cost 5,014 5,873 7,228 6,949 6,171 - Refurbishment cost 989 998 1,010 1,010 989 Variable cost 3,207 3,579 4,049 4,763 2,254 Total 9,210 10,450 12,287 12,722 9,415 Source: Authors’ original research. Note: We assume that there are no variable costs related to the new T&D assets. 64 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.16 Annualized costs as a share of GDP in WAPP, 2015 Percent Trade-expansion scenario Trade- Low-growth stagnation scenario scenario 2005 access Regional National National National rate target targets for targets for targets for access rate access rates access rates access rates, trade expansion Generation Investment cost 1.0 1.1 1.2 1.2 1.3 Refurbishment cost 0.1 0.1 0.1 0.1 0.1 Variable cost (fuel, O&M) 0.6 0.8 0.9 1.2 0.7 - New capacity 0.1 0.1 0.1 0.2 0.1 - Existing capacity 0.6 0.7 0.8 1.0 0.6 T&D and connection Investment cost 0.7 1.0 1.3 1.2 1.1 - Cross-border 0.0 0.0 0.0 0.0 0.0 - Distribution grid 0.5 0.5 0.5 0.5 0.3 - Urban connection 0.2 0.2 0.3 0.3 0.4 - Rural connection 0.0 0.2 0.4 0.4 0.5 Refurbishment cost 0.3 0.3 0.3 0.3 0.3 Variable cost (existing capacity) 0.5 0.5 0.5 0.5 0.5 Total Capital cost 2.1 2.4 2.8 2.7 2.8 - Investment cost 1.7 2.0 2.5 2.4 2.4 - Refurbishment cost 0.3 0.3 0.3 0.3 0.4 Variable cost 1.1 1.2 1.4 1.6 1.3 Total 3.2 3.6 4.2 4.4 4.0 Source: Authors’ original research. Note: We assume that there are no variable costs related to the new T&D assets. Cost effects of raising access rates Regional target rate: Electricity access of 53 percent on average Compared with the 2005 access rate, meeting the regional target for electricity access (53 percent on average) requires an additional investment of almost $7 billion. This figure corresponds to about $1.25 billion in annualized capital costs (see tables 5.13 to 5.16). The largest contributors to this increase are the costs of transmission, distribution, and connection: connecting new households to the grid involves extra costs of over $5 billion ($600 million in annualized costs). Almost half of this amount is spent in rural areas. 65 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Additional power-generating capacity is also needed to meet increased demand: investment costs are $1.7 billion higher ($200 million in annualized costs) than in the 2005 access rate scenario. The costs of operating the system (variable costs) are 12 percent higher—almost $400 million annually—since part of the new generation capacity is fossil-fueled (diesel in rural areas and refurbishment of gas-fired power plants in Nigeria). The annualized costs of system expansion are 2.4 percent of GDP in 2015 for the region. Including variable costs of existing capacity lifts the total annualized cost of system expansion and operation to 3.6 percent of GDP. National targets for electricity access Meeting national targets requires almost $18 billion more than keeping the access rate constant at 2005 rates. This corresponds to about $3 billion in annualized costs (see tables 5.13 to 5.16). The largest contributors to this increase are the costs of transmission, distribution, and connection. For example, connecting new households to the grid involves an extra investment of $12.5 billion ($1.6 billion in annualized costs). More than half of new connection costs are spent in rural areas, compared with only 15 percent in the current-access scenario. Additional power generation capacity is also needed to meet the increased demand: investment costs are more than $5 billion higher ($650 million in annualized costs) than in the 2005 access rate scenario. The additional costs of operating the system (variable costs) are $850 million annually. The annualized costs of system expansion are 2.9 percent of GDP in 2015 for the region. To include variable costs of current capacity increases the total annualized costs of system expansion and operation to 4.2 percent of GDP. The effect of trade Detailed figures are reported in tables 5.13 to 5.16. Dispersed hydropower and gas-fired power plants substitute for hydropower in Guinea The most visible effect of restricting investments in cross-border transmission lines is reduced hydropower development in Guinea (3,720 MW of forgone investments). As there are no lines from Guinea to neighboring countries today, Guinea will only produce enough to meet domestic demand (figure 5.12). Countries that have no international transmission lines today—Guinea-Bissau, Liberia, Sierra Leone—cannot benefit from imports from Guinea and must be self-sufficient. These countries invest in domestic hydropower to compensate for the lack of imports. In the trade-expansion scenario, electricity flowed from Guinea through Cote d’Ivoire to Burkina Faso and Ghana. Without this flow, Cote d’Ivoire increases investments (193 MW) and production (2 TWh) to export to Burkina Faso and Ghana on existing lines (figure 5.13). Similarly, Togo (a net importer in the trade-expansion scenario ) now exports to Ghana and Benin. Also Senegal, Mauritania, 66 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA and Mali benefited from imports from Guinea in the trade-expansion scenario. In the trade-stagnation scenario, Senegal becomes a net importer from Mauritania and Mali.19 In total, hydropower investments increase 1,270 MW in countries other than Guinea, so the total decrease in hydropower capacity in the region is 2,450 MW. More thermal power investments replace hydropower. Additional gas-fired power plants in Ghana, Benin, Togo, and Mauritania total 700 MW. Nigeria, on the other hand, reduces its exports to Niger and Benin and refurbishes fewer of its gas-fired power plants. Since thermal power plants can produce more power per MW installed compared to hydropower, total investments (measured in MW) are reduced by 1,700 MW. The total production, however, remains the same. The trade-expansion scenario leads to a generation mix with a somewhat higher proportion of renewable energy: hydropower and natural gas supply 77 percent and 16 percent of total capacity, respectively, in the trade-expansion scenario, compared to 73 percent and 19 percent in the trade- stagnation scenario (figure 5.11). Table 5.17 shows the impact of trade expansion on each country’s power balance. Figure 5.11 Capacity mix in WAPP in 2015, with different trade assumptions 19 Imports from Mali include power from the Manantali hydropower plant (located in Mali but co-owned by Mali, Mauritania, and Senegal). 67 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.12 Production in WAPP in 2015, with different trade assumptions Percentage of regional total Figure 5.13 Trade flows with trade expansion (left) and trade stagnation (right) in WAPP in 2015, TWh Source: Authors’ original research. 68 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.17 Net imports (+)/exports (–) as share of domestic demand Percent Trade-expansion Trade-stagnation scenario scenario Benin 45 18 Burkina Faso 58 58 Cote d'Ivoire -12 -47 Gambia -19 0 Ghana 52 14 Guinea -564 0 Guinea-Bissau 77 0 Liberia 89 0 Mali 79 -14 Mauritania 55 -23 Niger 86 20 Nigeria -3 0 Senegal 30 13 Sierra Leone 60 0 Togo 48 -27 Source: Authors’ original research. Costs and gains Overnight investment costs. Total overnight capital costs are $2.1 billion higher in the trade- expansion scenario. The costs related to transmission grid expansion are about $0.9 billion higher (since there are additional investments in cross-border transmission lines), and costs related to investments in generation capacity are $1.2 billion higher. Annualized costs. Although overnight investment costs are higher under trade expansion, annualized costs are lower since they account for savings on variable costs. While annualized investments costs are $280 million higher in the WAPP region under trade expansion, the total annualized costs of system expansion are almost $450 million lower. Both generation and transmission investments are higher in the trade-expansion scenario (due to investments in international transmission lines and the higher volume of hydropower investments required). But the operation of the integrated system in the trade-expansion scenario requires significantly lower variable costs (that is, fuel costs and variable O&M costs). These cost savings (over $700 million) stem from the greater proportion of hydropower in the trade-expansion scenario). Returns on investing in trade. The gains from trade can be regarded as returns on investment. Trade- expansion requires $2.1 billion more in investment costs than the trade-stagnation scenario. Annual variable costs, however, are more than $700 million lower in the enlarged system. Therefore, the extra investment cost is recouped in three years. This is equivalent to an annual return on investment of 33 percent. Annual savings from trade expansion are 5.1 percent of the annualized cost of expansion. 69 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.18 Long-run marginal costs in WAPP $ per MWh Trade-expansion scenario Trade-stagnation scenario Cost of Cost of Cost of Cost of generation and generation and domestic Total long-run domestic Total long-run international international transmission marginal costs transmission marginal costs transmission transmission and distribution and distribution lines lines Benin 79 111 190 81 111 192 Burkina Faso 72 181 253 79 181 260 Cote d’Ivoire 69 78 147 76 78 154 Gambia 63 17 80 58 17 74 Ghana 73 23 96 80 23 103 Guinea 58 13 70 47 13 60 Guinea-Bissau 63 22 85 134 22 156 Liberia 66 15 81 126 15 141 Mali 64 182 246 97 182 279 Mauritania 69 67 136 78 67 145 Niger 79 168 247 136 168 304 Nigeria 76 52 128 76 52 128 Senegal 66 368 434 100 368 468 Sierra Leone 61 24 86 73 24 97 Togo 76 27 103 80 27 106 Average 72 111 183 80 111 191 Note: Average is weighted by annualized cost. In some cases power exporting countries report higher LRMC under trade expansion. Even if the cost of meeting domestic power consumption may be higher with trade than without; the higher revenues earned from exports would more than compensate for that increment. Table 5.18 compares the long-run marginal cost of power in WAPP countries under trade expansion and trade stagnation. The cost figures illustrate how trade expansion increases the long-run marginal cost of power generation for exporters while the opposite is true for importers (see figure 5.12). For example, the cost of power generation (column 2) is lowest in Guinea (a net exporter); however, the costs of international transmission lines for each country through which power flows (for example via Sierra Leone to Liberia or via Cote d’Ivoire to Burkina Faso, Ghana, or Togo) adds to the long-run marginal cost. On the other hand, countries that import in the trade-expansion scenario experience much higher long-run marginal costs of power generation (column 5) under trade stagnation due to the cost of expanding domestic power-generation capacity. 70 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA CAPP Maintaining electricity access at current rates Investment needs To keep up with demand growth resulting from general economic development and population growth, 3,850 MW of new capacity are needed in the CAPP region over ten years. Virtually all new capacity investment is in hydropower:20 2,430 MW in Cameroon, 1,320 MW in the Republic of Congo, 85 MW in Gabon, and 25 MW in the Central African Republic. In this scenario, Cameroon fully exploits its available hydropower resources.21 In addition, more than 900 MW of existing capacity must be refurbished (600 MW in Cameroon and the remainder in Gabon, the Republic of Congo, and the Central African Republic). Table 5.19 provides an overview of generation capacity in 2015, ten-year refurbishment and investment volumes, and the resulting capacity mix for all scenarios. Table 5.19 Investment needs and generation mix in CAPP, 2015 Trade- Low-growth Trade-expansion scenario stagnation Scenario scenario 2005 access rate Regional target National targets National targets National targets access rate for access rates for access rates for access rates, trade expansion Generation capacity (in MW) - Installed 260 260 260 260 260 - Refurbishment 906 906 906 1,081 906 - New investments 3,856 4,143 4,395 3,833 3,915 Generation capacity mix (%): - Hydro 97 97 97 83 97 - Coal 0 0 0 0 0 - Gas 0 0 0 0 0 - Other 2 3 3 17 3 Source: Authors’ original research. Note: Installed capacity refers to capacity in place in 2005 that is not refurbished before 2015. Existing capacity that is refurbished before 2015 is not included in the installed capacity figure, but in the refurbishment figure. The Republic of Congo accounts for more than half (54 percent) of electricity demand in CAPP in 2015, and Cameroon for one-third. Therefore, the development of these countries is an important driving 20 There are very small investments in off-grid technologies in rural areas. 21 The term “fully exploited” refers to the assumed maximum potential for hydropower in the model. In most cases, this maximum potential has been set equal to identified projects and plans, even though the full hydropower potential of a country may be much larger. The identified projects serve as a proxy for realistic development in the timeframe considered (before 2015). 71 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA force for the rest of the region. Gabon has 10 percent of the region’s total demand, while the other CAPP countries demand only small amounts of power. Cameroon accounts for 64 percent of total electricity production in the region in 2015, compared to only 29 percent for the Republic of Congo. Cameroon exports more than a third (5.6 TWh) of its production to the Republic of Congo, as well as small amounts to Gabon, Chad, and Equatorial Guinea. We assume that imports from the Democratic Republic of Congo to the Republic of Congo remain at 2005 levels, although this is a small volume (less than 0.5 TWh/year). Figure 5.14 shows demand and production in each CAPP country as a share of the region’s total. Figure 5.14 Demand and production in CAPP countries in 2015 Current access scenario; percentage of regional total Overnight investment costs Generation costs. The investment costs of expanding the generation system by 2015 are almost $6 billion (table 5.20), the largest share of which is investment in new capacity ($5.6 billion). Refurbishment accounts for only $0.3 billion of total costs. Transmission, distribution, and connection costs are much lower than the costs of building new power plants, accounting for less than 20 percent of total investment costs: in the expansion and refurbishment of the existing grid requires $1.3 billion in investment (table 5.20). Investments in new transmission and distribution lines account for most of this sum (over $1 billion), of which a full third is investment in international transmission lines. Maintaining the access rate at the 2005 level (aligned with population growth) requires $0.4 billion (40 percent of total grid investment) to cover the costs of connecting new customers to the grid, 98 percent of which would be spent in urban areas. 72 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Total overnight investment costs. In the 2005 access rate scenario, the total overnight investment is slightly more than $7 billion: generation costs are $6 billion and grid-related costs are $1.3 billion. Table 5.20 Overnight investment costs in CAPP, 2005–15 $ millions Trade- Low-growth Trade-expansion scenario stagnation scenario scenario National targets 2005 access Regional target National targets National targets for access rates, rate access rate for access rates for access rates trade expansion Generation Investment cost 5,645 6,157 6,615 5,981 5,766 Refurbishment cost 272 272 272 301 272 T&D and connection Investment cost 1,057 1,648 2,348 2,036 2,311 - Cross-border 349 317 312 0 355 - Distribution grid 286 286 286 286 205 - Connection cost (urban) 412 753 1,010 1,010 1,010 - Connection cost (rural) 10 292 740 740 740 Refurbishment cost 222 222 222 222 222 Total 7,196 8,299 9,457 8,540 8,570 Source: Authors’ original research. Annualized costs The annualized capital cost of meeting market demand and maintaining current access over ten years is almost $1 billion: this includes almost $160 million in transmission, distribution, and connection costs and $780 million in generation costs (table 5.21). The annual variable costs of operating the system (that is, fuel and variable operation and maintenance costs) amount to $150 million. About $50 million of this is related to operating new power plants, while the rest is related to operating existing and refurbished power plants ($30 million) and the grid ($70 million). The total annualized cost of system expansion in the 2005 access rate scenario is about $1 billion, or 1.4 percent of the region’s GDP in 2015 (table 5.22). Adding the variable operation costs of existing capacity, the annualized costs of system expansion and operation are 1.6 percent of GDP. The varied investment patterns of CAPP countries lead to different annualized costs. For example, countries with significant hydropower development have higher costs of generation capacity expansion: 3 percent of GDP in the Republic of Congo and 1.6 percent in Cameroon. Grid-related costs (investments, refurbishment, and operation) in the Republic of Congo are 0.3 percent of GDP. This high total results from the new cross-border lines needed to accommodate the country’s substantial imports. Cameroon also invests 0.3 percent of GDP in grid-related costs, composed mainly of the cost of connecting new customers to the grid and investments in domestic and cross-border grid. 73 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Other countries in CAPP—Chad and Equatorial Guinea—do not invest in new generation capacity. Investment in these countries consists only of grid extension, maintenance, and new connection, the costs of which are quite low relative to new generation capacity. Annualized cost figures for the region are therefore lower than what might be expected. Detailed country tables can be found in annex 1 (separately bound). Table 5.21 Annualized costs of system expansion in CAPP, 2015 $ millions Trade-stagnation Low-growth Trade-expansion scenario scenario scenario National targets 2005 access Regional target National targets National targets for access rates, rate access rate for access rates for access rates trade expansion Generation Investment cost 734 800 860 798 751 Refurbishment cost 48 48 48 53 48 Variable cost (fuel, O&M) 84 88 92 347 84 - New capacity 51 55 61 259 51 - Existing capacity 33 33 31 89 33 T&D and connection Investment cost 131 205 292 253 287 - Cross-border 43 39 39 0 44 - Distribution grid 35 35 35 35 25 - Urban connection 51 93 125 125 125 - Rural connection 1 36 92 92 92 Refurbishment cost 28 28 28 28 28 Variable cost (existing capacity) 67 67 67 67 67 Total Capital cost 941 1,081 1,227 1,131 1,113 - Investment cost 865 1,005 1,151 1,051 1,037 - Refurbishment cost 76 76 76 81 76 Variable cost 151 156 159 414 151 Total 1,092 1,236 1,386 1,546 1,264 Source: Authors’ original research. Note: We assume that there are no variable costs related to the new T&D assets. 74 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.22 Annualized costs as a share of GDP in CAPP, 2015 Percent Trade- Low-growth Trade-expansion scenario stagnation scenario scenario National National National targets for 2005 access Regional target targets for targets for access rates, rate access rate access rates access rates trade expansion Generation Investment cost 1.1 1.1 1.2 1.1 1.2 Refurbishment cost 0.1 0.1 0.1 0.1 0.1 Variable cost (fuel, O&M) 0.1 0.1 0.1 0.5 0.1 - New capacity 0.1 0.1 0.1 0.4 0.1 - Existing capacity 0.0 0.0 0.0 0.1 0.1 T&D and connection Investment cost 0.2 0.3 0.4 0.4 0.5 - Cross-border 0.1 0.1 0.1 0.0 0.1 - Distribution grid 0.1 0.1 0.1 0.1 0.0 - Urban connection 0.1 0.1 0.2 0.2 0.2 - Rural connection 0.0 0.1 0.1 0.1 0.1 Refurbishment cost 0.0 0.0 0.0 0.0 0.0 Variable cost (existing capacity) 0.1 0.1 0.1 0.1 0.1 Total Capital cost 1.3 1.5 1.8 1.6 1.7 - Investment cost 1.2 1.4 1.7 1.5 1.6 - Refurbishment cost 0.1 0.1 0.1 0.1 0.1 Variable cost 0.2 0.2 0.2 0.6 0.2 Total 1.6 1.8 2.0 2.2 2.0 Source: Authors’ original research. Note: We assume that there are no variable costs related to the new T&D assets. Cost effects of raising access rates Regional target for access rate: electricity access of 44 percent on average Compared with the 2005 access rate, meeting the international target for electricity access in CAPP (44 percent on average) requires an additional investment of $1.1 billion. This equates to about $140 million in annualized capital costs (see tables 5.19 to 5.22). The largest contributors to this increase are the costs of transmission, distribution and connection: connecting new households to the grid involves additional costs of about $0.6 billion ($80 million in annualized costs). Almost 30 percent of the total connection costs are spent in rural areas, compared to only 2 percent in the 2005 access scenario. 75 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Meeting the regional target for access rates also requires additional power generation to satisfy the increased demand: investment in generation capacity is $0.5 billion higher ($66 million in annualized costs) compared to the 2005 access-rate scenario. Since some of the new generation capacity in rural areas is based on off-grid diesel generators, the additional costs of operating the system (variable costs) increase slightly. The annualized costs of system expansion are about 1.6 percent of the region’s GDP in 2015. Including variable costs of existing capacity lifts the total annualized cost of system expansion and operation to 1.8 percent of GDP. National targets for electricity access Meeting national targets for access rates requires $2.3 billion more in investment compared to maintaining the access rate constant at 2005 levels. This corresponds to $300 million more in annualized costs (see tables 5.19 to 5.22). The largest contributors to this increase are the costs of transmission, distribution, and connection. For example, connecting new households to the grid involves an extra cost of about $1.3 billion ($165 million in annualized costs). More than 40 percent of this total is spent in rural areas, compared with only 2 percent in the 2005 access-rate scenario. Investment in generation capacity is also almost $1 billion higher ($126 million in annualized costs) than in the 2005 access rate scenario to meet increased demand. The annualized costs of system expansion are 1.8 percent of the region’s GDP in 2015. Including variable costs of current capacity increases the total annualized costs of system expansion and operation to 2 percent of GDP. The effect of trade Detailed figures are reported in tables 5.19 to 5.22. Thermal power substitutes for hydropower In the trade-stagnation scenario, all CAPP countries must rely on domestic power generation. Therefore, the Republic of Congo, Chad, and Gabon increase their production considerably while Cameroon reduces production correspondingly (figure 5.16 and figure 5.17). The Central African Republic was self-sufficient in the trade-expansion scenario, and production therefore remains the same. The most visible effect of restricting investments in cross-border transmission lines is reduced hydropower development, especially in Cameroon (1,450 MW of forgone investments). These investments contributed to exports to the Republic of Congo, Chad, Gabon, and Equatorial Guinea in the trade-expansion scenario. In the trade-stagnation scenario, CAPP countries cannot import power to meet demand. Therefore, investments in domestic generation capacity in the Republic of Congo, Chad, Gabon and Equatorial Guinea replace investments in exportable capacity in Cameroon. The new generation capacity is composed of HFO (over 700 MW new and refurbished capacity in the four countries combined) and hydro in the Republic of Congo (330 MW) and Equatorial Guinea (30 MW). In addition, production in existing HFO-fueled plants in Gabon increases considerably. 76 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Total capacity is also reduced (400 MW), since thermal power plants produce more per MW installed than hydropower. Total production is unchanged since demand is the same in the trade-expansion and trade-stagnation scenarios. Figure 5.15 Capacity mix in CAPP in 2015 with different trade assumptions The trade-expansion scenario leads to a generation mix with a higher proportion of renewable energy and lower proportion of thermal energy: 97 percent of total capacity is supplied by hydropower and 3 percent by HFO in the trade-expansion scenario, compared to 83 percent and 17 percent, respectively, in the trade-stagnation scenario (figure 5.15). The impact of trade on the individual Table 5.23 Net imports (+)/exports (-) as share of domestic demand countries’ power balance is illustrated in Percent figure 5.16 and table 5.23. Under trade Trade-expansion Trade-stagnation expansion, Cameroon exports about half of Scenario scenario its production, while Chad and Equatorial Cameroon -84 0 Guinea import all of their consumption. Central African Republic 0 0 The Republic of Congo imports about one Chad 102 0 third of its consumption and Gabon almost Congo, Rep. 34 0 a half. The Central African Republic is Equatorial Guinea 100 0 self-sufficient. Cameroon is the largest Gabon 42 0 electricity supplier in the CAPP region. Source: Authors’ original research. 77 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 5.16 Production in CAPP in 2015 with different trade assumptions Percentage of regional total Figure 5.17 Trade flows with trade expansion (left) and trade stagnation (right) in CAPP in 2015, TWh Source: Authors’ original research. Costs and gains Overnight investment costs. Total overnight capital costs are $920 million higher in the trade- expansion scenario than in the trade-stagnation scenario. The costs related to transmission grid expansion are about $310 million higher (since there are additional investments in cross-border transmission lines). 78 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The costs related to investments in generation capacity are $630 million higher in the trade-expansion scenario. Annualized costs. Although overnight investment costs are higher under trade expansion, annualized costs are lower since they account for savings on variable costs. The total annualized costs of system expansion and operation in the CAPP region are $160 million lower in the trade-expansion scenario than in the trade-stagnation scenario. Annualized capital costs are $100 million higher in the trade-expansion scenario. Both generation and transmission investments are higher in the trade-expansion scenario (due to the higher volume of hydropower needed and the lack of international transmission lines in the trade- stagnation scenario). But the operation of the integrated system in the trade-expansion scenario requires significantly lower variable costs (that is, fuel costs and variable O&M costs). These cost savings ($255 million) stem from the higher proportion of hydropower in the trade-expansion scenario. Returns on investing in trade. The gains from trade can be regarded as returns on investment. Capital costs in the trade-expansion scenario are $920 million higher than in the trade-stagnation scenario due to the costs of replacing thermal power with hydropower and constructing transmission lines. On the other hand, the annual costs of the enlarged system (variable costs) in 2015 are $255 million lower in the enlarged system. Therefore, the additional capital investment is recouped in less than four years through savings in variable cost expenditures. This is equivalent to a return on investment of 28 percent. Annual savings are 11.5 percent of the annualized cost of expansion. Long-run marginal cost. Table 5.24 compares the long-run marginal cost of power in CAPP countries under trade-expansion and trade-stagnation. The cost figures illustrate how trade expansion increases the long-term marginal cost of power generation for exporters while the opposite is true for importers (see figure 5.17): the cost of power generation (column 2) is the lowest in Cameroon; however, the costs of international transmission lines needed to export power (which increase with distance) add to the long-run marginal cost. On the other hand, countries that import in the trade-expansion scenario experience much higher long-run marginal costs of power generation (column 5) under trade stagnation due to the cost of expanding domestic power-generation capacity (especially Chad, but also the Republic of Congo and Equatorial Guinea). 79 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table 5.24 Long-run marginal costs in CAPP $ per MWh Trade-expansion scenario Trade-stagnation scenario Cost of Cost of Cost of Cost of generation generation domestic Total long-run domestic Total long-run and and transmission marginal transmission marginal international international and costs and costs transmission transmission distribution distribution lines lines Cameroon 44 24 69 40 24 64 Central African Republic 48 63 111 48 63 111 Chad 49 20 68 90 20 109 Congo, Rep. 54 2 56 79 2 81 Equatorial Guinea 48 27 75 70 27 97 Gabon 49 16 65 59 16 74 Average 49 22 70 70 22 91 Note: Average is weighted by annualized cost. In some cases power exporting countries report higher LRMC under trade expansion. Even if the cost of meeting domestic power consumption may be higher with trade than without; the higher revenues earned from exports would more than compensate for that increment. CO2 emissions and their reduction in Sub-Saharan Africa Throughout this study we have reiterated that least-cost trade expansion in Sub-Saharan Africa will result in the development of more hydropower plants, which have lower emissions than thermal power plants. Total reductions amount to about 70 million tons CO2 annually in Sub-Saharan Africa (table 5.25). It is quite interesting that in the SAPP region, trade results in CO2 reductions that are about as large as those that would be produced by the Clean Development Mechanism (CDM; see chapter 6). Table 5.25 Difference in electricity production and CO2 emissions in the trade-expansion and trade-stagnation scenarios SAPP EAPP WAPP CAPP Total Production difference (TWh) Coal -41.5 0.7 -40.8 Diesel -0.3 0.3 -0.8 -0.8 Gas -5.3 -42.4 -9.2 -56.8 HFO 0.4 0.2 -4.9 -4.3 Hydro 47.5 43.4 11.5 5.1 107.4 Total production difference (TWh) 0.5 2.4 1.6 0.3 4.7 Emissions savings (M ton) Coal -37.8 0.6 -37.2 Diesel -0.2 0.2 -0.6 -0.6 Gas -2.7 -21.5 -4.7 -28.9 HFO 0.3 0.1 -3.6 -3.2 Hydro 0.0 Total emission savings (M ton) -40.7 -20.4 -5.2 -3.6 -69.9 Source: Authors’ original research. 80 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Note: The emission factors are based on the carbon content of the fuel and on the efficiencies of different power plants as assumed in the model. The following emission factors have been used: coal 0.912 kg/kWh, diesel 0.788 kg/kWh, natural gas 0.508 kg/kWh, and HFO 0.745 kg/kWh. Emissions associated with hydro reservoirs are not calculated. The International Energy Agency (2008) estimated that 2006 emissions from power and heat production in Africa were 360 million tons. The savings from trade expansion in Sub-Saharan Africa amount to 20 percent of this volume. Our estimate does not, however, include greenhouse gas emissions from hydropower in the form of methane from dams. 81 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA 6 Sensitivity analyses The accuracy of the results presented in chapter 5 depend heavily on assumptions of parameter values. We have therefore tested the sensitivity of our results to eight variables: low economic growth, higher fuel prices, climate change (assumed to affect the reliability of hydropower), the price of credits available under the Clean Development Mechanism (CDM), barriers to hydropower development, higher unit investment costs of coal-fired generation and hydropower, and more imports from the Democratic Republic of Congo to the CAPP region. Low economic growth This section presents a scenario in which each region experiences lower economic growth. Economic growth directly influences market-driven demand by raising income and indirectly by leading to structural adjustments (which then influence demand). We examine the extent to which electricity sector investments will differ if per capita economic growth is reduced by 50 percent compared to the baseline scenario. Detailed results for this sensitivity analysis were included in the tables in chapter 5. SAPP A 50 percent lower rate of economic growth lowers electricity demand growth in the SAPP region, significantly reducing investment requirements (by almost 40 percent). The largest reduction is in South Africa, where 12,000 MW less is invested in new coal-fired power plants. South Africa’s large impact on investments is expected—since South Africa accounts for about 80 percent of total power demand in the SAPP region, any change there will affect the region’s power balance. Hydropower investments in the region are reduced by only 400 MW, entirely in the Democratic Republic of Congo. The distance between the region’s largest load center (South Africa) and the Democratic Republic of Congo is too large enough for cost-effective investments in hydropower and the accompanying transmission lines. The result confirms that lower demand growth has a smaller effect on hydropower investments than investments in thermal power. About 97 percent of the hydropower capacity that is built in the normal growth scenario remains in the low growth scenario. The reduction of hydropower investments in the Democratic Republic of Congo is only about 5 percent of total investment. This result confirms the importance of trade: even if demand grows at a slower pace, power trading enlarges the overall market enough to make expansion of hydropower capacity preferable to thermal power options. Intuitively, the model confirms that lower economic growth does not affect the refurbishment of aging infrastructure, which is usually the most economical way to increase generating capacity. 82 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA EAPP In EAPP, lower demand growth reduces investments in gas-fired power plants in Egypt (by 1,921 MW) and in Tanzania (by 180 MW). Similar to South Africa in SAPP, Egypt is by far the EAPP’s largest power consumer, accounting for 70 percent of the region’s demand. Lower production in Tanzania also reduces net exports from there. WAPP In the WAPP region, 80 percent of the total demand reduction (14.6 TWh) occurs in Nigeria, and another 14 percent in Ghana. Refurbishments of gas-fired power plants are reduced by 1,440 MW in Nigeria, resulting in a 13 TWh decrease in total production. Exports from Nigeria to Benin are also reduced (but are partly replaced by imports from Togo). Since Nigeria is a net exporter, lower domestic demand there does not influence investments in neighboring countries. Hydropower investments are also lower by 800 MW in Cote d’Ivoire, accounting for a fall in production of 3.6 TWh. This production was exported to Ghana. In the normal growth scenario, Cote d’Ivoire was a huge transit country and a net exporter. With lower domestic investments, Cote d’Ivoire becomes a net importer. But power still flows through Cote d’Ivoire, mainly to Ghana and, farther out, to Togo and Benin. Total capacity reduction is 2,250 MW in the WAPP region. CAPP Total electricity demand in CAPP is 9 percent (2.25 TWh) lower, mostly due to decreases in demand in the Republic of Congo (10 percent) and Cameroon (9 percent). Lower demand growth reduces hydropower investments in the Republic of Congo by almost 30 percent (470 MW). The Republic of Congo replaces lower domestic investment with imported power from Cameroon, which has lower unit costs for hydropower and does not experience a change in investments or production, despite lower domestic demand. Exports from Cameroon to Congo increase by 0.8 TWh. Total generation capacity in the CAPP region is 10 percent lower. The impact of fuel prices in EAPP The simulations presented in chapter 5 revealed large differences in investment patterns and costs in the trade-expansion and trade-stagnation scenarios; for example, more hydropower capacity is built in the trade-expansion scenario. Hydropower has substantially higher investment costs than thermal power plants; therefore, up-front investment costs are higher in the trade-expansion scenario. On the other hand, hydropower plants have longer economic lifetimes, so the investment costs can be spread over more years of operation. The difference in annualized capital cost is therefore less than the difference in investment cost. 83 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Moreover, trade expansion leads to large savings in operating costs due to the higher share of hydropower. These savings are often large enough to justify hydropower’s higher up-front investment costs. The total savings in operating costs depend on the fuel prices of thermal power plants. We use the price of crude oil as a signal of fuel prices for thermal power plants since it is closely correlated to the price of natural gas, heavy fuel oil and diesel. In our base case (presented in chapter 5), we used a crude oil price of $46/barrel, based on 2005 prices. Oil prices are highly volatile, however, making them difficult to predict: by May 2008 the price of crude oil had risen far above $46/barrel, before falling below that mark in January 2009. Therefore, we found it important to check the sensitivity of our results to changes in oil prices. In our analyses, we used a crude oil price in the range of $25–$75/barrel, using historical data to estimate the correlation between the price of crude oil and the prices of other oil products and natural gas. Running the least-cost expansion model for the EAPP region with different assumptions about crude oil prices reveals an interesting pattern: • The total investment volume increases with the oil price during both trade expansion and trade stagnation (table 6.1) • The total investment level is higher in the trade-expansion scenario than in the trade-stagnation scenario for all oil price levels • The share of hydropower investments is higher in the trade-expansion scenario; investments in geothermal power also become more profitable with increasing oil prices While higher oil prices make investments in thermal power plants less attractive, they also contribute to the faster replacement of existing gas, oil, and diesel-fueled power plants by making them more expensive to operate. The possibility of trade also encourages the development of hydropower capacity in countries that are rich in hydropower, avoiding expensive thermal power. Eventually, however, countries approach full exploitation of their hydropower resources, diminishing the potential for replacing fossil fuels with hydropower. Further increases in oil prices therefore have a diminishing impact on investments in hydropower. 84 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Under trade expansion, hydropower is the Table 6.1 Investments with different oil price assumptions least-cost choice for expanding generation MW capacity despite higher up-front costs. Total investments Hydropower Therefore, total overnight investment costs are Crude oil price Trade- Trade- Trade- Trade- much more sensitive to oil price increases ($/barrel) expansion stagnation expansion stagnation under trade expansion compared to trade scenario scenario scenario scenario stagnation (columns 2 and 3 of table 6.2). On 25 9,912 8,276 5,646 3,074 35 23,413 17,434 10,492 4,150 the other hand, annualized costs are lower in 46 25,637 17,972 15,235 4,689 the trade-expansion scenario for all prices of 55 25,787 18,304 15,308 5,271 crude oil (columns 4 and 5). Since hydropower 65 26,096 19,324 15,931 5,308 is a less feasible alternative to thermal power 75 26,400 19,324 16,031 5,308 under trade stagnation, increases in fuel costs Source: Authors’ original research. affect electricity production costs much more Table 6.2 Investment cost sensitivity to oil price than under trade expansion. Therefore, the total annualized costs of expansion (which take into $ millions Total overnight Annualized cost of power account both annualized capital costs and the investment system expansion and annual costs of electricity production) also Crude oil cost of expansion operation increase more rapidly under trade stagnation. price Trade- Trade- Trade- Trade- ($/barrel) expansion stagnation expansion stagnation Thus, the savings from electricity trade scenario scenario scenario scenario increase with the price of crude oil. Figure 6.1 25 47,392 46,726 11,454 11,587 plots the differences in overnight investment 35 59,898 53,500 13,296 13,721 cost on the left axis and the annual savings on 46 69,928 54,870 15,004 16,003 the right axis. 55 70,570 56,392 15,957 17,513 65 72,129 57,597 17,101 19,333 75 73,020 57,597 18,201 21,115 Source: Authors’ original research. 85 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure 6.1 Difference in overnight investment costs and annual savings with different oil prices Climate change and less reliable hydropower in EAPP In this scenario we test how climate change—assumed to reduce the reliability of hydropower— influences the optimal (least-cost) investment solution in EAPP. According to some, climate change will lead to more unstable weather patterns, including drought in some regions and more storms, heavy rainfall, and floods in others. It may not be possible to utilize the increased precipitation in some regions for hydropower production. According to the authoritative review by Boko and others (2007), East and West Africa will likely experience decreased water flow, while water flow will probably increase in southern Africa. The possible impact of climate change throughout Africa ranges from a 15 percent decrease to a 5 percent increase in water flows by 2050 compared to the 1961–90 baseline. It is difficult to predict the short-term (before 2015) impact of climate change on water flows. Therefore, we conducted simulations in which climate change reduces firm hydropower production (that is, production in GWh per MW installed) of both existing and new capacity by between 5 percent and 25 percent. 86 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The results of our simulation indicated that Table 6.3 Investments with different assumptions investments in natural-gas-fired power plants in Egypt about hydropower availability and Tanzania replace investments in the less reliable MW hydropower in EAPP (table 6.3). Because natural-gas- Hydro- Hydro- Natural gas- fired power plants produce more energy (in GWh) per power Total power fired availability MW installed, the total investment level is reduced. As Base 25,637 15,235 8,577 the gas resources in Egypt and Tanzania are almost -5% 24,233 11,598 10,809 fully exploited,22 geothermal power in Djibouti -10% 24,636 11,589 11,221 becomes a more profitable option. In rural areas, diesel- -15% 25,121 11,657 11,532 fired power plants become an alternative to minihydro. -20% 25,351 11,548 11,831 Lower capacity investment and a shift away from -25% 24,976 11,933 11,007 Source: Authors’ original research. hydropower reduces total overnight investment costs; as previously explained, hydropower is more capital intensive and requires larger up-front investments. Additionally, the lower investment in hydropower reduces the need for transmission lines, since thermal power plants can be built in the vicinity of load centers. On the other hand, as hydropower resources are increasingly more constrained, substitutes become increasingly expensive. Ultimately, investment costs are reduced, but only slightly. On the other hand, when taking into Table 6.4 Power sector costs with different assumptions about account the operating costs and the shorter hydropower availability economic lifetime of gas-fired power plants, $ millions annualized costs increase steadily as Total overnight investment cost of power Annualized hydropower investments are replaced by sector ($ millions) cost of Hydro- system other technologies. According to our power expansion availability estimates, annualized costs increase 9 Trans- and Total Generation mission operation percent when hydropower availability is Base 69,928 32,667 1,013 15,004 reduced 25 percent (table 6.4). A 9 percent -5% 62,651 25,058 1,346 15,311 increase is substantial, but, in light of -10% 62,910 25,363 1,299 15,566 warnings about the potential damage of -15% 63,615 26,072 1,294 15,824 climate change, it is in our view a -20% 63,793 26,236 1,309 16,094 surprisingly small effect. -25% 62,585 25,176 1,161 16,367 Source: Authors’ original research. Note: Investment cost of subnational distributional network not specified since Clean Development Mechanism it does not respond to hydropower availability. (CDM) in the SAPP More focus on climate change would increase the price of CDM credits and contribute to investments in emission-free technologies such as hydropower and solar PV. The CDM is meant to cover the difference between the polluting, least-cost technology and the cost of the more expensive, clean project under consideration. The CDM difference is divided by emissions 22 The potential for gas-fired power plants is assumed to be the same in all scenarios. 87 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA saved to calculate the cost of certified emission reduction credits (CERs) associated with the project under consideration. The process can also be done in reverse: starting with the price of CERs, the admissible additional cost of a clean project, and subsequently the associated increase in clean technology, can be calculated. By doing this for several CER prices, it is possible to map the increase in clean technologies as a function of the CER price. We do this for the SAPP region, focusing on large-scale hydro. Off-grid technologies such as minihydro and solar PV are not considered. In our simulations for the SAPP region, additional Table 6.5 Investments with different investments in hydropower replace investments in coal- assumptions about CDM prices fired power plants in South Africa (table 6.5). As MW expected, the increased investment costs in hydropower Coal-fired CO2 cost Hydro- exceed the reduced investment cost in the coal-fired power Total power ($/ton) power plants plants (table 6.7). The difference is the macro equivalent Base 33,343 16,134 13,867 of each project’s CDM component. 5 34,159 18,691 12,125 Investments in hydropower increase in the Democratic 10 35,349 9,225 22,781 Republic of Congo, Malawi, Zambia, and Namibia. In 15 35,716 24,131 8,242 other countries with hydropower potential (Mozambique, Source: Authors’ original research. South Africa, and Zimbabwe), hydropower resources were already fully exploited in the base case, although this is only a model assumption. In reality, the CDM could make additional investment in hydropower feasible. In all, the CDM generates 8,000 MW of additional hydropower (producing 42 TWh) at a CER price of $15/ton CO2 (tables 6.5 and 6.6; figure 6.2). We estimate that a CER price of $15 has the Table 6.6 Production change and CO2 savings potential to eliminate 36 million tons of CO2 (table compared to base case (no CDM price) 6.6). Note that greenhouse gas emissions other than Production change (TWh) CO2 (for example, methane from hydropower CO2 cost Hydropower Coal-fired CO2 savings dams) are not calculated. According to the (US$/ton) power plants (M ton) International Energy Agency (2008), CO2 5 16 -15 -14 emissions related to production of heat and power 10 36 -34 -31 in Africa stood at 360 million tons in 2006. In 15 42 -39 -36 other words, at a CER price of $15, CDM projects Source: Authors’ original research. Note: Excluding emissions from hydropower dams. in SAPP alone are capable of reducing Africa’s power sector emissions by 10 percent. 88 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The underlying assumption Table 6.7 Costs with different assumptions about CO2 cost and CDM price of the CDM is that CDM Annualized cost of expansion and operation finance should neutralize the Total ($ millions) higher generation costs of CO2 cost overnight ($/ton) investment Total Trans- cleaner projects. This would cost ($) Generation Total excluding mission CER price imply that system costs do not Base 98,638 4,554 380 18,500 18,500 increase after deducting CDM 5 101,161 4,508 722 19,800 19,100 finance and the price of CERs. 10 104,893 4,715 944 21,100 20,800 The results of the model, 15 106,271 4,802 1,014 22,300 21,700 however, suggest that system Source: Authors’ original research. costs increase substantially even after deducting CDM due to the greater transmission and distribution costs needed for the cleaner hydropower projects. These costs are unaccounted for by the mechanism. There are savings on maintenance, but cost additions are higher than the cost deductions. The model is able to reveal the hidden costs of CDM because it treats generation, transmission,and distribution in an integrated way. There are two interpretations of the increase in expansion and operation costs after deducting CDM. First, Africa should be paid more for its CO2 reductions, and the rest of the world underpays for the benefit of reduced emissions. This is a valid point since our simulations show that Africa does indeed end up with a higher bill after CDM than before. Second, the cost of a given CDM reduction is higher than suggested by generation costs in isolation. Therefore, the given price of CERs should lead to fewer projects and lower CO2 savings. This is also a valid point since each unit of CO2 reduction is more expensive when taking into account all costs than when only considering generation costs. Figure 6.2 Investments with different CO2 costs 89 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Barriers to hydropower development in WAPP: Less hydropower in Guinea In our trade-expansion scenario, cheap hydropower in Guinea supplies much of the power in the WAPP region (except in Nigeria). Yet most of Guinea’s hydropower capacity was undeveloped as of 2005. Therefore, this sensitivity analysis examines a scenario in which Guinea is unable to develop its maximum hydropower potential originally assumed in the model. Which other producers emerge when exports from Guinea are limited? As of 2005, Guinea had 103 MW of hydropower capacity and 108 MW of thermal power capacity. A portion of this will require refurbishment before 2015, so we assume that only 119 MW of Guinea’s capacity will be operational capacity at that time. According to estimates, however, Guinea has economically profitable hydropower potential of about 6,000 MW. Based on our review of all planned hydropower projects in the country, we originally included a maximum potential of 4,287 MW hydropower in the model. Since the assumed unit cost of hydropower in Guinea is among the lowest in the region, all of this potential is exploited in our base case model runs and Guinea becomes a large exporter of power in WAPP. This is a huge amount of hydropower to be developed in one country over such a short time horizon (ten years). Therefore, we tested the effect of constrained hydropower investment in Guinea on development in the WAPP region. In this modified scenario, we include two power plants (Kaléta I and Sambangalou) with a combined capacity of 165 MW that are scheduled to come online in 2008–9. Among the other planned hydropower projects in Guinea, we assume that only Kaleta II (210 MW) can be completed before 2015. In other words, we reduce the maximum amount of hydropower that can be developed in Guinea to 375 MW. In this scenario, new trade patterns emerge in the WAPP region. Guinea becomes a net importer, despite being a significant exporter in the base case. Côte d’Ivoire emerges as the major power exporter, while Ghana increases domestic production considerably to reduce net imports. Mauritania and Sierra Leone also become net exporters. Hydropower investments in Côte d’Ivoire increase by just under 200 MW, but production from existing gas-fired power plants also increases; total production increases 3 TWh. Production in Ghana almost doubles, increasing to 16 TWh, although only 370 MW new capacity comes online. New (gas- fired) power plants and increased production in existing plants supply the additional power. Sierra Leone invests over 500 MW in hydropower and Mauritania 85 MW in gas-fired power plants to export to neighboring countries. Total annualized costs increase by only 3 percent (just above $300 millions). There is, however, a huge trade-off between lower capital costs and higher variable costs: while capital costs are $500 million lower (mainly due to lower generation investments), variable operation costs of production are $850 millions (30 percent) higher. In addition, more of the existing lower-efficiency thermal capacity is put into use. 90 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Unit investment cost of coal-fired power generation in SAPP The unit investment (capital) cost of coal-fired power generation has increased recently. In particular, the unit cost of new coal-fired power plants in South Africa is more likely $1,500/kW than the $1,100/kW assumed in our main scenarios. We perform a sensitivity analysis to test how a higher investment cost of coal-fired power plants in SAPP would influence the results in the trade-expansion scenario; more specifically, the unit investment cost of coal generation is raised from $1,100 to $1,500/kW. The increase applies to all countries, but only South Africa invests in new coal capacity in the original scenario, while South Africa, Mozambique, Namibia and Zimbabwe refurbish their existing coal capacity. We find that with higher coal prices, coal investments in South Africa are reduced by 33 percent, from 13,900 MW to 9,300 MW. Incidentally, the 36 percent increase in cost results in a 33 percent decrease in investment.23 Refurbishment is not affected. Since demand is inflexible to price, the shortfall in investment must be covered elsewhere: hydropower investments in the Democratic Republic of Congo, Zambia, and Malawi increase 6,500 MW. This leads South Africa to depend more heavily on imports from these countries, which increase by 32 TWh. New transmission lines are also needed to transport power from the Democratic Republic of Congo, Zambia, and Malawi to South Africa. The environmental implications of this scenario are positive and similar to an increase in the cost of coal or a decrease in the price of hydropower for CDM-related reasons.24 Financially, the increase of the unit capital cost of coal-fired power plant increases overnight capital costs by $9.5 billion, a 15 percent increase. This increase is perhaps lower than expected, given that in the original scenario 60 percent of new capacity in SAPP is coal-fired, and 60 percent of 36 (the unit cost increase) is 22 percent. Out of the $9.5 billion increase, generation investments account for almost $6 billion. Finally, annualized costs of expansion and operation in SAPP increase from $18.4 to $19.0 billion, an increase of only 3 percent. The reasons include the savings in fuel costs generated by replacing coal with hydropower. Also, annualized costs of expansion and operation include several fixed items such as maintenance of the distribution network, distribution, and connection costs of social demand, which prevent any single parameter from reducing annualized costs by a significant margin. How robust are model results to a parameter such as the unit investment cost of coal-fired power plants? Macro results, in particular the annualized cost of expansion and operation, are robust. But the closer one moves to the area of the model surrounding the parameter, the more vulnerable the results become. In this case, investment in coal-fired power plants in South Africa and imports to South Africa are two vulnerable aspects of the model. 23 There is no supply curve for coal in the model, so this—close to the Cobb-Douglas supply curve—is brought about by the links and constraints of the simultaneous system. 24 In this and similar discussions, it is important to remember that when demand is price insensitive, the level of production is not affected by cost increases (for example, in coal production). But the shares of production covered by coal and other technologies are affected. Also, from a least-cost perspective, it is the sum of annualized capital costs and operating costs that matters. 91 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Unit investment cost of hydropower in WAPP Hydropower development costs are project-specific but are still influenced by the world market conditions. Construction costs have risen lately and there is always the risk of cost overruns. We have tested to what extent the investment and trade pattern in the WAPP region depends on the costs of hydropower development. We have chosen to increase the unit cost of hydropower in all countries by 50 percent. In this way, the model retains the relationship between hydropower development costs in the different countries, but hydropower becomes relatively more expensive than thermal power compared to in the base case scenarios. With hydropower relatively more expensive, net exports from Guinea, the large hydropower supplier in the WAPP region, fall from 17.4 TWh to 5.4 TWh. Exports to Guinea-Bissau and Sierra Leone remain almost the same, but exports to Mali fall from about 13 TWh to 1.2 TWh. While in the base case, power flowed through Mali to the other WAPP countries (via Côte d’Ivoire to Ghana and farther), these power flows now cease completely and Guinea only supplies its closest neighbors. In this scenario, it is no longer profitable to invest in hydropower and the associated transmission lines while thermal power can be built closer to consumption centers. Just 1,600 MW of hydropower investments are carried out in Guinea (compared to 4,300 MW in the base case), in addition to some 65 MW of refurbishments of HFO- fueled power plants. Instead of relying on Guinea, countries in the eastern part of the WAPP region—Nigeria, Côte d’Ivoire, and Togo—increase their exports, and Ghana develops more of its domestic resources. Due to the relatively higher cost of hydropower, gas-fired power plants completely replace hydropower investments in Côte d’Ivoire, Ghana, and Togo. In addition, more gas-fired power plants are refurbished in Nigeria and Côte d’Ivoire. In total, hydropower capacity is 5,200 MW lower and gas-fired capacity 1,800 MW higher than in the base case.25 The capacity share of hydropower falls from 77 percent to 66 percent. This switch of investments increases annualized costs by 10 percent ($1.2 billion). The savings related to lower investments (in generation and transmission lines) are only 4 percent (less than $300 million), while the variable costs of generation increase by a significant 55 percent ($1.5 billion). While a 10 percent increase in annualized costs is fairly high, it may be lower than expected given the 50 percent unit cost increase in 77 percent of the system. More imports from the Democratic Republic of Congo to CAPP The Central African region (the Republic of Congo and Central African Republic) borders the Democratic Republic of Congo and could thus be expected to benefit from hydropower development 25 As thermal power plants can produce more kWh per MW installed, less capacity is required to produce the same amount of power. But with lower power flows, power losses are also smaller. Therefore, total production is slightly (2 percent) lower in the sensitivity with higher hydropower costs. 92 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA there, especially from the Inga project, which is located relatively close to Brazzaville. In this study, the Democratic Republic of Congo is part of SAPP and a large hydropower development will indeed benefit the SAPP region, with exports of 52 TWh. This sensitivity analysis illustrates how imports from the Democratic Republic of Congo might influence the power balance in the CAPP region. Since demand in the CAPP region is relatively low compared to that of SAPP, the Democratic Republic of Congo could redirect some exports to the north (to CAPP) while exporting to the south. (Additional investments are of course possible, but this is outside the scope of this sensitivity analysis.) Note, however, that we cannot test whether importing from the Democratic Republic of Congo is more profitable than developing the domestic capacity in CAPP countries. We tested a scenario in which the Republic of Congo can import 5 TWh (equal net imports from other CAPP countries to the Republic of Congo in the base case) from the Democratic Republic of Congo.26 Imports of 5 TWh to the CAPP region reduce hydropower investments in both Cameroon (550 MW less, due to less exports to the Republic of Congo) and the Republic of Congo itself (300 MW less). Imports from Cameroon completely replace hydropower investments in Gabon. Even though imports from Democratic Republic of Congo to the Republic of Congo replace imports from Cameroon to some extent, they do not eliminate them: Congo still imports 1.3 TWh from Cameroon. Instead, the new imports replace production and investments in new capacity in the Republic of Congo. Allowing the Democratic Republic of Congo to export to the CAPP provides overall annual savings of $220 million. Of this total, $180 million arises from reduced investment in generation capacity in the CAPP region, while the rest is due to lower operating costs and less cross-border transmission investments. Note, however, that we do not account for costs incurred in the Democratic Republic of Congo. 26 Note that in the base case, current imports (0.4 TWh annually) from the Democratic Republic of Congo to the Republic of Congo were included. These imports are included in the 5 TWh of the sensitivity. 93 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA References This section lists references cited in the main document. Additional references to official websites and master plans made in the annexes are not included here. Arthur Energy Advisors. 2007. “WAPP Emergency Power Supply Plan to Curtail Power Shortages Faced by ECOWAS Member States.” Draft Study Report. August 2007. Bergman, S., and I. Davies. 2005. “Mainstreaming Low-Cost Innovations in Electricity Distribution Networks.” Study prepared for Energy Sector Management Assistance Program, World Bank, Washington, DC. Bogetic, Z., and J. W. Fedderke. 2005. “Forecasting Investment Needs in South Africa’s Electricity and Communications Sector.” World Bank Policy Research Paper 3829, Washington, DC. Boko, M., I. Niang, A. Nyong, C. Vogel, A. Githeko, M. Medany, B. Osman-Elasha, R. Tabo, and P. Yanda. 2007. “Africa.” In M. L. Parry, O. F. Canziani, J. P. Palutikof, P. J. van der Linden and C. E. Hanson (eds.), Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (433–467). Cambridge: Cambridge University Press. http://www.ipcc.ch/pdf/assessment- report/ar4/wg2/ar4-wg2-chapter9.pdf. CORE. 2003. “Issues and Options for Rural Electrification in SAPP Member Countries and Rural Electrification Planning in Lesotho.” Unpublished desk study. DECON/SWECO. 2005. “Tanzania Rural Electrification Study.” Econ Pöyry. 2005. “Reform and Development of the Angolan Electricity Sector.” Discussion document prepared for the Angolan Ministry of Energy and Water. ———. 2006. “Tanesco Pre-Feasibility Grid Project.” Report prepared for Tanzanian Ministry of Energy and Mines. ECOWAS. 2006. “White Paper for a regional policy geared towards increasing access to energy services for rural and periurban populations in order to achieve the Millennium Development Goals.” ECOWAS/UEMOA. EIA. 2006. Annual Energy Outlook. http://www.eia.doe.gov/ Fay, M. 2001. “Financing the Future: Infrastructure Needs in Latin America 2000–05.” Unpublished paper, World Bank, Washington, DC. Fay, M., and T. Yepes. 2003. “Investing in Infrastructure: What is Needed from 2000 to 2010?” World Bank Policy Research Working Paper 3102, Washington, DC. Gaunt, C. T. 2005. “Meeting Electrification’s Social Objectives in South Africa, and Implications for Developing Countries.” Energy Policy 33: 1309–1317. 94 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA International Energy Agency. 2005. “Projected Costs of Generating Electricity.” Paris. ———. 2008. “CO2 emissions from fuel combustion, 2008 Edition.” Paris. Karezeki, S., and L. Majoro. 2002. “Improving Modern Energy Services for Africa’s Urban Poor?” Energy Policy 30: 1015–1028. Kongsamut, P., S. Rebelo, and D. Xie. 1997. “Beyond Balanced Growth.” NBER Working Paper W6159. National Bureau of Economic Research (NBER), Cambridge, MA. Meschies. 2005. “International Fuel Prices 2005.” 4th edition. Report commissioned by Federal Commission for Economic Cooperation and Development, Germany. Modi, V. 2004. “Energy and Transport for the Poor.” Paper commissioned for the Millennium Project Task Force 1. Nexant. 2007. “SAPP Regional Generation and Transmission Expansion Plan Study.” October. Royal Academy of Engineering. 2004. The Costs of Generating Electricity. London. Sanghvi, A. 2005. “Rural Electrification: Costs of Connection, Willingness and Ability to Pay.” Presentation at Consultative Workshop on Public Private Partnership in Rural Electrification. World Bank. 2006. “An Investment Framework for Clean Energy and Development.” UNEP. 2002. Widening access in the context of power sector reform – an overview of the institutional challenges in Africa. Power sector reform and sustainable development. Brainstorming meeting of UNEP and IEA, May 21-22, 2002, Paris. USAID (U.S. Agency for International Development). 2004. “Innovative Approaches to Slum Electrification.” Washington, DC. Zomers, A. 2001. “Rural Electrification.” PhD dissertation, University of Twente. 95 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Appendix 1 Documentation of asset stock and refurbishment needs of thermal power generation This appendix gives an overview over the existing asset stock and refurbishment needs of thermal power generation in Sub-Saharan Africa. These assumptions are used in the Least Cost Expansion Model. Refurbishment refers to major refurbishments, to prolong the life of an outdated plant whose operating life is coming to an end by restoring it to full operational status, or to repair generating assets that have been seriously damaged due to for example war. We do not include costs for ordinary maintenance and repair work. We will here describe the assets that need to be refurbished. The actual amount of refurbishment will be determined as an endogenous variable in the Least Cost Expansion Model. However, due to the lower costs per MW of refurbishment (compared to new investments), all assets in need of refurbishment are usually refurbished. CAPP Region Cameroon 98.2 MW of total capacity to be refurbished before 2015. Table A2.1 Thermal Generation in Cameroon (2005) Plant Type Commissioning date Installed capacity (MW) Available capacity (MW) Limbé Diesel 2004 85 85 Bassa Diesel 2002 24.6 9.6 Logbaba Diesel 2002 17.6 17.6 Mefoue Diesel 1980 13 13 Oyomabang (old) Diesel 1980 19.6 19.6 Oyomabang (new) Diesel 2002 16 16 Bafoussam (old) Diesel 1979 12 12 Bafoussam (new) Diesel 2002 4.7 4.7 Garoua Diesel 1980 19 19 Isolated mini generators Diesel 1980 10 10 Total Cameroon 221.5 206.5 96 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Central African Republic The entire capacity of 24 MW is to be refurbished before 2015. Table A2.2 Thermal Generation in Central African Republic (2005) Plant Type Commissioning date Installed capacity (MW) Available capacity (MW) Bangui Diesel 1980 18.2 18.2 Many small ones Diesel 1983 5.8 5.8 Total CAR 24 24 Chad The entire capacity of 29 MW is to be refurbished before 2015. Table A2.3 Thermal Generation in Chad (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) N'Djamena Diesel 1980 22 22 Other Diesel 1980 7 7 Total Chad 29 29 Republic of Congo The entire capacity of 28.4 MW is to be refurbished before 2015. Table A2.4 Thermal Generation in Republic of Congo (2005) Plant Type Commissioning date Installed capacity (MW) Available capacity (MW) Pointe-Noire Diesel 1980 19 19 Brazzaville Diesel 1980 9.4 9.4 Total Republic of Congo 28.4 28.4 Equatorial Guinea Out of the total capacity 34 MW is to be refurbished before 2015. Table A2.5 Thermal Generation in Equatorial Guinea (2005) Plant Type Commissioning date Installed capacity (MW) Available capacity (MW) Punta Europa Gas 1999 10.4 10.4 In Malabo (name unknown) Diesel 2000 30 30 Name unknown Diesel 1980 4 4 Total Equatorial Guinea 44.4 44.4 97 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Gabon The entire capacity of 244 MW is to be refurbished before 2015. Table A2.6 Thermal Generation in Gabon (2005) Plant Type Commissioning date Installed capacity (MW) Available capacity (MW) Owendo Diesel 1980 33 33 Name not known Diesel 1995 203 203 Other Diesel 1980 8 8 Total Gabon 244 244 EAPP – Nile basin region Burundi No thermal refurbishment is expected in Burundi. Almost all electricity generated in Burundi is hydropower provided either by the national Regie de Production et Distribution d'Eau et d'Electricité (Regideso) or the three nation cooperation Société Internationale des Pays des Grands Lacs that operates hydropower generation with supply to Burundi, Rwanda and DRC. There is also the Direction Generale de l’Hydraulique et des Energies Rurales (DHER) that independently develops rural electrification projects. There is installed back-up capacity of 5.5 MW diesel generation in Bujumbura. These diesel generator sets are mostly from 1996, and we assume that they are still operational in 2015. Djibouti No thermal refurbishment is expected in Djibouti. The publicly owned Electricité du Djibouti, EdD, is responsible for power generation, transmission and distribution in Djibouti. All generation capacity (85 MW) is thermal, mainly oil-fired. The World Bank has provided grants and loan for the establishment of a transmission interconnection from Ethiopia to Djibouti in 2009, so that future hydropower from Ethiopia can replace Djibouti’s expensive oil-fueled generation. It is further planned a geothermal capacity in Assal of 30 MW, and Djibouti is preparing a BOO scheme for this. The government has also sourced financing from the Arabic countries to replace aging rural diesel generation with renewable energy (photovoltaic and wind) for water pumps. The residual electricity from these installations will be used for limited rural electrification. Egypt Egypt’s thermal generating capacity, under the supervision of Egyptian Electricity Holding Company (EEHC) is divided into four regions: Cairo, East Delta, West Delta and Upper Egypt. While much of the thermal generating capacity is old, Egypt has recently completed an ambitious program enabling a large majority of old oil-fired power plants to run on natural gas. This reduces the refurbishment requirements of these older units. We will go through each region below, highlighting refurbishment requirements. The total installed thermal capacity is 16,475 MW, of which 2,093 MW must be refurbished before 2015. 98 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table A2.7 Cairo Region: Installed thermal capacity Plant Fuel Commissioning date Installed capacity (MW) Shoubra El-Kheima Nat. gas/HFO 1984–1988 1260 Shoubra El-Kheima Nat. gas/HFO 1986 35 Cairo West Nat. gas/HFO 1966–1979 350 Cairo West Extension Nat. gas/HFO 1995 660 Cairo South I** Nat. gas/HFO 1957–1989 570 (700) Cairo South II Nat. gas 1995 165 Cairo North Nat. gas/HFO 2004 750 Wadi Hof Nat. gas/HFO 1985 100 Tebbin* Nat. gas/HFO 1979 46 Tebbin* HFO 1958–1959 45 * To be replaced by a new 750 MW plant. Funding secured from World Bank. ** To be replaced by 2*350 MW. For the Cairo region, all thermal plants except parts of the aging Tebbin plant now run on natural gas. The Tebbin plants are, however, to be replaced by a new plant at the same site, therefore we will remove it from the 2015 available capacity. The largest plant is Shoubra El- Kheima, which has been in operation for 20 years. However, Siemens recently won a major contract to rehabilitate and upgrade the 1,260 MW plant and therefore we estimate that it will not need any further refurbishment. Three remaining power stations were commissioned prior to 1995 and could be in need of refurbishment: the Cairo West station (4*87.5 MW), the Cairo South I (3*110 MW + 4*60 MW), the Wadi Hof (3*33 MW). The Cairo West station will be partly refurbished by Alstrom but we estimate that half of the capacity in that plant would need refurbishment prior to 2015. Alstrom also recently announced (January 2007) that they will supply 2*350 MW gas turbines to the Cairo South power plant, raising the capacity of that plant. It would be natural to assume that the aging units in that plant would be retired and we will therefore count the plant as having a 700 MW capacity in 2015 but not needing refurbishment. Also the Wadi Hof Station is to be partly refurbished by Alstrom, who won a contract in 2006 to rehabilitate one of the three 33 MW gas turbines that are installed. The station is old and with a load factor of 10% in 2005 the plant can be said to be a peak load plant. We still assume that 66 MW will need refurbishment prior to 2015. Table A2.8 Upper Egypt Region: Installed thermal capacity Plant Fuel Commissioning date Installed capacity (MW) Walidia HFO 1992–1997 600 Kuriemat Nat. gas/HFO 1998–1999 1254 Assiut HFO 1966–1967 90 In Upper Egypt two of the plants, Walidia and Kuriemat, are relatively modern and according to information in the EECH Annual Report for 2004/2005, they seem to be working well (high 99 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA availability factor, 85% or higher). The aging Assiut plant had a major refurbishment in 1996. Therefore we envision no refurbishment requirements for this region and the full installed capacity will be available in 2015. Table A2.9 East Delta Region: Installed thermal capacity Plant Fuel Commissioning date Installed capacity (MW) Damietta Nat. gas/LFO 1989–1993 1125 Talkha* Nat. gas/LFO 1979–1989 284 Talkha* 210 Nat. gas/LFO 1993–1995 410 Ataka Nat. gas/LFO 1985–1987 900 Abu Sultan Nat. gas/LFO 1983–1986 600 Shabab Nat. gas/LFO 1982 100 Port Said Nat. gas/LFO 1977–1984 64 Arish HFO 2000 66 Oyon Mousa Nat. gas/LFO 2000 640 Suez Gulf (BOOT) Nat. gas/HFO 2002 682.5 Port Said East (BOOT) Nat. gas/HFO 2003 682.5 * Replaced by a new 750 MW plant at the existing site Note: Excluded are the old power plants in Sharm El Sheik and Hurghada which are not in use. In the East Delta Region, the plants Arish, Oyon Mousa, Suez Gulf and Port Said East (the last two are IPPs) are relatively new and are not estimated to need refurbishment. The Talkha plant is to be decommissioned and replaced by a new 750 MW plant at the existing site. We will therefore remove the current plant from 2015 available capacity, but add 750 MW of capacity instead and assume no further refurbishment requirements of this plant until 2015. The Ataka and Abu Sultan plants have received refurbishment during the past few years (2002 and 2004 respectively). The contract for rehabilitating the Abu Sultan power station went to Honeywell while Emerson landed the Ataka project. They are not assumed to need further refurbishment. Finally, the aging Shabab and Port Said plants with a total capacity of 164 MW we assume need refurbishment prior to 2015. Table A2.10 West Delta Region: Installed thermal capacity Plant Fuel Commissioning date Installed capacity (MW) Kafr El-Dawar Nat. gas/HFO 1980–1986 440 Mahmoudia I Nat. gas/LFO 1981–1982 135 Mahmoudia II Nat. gas/HFO 1983, 1995 308 Damanhour 300 Nat. gas/HFO 1991 300 Damanhour Extension Nat. gas/HFO 1968–1969 195 Damanhour (old)* Nat. gas/LFO 1960 30 Damanhour CC 153 El-Sieuf Nat. gas/LFO 1985–1995 200 El-Sieuf HFO 1981–1984 113 Karmouz LFO 1980 25 Abu Kir Nat. gas/HFO 1983–1991 900 100 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Abu Kir* LFO 1983 25 Sidi Krir 1&2 Nat. gas/HFO 1999–2000 640 Nubari Nat. gas/LFO 2005 500 Matrouh Nat. gas/HFO 1990 60 Sidi Krir 3 & 4 (BOOT) Nat. gas/HFO 2002 682.5 * Retired The West region has several old thermal plants with poor operating performance (Egyptian Electricity Holding Company Annual Report 2004/2005). While the Mahmoudia II, Damanhour CC, Abu Kir, Sidi Krir and Nubari plants are either new or refurbished, Kafr El-Dawar, Mahmoudia, Damanhour Extension and El-Sieuf power plants are assumed to need refurbishment prior to 2015, based on age and low availability factors and/or efficiency factors. Total refurbishment requirements are thus 1,105 MW. The Abu Kir LFO plant has been retired, and the Matrouh plant was not in operation in 2005. These plants will be taken out of 2015 generating capacity, as will the old oil-fired Karmouz plant. In addition Egypt has over 50 isolated thermal power plants with an Installed capacity of 280 MW. The majority is older diesel units and although some of the isolated plants were strengthened in 2003/2004, we still estimate that many of these isolated plants would need to be replaced before 2015. A conservative estimate would be 100 MW. Ethiopia The bulk of Ethiopia’s generating capacity is hydro, but there are also about 43.3 MW of thermal (diesel) capacity spread out over 12 plants and one 7.3 MW geothermal plant connected to the grid. In addition, there are 25.1 MW of stand-alone diesels. The Ethiopian Electric Light and Power Authority (EELPA) official homepage acknowledges that due to aging of plants the available capacity is lower than the installed capacity (approximately 84%). The table below presents a list of the thermal units in Ethiopia. The installed thermal capacity is 76 MW, but due to aging of the plants the available capacity is lower. We estimate that the newer Aluto Langano, Adwa, Awash 7 Kilo and Kaliti plants will be operational in 2015, as well as 10 MW of the stand-alone diesel plants. Available capacity in 2015 is thus 51 MW. 101 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table A2.11 Thermal power plants in Ethiopia Plant Type Commissioning Date Installed capacity MW Aluto Langano Geothermal 1999 7.3 Alemaya Diesel 1958 2.3 Dire dawa Diesel 1965 0.3 Adigrat Diesel 1992-95 1.1 Axum Diesel 1972,92 0.55 Adwa Diesel 1998 3.0 Mekele Diesel 1984,91,93 1.3 Shire Diesel 1975,91,95 0.8 Jimma Diesel - 0.1 Nekempt Diesel 1984 1.7 Awash 7 Kilo Diesel 2004 22.4 Kaliti Diesel 2004 9.0 Ghimbi Diesel 1962,84 0.3 Stand Alone Units Diesel 1967-1998 25.1 Total 75.7 Source: Ethiopian Electric Light and Power Authority (EELPA) official homepage Kenya Kenya has 10 thermal power plants with a total installed thermal capacity of 520 MW and an operational capacity of approximately 470 MW. Below we present more detailed information on the plants:27 • Kipevu (steam), the currently installed units with capacity of 63 MW were installed in 1972 and 1976, the actual capacity is 26 MW. • Kipevu (diesel), commissioned in 1999. Installed capacity 73 MW, effective capacity 70 MW. • Kipevu (gas), two gas turbines commissioned in 1987 and 1999. Installed capacity 63 MW, effective capacity 60 MW. • Nairobi South (gas), commissioned in 1972. Installed capacity 13.5 MW, effective capacity 10 MW. • Iberafrica (diesel), IPP commissioned in 1997. Installed and effective capacity 57 MW. • Westmount (gas), IPP commissioned in 1997. Installed capacity 46 MW; effective capacity 43 MW. • Tsavo (diesel), IPP commissioned in 2001. Installed capacity 75 MW, effective capacity 74 MW. 27 SNC Lavalin (2005): “Strategic/Sectoral, Social and Environmental Assessment of Power Development Options in the Nile Equatorial Lakes Region”; Kengen official home page; Eberhard, A. and K. Gratwick (2005): “The Kenyan IPP Experience”, Working paper University of Cape Town 102 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA • Olkaria I (geothermal), 3*15 MW commissioned in 1981, 1982 and 1985. Installed and effective capacity 45 MW. • Olkaria II (geothermal), commissioned in 2000-2001. 70 MW installed and effective capacity. • Olkaria III / OrPower4 (geothermal), commissioned in 2001. Installed capacity 13 MW, effective capacity 12 MW. The IPPs are new and modern and do not need refurbishment. Also the geothermal plants are according to the above-mentioned sources fully functioning. We estimate that the aging Kipevu steam plant and Nairobi South gas-plant needs refurbishment prior to 2015. Available capacity is then 434 MW and refurbishment requirements are 76.5 MW. Rwanda Rwanda has only two thermal stations with combined installed capacity of approximately 20 MW.28 Both run on diesel. • Gatsata (diesel), originally three old 850 kW units with one 4.77 MW diesel unit added in 2004. • Kigali (diesel), 12.2 MW of which 8 MW was installed in 2004 by Lahmeyer. The Kigali plants have been extended with new units as a part of a 2004 Government Fast Track Emergency Project. We remove the three aging diesel units in Gatsata from generating capacity so available thermal capacity in 2015 is approximately 17 MW. Sudan Sudan’s National Electricity Corporation (NEC) operates 15 grid-connected thermal plants and 14 stand alone diesel systems. In addition there is a newer IPP plant commissioned in 2004. Below we provide information on the power plants. Table A2.12 Thermal generation in Sudan Plant Type Commissioning Date Installed capacity (MW) Available capacity (MW) Grid Connected: Sherief Steam 1 Steam 1984 60 53 Sherief Steam 2 Steam 1994 120 103 Atbara Diesel 1985 17.2 10.4 Girba Diesel 1984 7 4 Kassala 1 Diesel 1982 4 2 Kassala 2 Diesel 1990 3.5 2 Kassala 3 Diesel 1990 4.8 3.9 Fao Diesel 2004 12.6 10 Sherief Gas 1 Gas 1986 20 14 Sherief Gas 2 Gas 1986 50 30 Kuku 1 Gas 1986 15 8 28 SNC Lavalin (2005): “Strategic/Sectoral, Social and Environmental Assessment of Power Development Options in the Nile Equatorial Lakes Region”; Lahmeyer homepage. 103 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Plant Type Commissioning Date Installed capacity (MW) Available capacity (MW) Kuku 2 Gas 1986 12 10 Garri 1 Combined Cycle 2003 210 172 Garri 2 Combined Cycle 2003 120 84 El Jaili Combined Cycle 2004 73 73* Dit Kilo X (IPP) Diesel 2004 257 257* Total Grid Connected 986 836 Stand Alone: Port Sudan Diesel 1983, 1997, 2003 59 37 Wadi Halfa Diesel 2004 1.8 1.5 Dongola Diesel 1981 10.2 6.6 Karema Diesel 2004 12.7 12 Elobied Diesel 1987 3.2 2.9 Umrawaba Diesel 1994 9.4 6.6 Nyala Diesel 1985 11.24 7.05 ElFasher Diesel 1983 3.12 2.24 ElGlnina Diesel 1989 2 1.4 Aldaeen Diesel 2004 2 1.6 Kadogly Diesel 2004 2 1.6 Elnehood Diesel 2004 2 1.6 Juba Diesel 1986 5 1.6 Malakal Diesel 2002 2.6 1.6 Wau Diesel 1983 1.6 1.4 Total Stand Alone - - 142 94 Data missing, the new 2004 plants are assumed not to need refurbishment and full Installed capacity is available Source: NEC official home page as of October 2006, Mbendi The installed thermal capacity is thus 1,128.5 MW, of which 931 MW is available. Refurbishment requirements of natural gas and HFO plants amount to 133 MW. It is assumed that it is not feasible to refurbish the diesel units (65 MW); these can rather be replaced by new units. Tanzania Although the bulk of Tanzania’s installed capacity is hydropower, the generating mix also includes gas and diesel fired thermal plants. The thermal plants connected to the system have an installed capacity of approximately 300 MW according to the East Africa Power Master Plan. The installed capacity can be divided into the newer IPP plants that are driven by natural gas from the Songo-Songo gas to electricity project and the diesel generators owned and operated by Tanesco: • The Ubungo gas-fired plant with 120 (180/200) MW installed capacity (IPP) • The Tegeta gas-fired plant with 100 MW installed capacity (IPP) • Older grid connected diesel plants of 82 MW installed capacity • Stand alone diesel units of 30 MW installed capacity 104 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The Ubungo and Tegeta gas-fired plants are modern plants and have recently been refurbished when they were converted from diesel-fired to gas-fired plants. Therefore we do not consider them to be subject to major refurbishments before 2015. The older diesel generators operated by Tanesco have effective capacity supply only half of their installed capacity, due to age, lack of spare parts and other problems, according to the information provided on Tanesco’s official homepage. However, with the Songo Songo gas to electricity project, the objective is to retire the aging diesel fleet and use capacity at Tegeta and Ubungo instead29; therefore, we exclude the older diesel plants from current generating capacity in the Least Cost Expansion Model. Uganda Just a few years ago, Uganda relied almost exclusively on hydro-power, except the 10 MW thermal (bagasse cogeneration) plant in Kakari. Due to low levels of water in lake Victoria and resulting acute power crises, Uganda leases two 50 MW diesel plants from Aggreko, with plans of additional 50 MW.30 The Kakari plant currently sells 6 MW to the grid and uses 4 MW for internal consumption. Installed thermal capacity is then 106 MW. The newly installed Aggreko plants are estimated not to need refurbishment prior to 2015. SAPP region Angola Although hydropower is the main contributor in the supply of electricity in Angola, there is also thermal power generation capacity installed in the country. The thermal power situation is summarized below (the figures do not include the 82 MW new gas turbine in Benguela). Table A2.13 Thermal generation assets in Angola System Type Installed capacity Available capacity Available capacity (Main and Isolated) (MW) (MW) (%) Main – North Gas turbines 93.2 24.4 26 Diesel plants 25.5 16.3 64 Main – Central Gas turbines 33.8 0 0 Diesel plants 38.9 22.3 57 Main – South Gas turbines - - - Diesel plants 16.5 12.9 78 Isolated Systems Gas turbines 10.4 10.4 100 Diesel plants 19.8 4.7 24 Total 238.1 91 38 Source: Power Sector Master Plan (2003) and cross-reference of Alfstad (2005) and EIA (2003) 29 East Africa Master Plan (2005) 30 First-Hand View of Africa’s Power Crises, World Bank, November 2, 2006 105 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The installed thermal capacity, 238 MW, described in the Power Sector Master Plan (2003) is well in line with the figures in Alfstad (2005)31 and EIA (2003) which report 222 MW and 235 MW respectively. From the table we can see that a mere 38% of the installed thermal generating capacity is operational. We assume that it is not profitable to refurbish the broken-down diesel units. It is assumed that the 103 MW gas turbines can potentially be refurbished. Botswana Botswana Power Corporation (BPC) has installed capacity of 132 MW (4*33 MW) at the coal- fired Morupole Power Station, commissioned in 1985. In addition, there are approximately 20 MW of stand alone generating units, mostly diesel and solar. One example is the Ghanzi diesel power plant, run as an IPP, from which BPC obtains approximately 1-2% of electricity generation. Total refurbishment requirement before 2015 is 12 MW. Malawi About 98% of the installed capacity in Malawi’s generation system is hydropower. Escom, the national utility responsible for generation, transmission and distribution, operates three diesel units:32 • Lilongwe, 4.3 MW (1.3 MW and 3.0 MW diesel units), commissioned in 1972 • Mzuzu, 1.1 MW diesel unit, commissioned in 1980 • Chichri, 16 MW gas-fired back-up generator The 2000 Escom annual report stated that the Chichri gas-fired generator was closed down for maintenance and is under repair by a UK firm. We assume that the ChiChri gas turbine is available without additional refurbishment also in 2015., while the diesel units are not refurbished. Mozambique Mozambique relies mainly on hydropower for electricity. The state-owned utility company EdM (Electricidade de Mozambique) has installed capacity of 204 MW of thermal capacity in 14 plants, but not one single plant was fully operational in 2005:33 • Angoche, 1.6 MW, diesel (available capacity 0.4 MW) • Beira, 12 MW, gas (available capacity 0 MW) • Inhambambe, 4.3 MW, diesel (available capacity 2.15 MW) • Lichigonga, 1.3 MW, diesel (available capacity 0.9 MW) • Lionde, 2.4 MW, diesel (available capacity 1.8 MW) 31 T. Alfstad (2005): “Development of a least cost energy supply model for the SADC region”, Master Thesis at University of Cape Town 32 http://www.escommw.com/aboutESCOM/factfig.asp 33 T. Alfstad (2005): “Development of a least cost energy supply model for the SADC region”, Master Thesis at University of Cape Town; EdM official web-site (http://www.edm.co.mz/) 106 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA • Maputo, 111 MW, coal/gas (available capacity 62 MW) • Tete, 0.4 MW, diesel (available capacity 0.3 MW) • Mocuba, 0.8 MW, diesel (available capacity 0.4 MW) • Nacala, 21 MW, diesel (available capacity 3.5 MW) • Nampula, 6.5 MW, (available capacity 5.2 MW) • Pemba, 8 MW (available capacity 6.7 MW) • Quelimane, 6.9 MW (available capacity 6.4 MW) • Xai-Xai, 2.6 MW (available capacity 1.4 MW) In 2005, the available thermal power generating capacity was thus only 50% of the installed capacity, pointing to huge needs for refurbishment of thermal power plants in the war-torn country. We assume that the coal and gas-fired power plants will either be operational or considered for refurbishment (80 MW) ten years from now, while the diesel plants will not be refurbished. Namibia Namibia’s thermal power generation resources amount to approximately 140 MW, summarized below.34 • Van Eck coal-fired power plant in Windhoek, 4*30 MW units, unit 1 and 2 commissioned in 1972, unit 3 commissioned in 1973 and unit 4 in 1979. Current operating peak load capacity is 76 MW. • The Paratus diesel-fired stand-by plant in Walvis Bay, 24 MW, commissioned in 1976. Current operating peak load capacity is 18 MW. The two thermal plants are high-cost options. They are currently in operation, mainly due the constrained supply situation in South Africa. However, the future supply plans of Namibia mainly concern gas-fired power plants and hydropower plants. The main projects are: • Kudu Gas-to-Electricity project, which plans to use gas from the Kudu gas fields of the coast of southern Namibia to fuel an 800 MW gas-fired power plant, • Two hydropower projects Epupa and Popa Falls. It is likely that existing thermal plants will be replaced by the newer power sources. According to the SAPP 2005 Statistics Information from the SAPP official home page, it is expected that the Kudu project will be commissioned in 2009 and the two hydropower projects in 2015. Since these dates are associated with large uncertainty, we will include the 120 MW of the aging Van Eck coal plant as potential refurbishment in the Least Cost Expansion Model. We remove the Paratus power plant due to its very high generating cost from the generating capacity in 2015 in the model. 34 Nampower generation information from official web-site (www.nampower.com.na) 107 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA South Africa South Africa is the country with the largest thermal power generation capacity in the region. About 90% of generation capacity is coal-fired, the rest is nuclear (Koeberg, 1,800 MW), gas turbines and hydropower/pumped storage (2,240 MW in total). ESKOM’s power plants ESKOM is by far the largest producer in the country. Let us start with an overview over ESKOM’s power plants.35 Gas turbines • Acacia Power Station: Gas turbine, 3*57 MW. Operational in 1976. • Port Rex Power Station: Gas turbine, 3*57 MW. Operational in 1976. Coal-fired power plants • Arnot Power Station: Coal-fired, 6*350 MW units, to be expanded to 6*380 MW. Fully operational in 1975. • Duvha Power Station: Coal-fired, 6*600 MW units. Fully operational in 1984 • Hendrina Power Station: Coal-fired, 10*200 MW units. Fully operational in 1976. • Kendal Power Station: Coal-fired, 6*868 MW units (3,840 MW net maximum capacity). Fully operational in 1993. • Kriel Power Station: Coal-fired, 6*500 MW units. Fully operational in 1979. • Lethabo Power Station: Coal-fired, 6*618 MW units. Fully operational in 1990. • Majuba Power Station: Coal-fired, 3*665 MW and 3*716 MW units. Fully operational in 2001. • Matimba Power Station: Coal-fired, 6*665 MW units. Fully operational in 1988. • Matla Power Station: Coal-fired, 6*600 MW units. Fully operational in 1983. • Tutuka Power Station: Coal-fired, 6*609 MW units. Fully operational in 1993. In addition, the following mothballed power plants are in the process of being recommissioned: • Camden Power Station: Coal-fired, 1,520 MW. • Grootvlei Power Station: Coal-fired, 1,128 MW. • Komati Power Station: Coal-fired, 909 MW. Most power stations are old, commissioned in the late 1960-ies to early 1980-ies, implying that by 2015 they are between 35 and 55 years old. Since coal plants have an estimated lifetime ranging from 30–40 years,36 this implies that most of the power plants need refurbishment. Some plants are 35 Net maximum capacity is usually lower than nominal capacity of a power plant. The overview here indicates the nominal capacity of the power plants, except for power plants where the net maximum capacity is significantly lower than the nominal capacity. 36 The International Energy Agency assumes a 40 year economic life-time of coal-fired power plants in the Projected Costs of Generating Electricity Report (2005). 108 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA newer and some have been refurbished already. Plants that became operational 1990 or later are assumed not to need major refurbishment, while all plants commissioned before 1990 are assumed to need refurbishment, except those that have been refurbished. This includes five of the ten units at the Hendrina power plant that were refurbished in 1996–1997, and one of the units at the Duvha power plant which was refurbished after a mechanical breakdown in 2003. We estimate that remaining ESKOM power stations will need refurbishment before 2015, even though they are currently in full operation. IPPs and municipal power plants In addition to Eskom’s generation capacity, there are also some municipal power stations and private power stations. With the exception of the Kelvin Power Station that was sold to an IPP in November 2001, the private power stations produce electricity exclusively for individual manufacturing companies. It is reasonable to assume that these companies keep their power plants in working order; hence, refurbishment needs of these plants will not be considered in the model. The municipal power stations are listed below, with installed (design) capacity. Note that only Rooiwal, Pretoria West and Roggebaai are currently operational. All of them are estimated to need refurbishment before 2015. • Rooiwal: Coal-fired, 300 MW (206 MW net maximum capacity) • Pretoria West: Coal-fired, 170 MW (100 MW net maximum capacity) • Athlone: Coal-fired, 180 MW (currently not operational) • Kroonstad: Coal-fired, 30 MW (currently not operational) • Swartkopf: Coal-fired, 240 MW (currently not operational) • Bloemfontein: Coal-fired, 103 MW (currently not operational) • Orlando: Coal-fired, 300 MW (currently not operational) • Roggebaai: Gas turbine, 50 MW • Athlone Gas: Gas turbine, 40 MW (currently not operational) • Port Elizabeth: Gas turbine, 24 MW (currently not operational) • Orland Gas: Gas turbine, 176 MW (currently not operational) • Pretoria West: Gas turbine, 24 MW (currently not operational) Zambia Zambia relies mostly on hydropower. According to information from the official web site of the national utility ZESCO, 99.1% of the utility’s production came from hydropower sources. ZESCO’s thermal power plants amount to 10 MW small diesel-fired power plants. Many of these units were installed in the 1980-ies and spare parts represent a problem. An intention is to replace these small diesel stations with a combination 109 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA of grid-extension and mini-hydro plants.37 Therefore, we will not include them in refurbishment potential in the Least Cost Expansion Model. The installed thermal capacity in Zambia is, however, larger than the few MW of small diesel- fired plants owned and operated by ZESCO, due to gas-fired turbines installed to provide electricity to the many copper mines in Zambia. The privately owned Copperbelt Energy Cooperation (CEC) is the largest such operator, and has around 80 MW of stand-by gas-fired generation turbines.38 According to the company’s web-site, the system is modern and maintained to the highest international standards. Consequently, we assume that these privately owned generation systems are in no need of refurbishment during the next ten years. Zimbabwe Zimbabwe’s thermal generating capacity comes from the following coal-fired power plants • Bulawayo, 120 MW • Harare, 135 MW • Munyati, 120 MW • Hwange I, 480 MW (4*120) commissioned in 1984 • Hwange II, 440 MW (2*220MW) commissioned in 1986 Refurbishment requirements of the thermal fleet are large, due to the difficult financial position of ZESA, the state-owned utility company. For example, in October 2006, four generators at Hwange power station broke down, while the remaining two had been idle for several month due to lack of funds for repair/refurbishments.39 The three remaining thermal plants, which are older than the Hwange system, have an available capacity far below the installed capacity and are in need of refurbishment. Therefore, before 2015 it is assumed that all of Zimbabwe’s thermal generating capacity needs refurbishment WAPP region Benin Out of the total installed capacity, 89.5 MW are to be refurbished before 2015. This includes all the existing plants except Cotonou, Vedoko and some small plants. 37 Core International (2004) 38 http://www.copperbeltenergy.com/ 39 IRIN News 110 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table A2.14 Thermal generation in Benin (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Cotonou (Akpakpa) Combustion Turbines 1980 56 56 Natitingou 1 Diesel 1998 4.5 4.5 Natitingou 2 Diesel 2004 10.1 10.1 Parakou 1 Diesel 1998 7.8 7.8 Parakou 2 Diesel 2004 26.8 26.8 SIIF Porto-Novo Diesel 2004 17.4 17.4 Small units (SBEE's) Diesel 2000 8.5 8.5 Small units (SBEE's) Diesel 2004 8.5 8.5 Vedoko (CEB) Gas turbine 1998 25 20 Total Benin 164.6 159.6 Source: WAPP Key Performance Indicators 2005 Burkina Faso Out of the total capacity 116.5 MW are to be refurbished before 2015. Table A2.15 Thermal Generation in Burkina Faso (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Ouaga 1 G1/G2 Diesel 1991 6.8 5 Ouaga 1 G5/G6 Diesel 2003 6.8 5.2 Ouaga 2 G1 Diesel 1975 4 2 Ouaga 2 G2/G3/G4 Diesel 1978 19.8 12 Ouaga 2 G5/G6 Diesel 1982 19.8 12 Ouaga 2 G7/G8/G9 Diesel 1999 12.2 9 Kossodo G1/G2/G3 Diesel 2000 20.9 15.5 Kossodo G4/G5 Diesel 2003 16.2 16.2 Kossodo G6/G7 Diesel 2005 20.1 20.1 Koudougou G4/G5/G6 A Diesel 1978 3.2 3.2 Koudougou G4/G5/G6 B Diesel 1993 26.3 26.3 Bobo I G1/G2/G3/G4 Diesel 1976 7.5 7.5 Bobo II G1/G2 Diesel 1988 9.5 9.5 Bobo II G3/G4/G5 Diesel 1995 9.5 9.5 Total Burkina Faso 182.5 152.9 Source: Western Africa Regional Transmission Stability Study: Master Plan (2004) 111 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Cote d’Ivoire CIE, the company responsible for generation, transmission and distribution, operates three thermal plants with a total installed capacity 612 MW. 114 MW of this must be refurbished before 2015. Table A2.16 Thermal Generation in Cote d’Ivoire (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Vridi 1 TAG 5000 Gas (CCGT) 1984 100 100 Vridi 2 Ciprel Gas (CCGT) 1995 216 210 Azito IPP Gas (GT) 2000 296 288 Total Cote d’Ivoire 612 598 Source: WAPP Key Performance Indicators 2005 Gambia Gambia has a total capacity of 102.8 MW, of which 81 MW must be refurbished before 2015. Table A2.17 Thermal Generation in Gambia (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Kotu G1 Diesel 1981 3.4 2 Kotu G2 Diesel 1981 3.4 2 Kotu G3 Diesel 1997 3.4 3 Kotu G4R Diesel 2001 6.4 6.4 Kotu G5 Diesel 1990 6.4 6.4 Kotu G6 Diesel 2001 6.4 6.4 Kotu G7 Diesel 2001 6.4 6 Kotu G8 Diesel 1997 11 0 Other Diesel 2000 56 56 Total Gambia 102.8 88.2 Source: WAPP Key Performance Indicators 2005 Ghana 30 MW of total capacity is to be refurbished before 2015. Table A2.18 Table: Thermal Generation in Ghana (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Takoradi 1 (TapCo) CCGT 1997 330 330 Takoradi 2 (TiCo) CCGT 2000 220 220 Tema Diesel 1970 30 30 Effasu Diesel 2007 125 125 Total Ghana 705 705 112 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Guinea Out of the total capacity of 108 MW, only 44 MW is available till 2015. 64 MW of total capacity is to be refurbished before 2015. Table A2.19 Thermal Generation in Guinea (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Tombo 1 - 13 G Diesel 1997 11 1.5 Tombo 1 - 14 G Diesel 1987 11 0 Tombo 1 - 15 G Diesel 1989 11 0 Tombo 1 - 16 G Diesel 1993 11 0 Tombo 2 - 23 G Diesel 1997 6.8 0 Tombo 2 - 24 G Diesel 1997 4 4 Tombo 3 - 31 G Diesel 1997 11 11 Tombo 3 - 32 G Diesel 1997 11 11 Tombo 3 - 33 G Diesel 1997 11 11 Tombo 3 - 34 G Diesel 1999 11 11 Isolated centres 1990 9.2 Total Guinea 108 49.5 Guinea Bissau The entire capacity of 22.2 MW is to be refurbished before 2015. Table A2. 20 Thermal Generation in Guinea Bissau (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Bissau 1 Diesel 1980 8.3 8.3 Bissau 2 Diesel 1990 14 14 Total Guinea Bissau 22.2 22.2 Liberia The entire capacity of 9 MW is to be refurbished before 2015. Table A2.21 Thermal Generation in Liberia (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Monrovia Diesel 1980 9 9 Total Liberia 9 9 113 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Mali 19.5 MW of the total capacity must be refurbished before 2015. Table A2.22 Thermal Generation in Mali (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Dar Salam Gas 1999 20 20 Dar Salam Diesel 1998 7 7 Dar Salam Diesel 1998 6 6 Balingue IV Diesel 1997 18.8 18.8 Balingue V Diesel 1997 4.8 4.8 Mopti Diesel 2001 6.2 6.2 Sikasso Diesel 2001 6.0 6.0 Gao Diesel 2001 2.1 2.1 Tombouctou Diesel 2001 1.9 1.9 Other small diesel Diesel 1995 15.3 15.3 Sugar refineries Diesel 7 7 Compagnie Malienne de Dev Textile Diesel 8 8 Huilerie Cotonnerie du Mali Diesel 4.5 4.5 Total Mali 107.6 107.6 Mauritania 72.9 MW of total capacity is to be refurbished before 2015. Table A2.23 Thermal Generation in Mauritania (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Nouakchott (Arafat) Diesel 1980 42 30 Nouakchott (ksar) Diesel 2000 10.5 7 Nouadhibou Diesel 2003 22.1 12 Atar Diesel 2003 3.7 1.2 Néma Diesel 2003 1.5 0.4 Timbédra Diesel 2003 0.8 0.4 Aîoun Diesel 2003 1 0.6 Tintane Diesel 2003 0.4 0.3 Kiffa Diesel 2003 2.4 0.5 Geurou Diesel 2003 0.8 0.3 Sélibaby Diesel 2003 1.2 0.6 M’Bout Diesel 2003 0.4 0.2 Tidjikja Diesel 2003 0.8 0.3 M. Lahjar Diesel 2003 0.6 0.3 Aleg Diesel 2003 1.1 0.4 Boutilimit Diesel 2003 1.1 0.7 Akjoujt Diesel 2003 1.4 0.5 Other Diesel 2004 18 18 Total Mauritania 110 74.1 114 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Niger The entire capacity of 105 MW is to be refurbished before 2015. Table A2.24 Thermal Generation in Niger (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Niamey II Diesel 36 36 Arlit Coal 37.7 37.7 Other Diesel 1999 31 31 Total Niger 105 105 Nigeria In Nigeria, 5,003 MW of total capacity is to be refurbished before 2015. Table A2.25 Thermal Generation in Nigeria (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Egbin/Lagos CCGT 1986 1320 1147.8 Sapele SCGT 1981 1020 104.7 Delta SCGT 1989 912 393.5 Afam SCGT 1982 623 221.2 Ijora SCGT 1978 40 0 Oji SCGT 1956 30 0 AES Lagos* Gas 2005 300 300 AGIP Okpai Gas 1986 480 262.3 Omoku Gas 1980 150 100 Ajaokuta Gas 1990 55 21.1 Geregu Gas 2005 414 41 Omotosho Gas 2007 335 335 Papalanto/Olurunsogo Gas 2007 335 335 Total Nigeria 6,014 3,531.5 Senegal 290.8 MW of total capacity is to be refurbished before 2015. Table A2.26 Thermal Generation in Senegal (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Bel Air C2 Steam 1961 51.2 41.6 Bel Air C2 GT 1999 35 30 Bel Air C1 Diesel 1990 9 0 Cap de Biches C3 Steam 1978 87.5 75 Cap de Biches C3 GT 1984 30 30 Cap de Biches C3 GT 1994 30 30 Cap de Biches C4 Diesel 1990 30.3 30.3 Cap de Biches C4 Diesel 1997 30.3 30.3 Cap de Biches C4 Diesel 2003 30.3 30.3 115 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Cap de Biches, IPP GTI Combined cycle 1999 52 52 Aggreko, Cap de Biches Diesel 2005 48 48 St. Louis Diesel 1979 6 6 Kahone Diesel 1982 9.4 7 Kahone Diesel 1988 9.4 7 Tambacounda Diesel 7 7 Boutoude Diesel 2006 15.2 15 Isolated Centers (25) Diesel 1990 15.4 10 Total Senegal 496.1 449.6 Sierra Leone 54.5 MW of total capacity is to be refurbished before 2015. Table A2.27 Thermal Generation in Sierra Leone (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) King Tom 1 K3 Diesel 1980 3 2 King Tom 2 S4 Diesel 1979 9.2 7.5 King Tom 3 S5 Diesel 2008 6 6 King Tom 4 C1 Diesel 2000 1.3 1.1 King Tom 4 C2 Diesel 2000 1.3 1.1 King Tom 4 C3 Diesel 2000 1.3 1.1 King Tom 5 M6 Diesel 1995 5 4.5 King Tom 6 M3 Diesel 2002 6.3 5.8 Bo Diesel 1987 5 3.8 Private (mining firms) 28 28 Total Sierra Leone 66.3 60.9 Togo The entire 60 MW of the total capacity is to be refurbished before 2015. Table A2.28 Thermal Generation in Togo (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Lome Gas 1980 20 20 Various plants Gas 1980 40 20 Total Togo 60 40 116 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Island states Cape Verde Cape Verde has only two thermal plants, one coal based and the other diesel. The Matiota plant is to be refurbished before 2015. Table A2.29 Thermal Generation in Cape Verde (2005) Plant Type Commissioning date Installed capacity Available capacity (MW) (MW) Matiota Coal 1980 0.8 0.8 Many small plants Diesel 1991 77 77 Total Cape Verde 78 78 Source: Western Africa Regional Transmission Stability Study: Master Plan (2004). Madagascar Jirama, the state owned utility company, has 292 MW of installed capacity, 187 MW of which is thermal.40 The generating capacity is divided between 112 small stand-alone diesel and mini- hydro stations, and three urban interconnected networks serving Antananarivo-Antsirabe, Toamasina and Fianarantsoa. On a general level the thermal generating units are old and in poor condition and it is expected that a large number of the generating stations will be replaced before 2010. A World Bank Energy Sector Development Project between 1996 and 2005 modernized 82 small stand-alone units with a combined capacity of 20 MW.41 In addition, six of the larger thermal plants were refurbished. Nonetheless, a World Bank Power/Water Sector Recovery and Restructuring Project plan for future (2007-2011) refurbishment of additional 4-5 power plants. We assume that no additional old diesel plants will be refurbished during the next ten years. Mauritius 482.5 MW of capacity is to be refurbished before 2015. Table A2.30 Thermal Generation in Mauritius (2005) Plant Type Commissioning date Installed capacity (MW) Available capacity (MW) Fort George Diesel 1980 138 137 St Louis Diesel 1980 113 77.4 Fort Victoria I Diesel 1980 24 14 Fort Victoria II Diesel 2005 18 18 Nicolay Kerosene 1990 78 76 IPP Cogeneration 2000 247 117.5 Other Diesel 0 Total Mauritius 618 440 40 World Bank PAD Madagascar (2006) 41 World Bank ICR 37163 117 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Appendix 2 Documentation of hydropower refurbishment needs and investments The objective of this appendix is to detail the assumptions, approximations and methodology adopted for deriving a simple and robust methodology for estimation of hydropower development costs at country level. These assumptions are used in the Least Cost Expansion Model. Hydropower characteristics Hydropower differs from thermal power generation in the sense that the investment costs are determined by site-specific, natural conditions. The “fuel cost”, however, is zero provided the station does not compete with alternative water use as irrigation etc. The fuel in a hydropower station, i.e. water, is a stochastic variable and the most used key figure for indication of the production capacity is the so-called firm energy. The definition varies, but is always linked up to the expected supply security, i.e. the probability of having enough water to supply the indicated amount of energy. However, paying main interest to firm energy does not mean that “excess energy” is worthless. In many situations such energy can replace other generation sources with high fuel cost. This is even more applicable for regions with cross boundary interconnectors and free power trade where difference in hydrological regimes, load patterns or generation mix can give opportunities for trade of excess energy. Hence, average energy production is also an important key figure. A hydropower station with relatively modest storage capacity can also convert off-peak power to peak power if the generation capacity is increased, enabling it to produce more energy over a shorter time span. Pumped storage schemes develop this strategy to the extreme as they are pumping the water in off- peak hours in order to generate power in peak hours. Hence, their energy contribution is negative, but they refine low cost power to high value power. Peak production capacity will decline with lower dam filling where such gives lower head for the turbines. This factor is very noticeable with low head schemes. Also, more water is needed per MWh produced energy when the reservoir level is low. However, a reservoir higher up in a river course can increase firm energy production in downstream power stations. Some of the best hydropower sites are so big, difficult and/or costly to construct in steps that the domestic demand cannot absorb the large expansion of production capacity, making the cost per unit of actual produced energy high in the initial phase. In such situations, interconnectors that open a larger market can contribute to make the project financially viable. The natural data that mostly influence the cost are meteorological, hydrological, topographical and geological conditions: • Rainfall data together with run-off characteristics and size of catchment area determine the amount of water available for power production. Seasonal variations give less water available as firm energy. However, where favourable topographical conditions exists, storage dams can be established for a 118 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA reasonable cost and this can convert non-firm energy into firm energy by retaining water in wet seasons and releasing it in dry seasons. • The amount of power generated from a water course is dependant on water volume and altitude variations (head). Large head over short distance give less expensive power stations whereas gentle sloping river courses increase the distance between the intake and outlet and hence the development costs. The scaling of the installations is normally a trade-off between the marginal increase in construction cost and the extra production potential. This determines which part of the river will be included in the hydropower scheme and which sort of storage facilities that will be developed. • Construction of power stations involves tunnelling and/or construction of large civil structures such as dams and powerhouses. The costs of such installations depend heavily on geological and topographical conditions. Being so dependant on site-specific conditions, the energy price can only be calculated by in situ determination of construction cost and production capacity. Sometimes the size of the installation can be adapted to the local demand, but downscaling of an installation leads normally to increased cost per unit. Care should also be taken not to block future development of an optimally sized power station. Another factor influencing the cost is that a hydropower station must be located at a specific site often far from the consumers so large transmission costs may have to be added. If the hydropower plant is part of an irrigation scheme, limitations on when the water can be released may reduce the firm energy capacity and increase the cost of produced energy. However, if hydropower is a side benefit to an irrigation scheme, marginal cost considerations can make cheap hydropower available. Hydropower can have a large influence on the environment. Environmental costs are normally incorporated in the feasibility study and a project is, in this report, regarded as potentially environmentally acceptable if they are incorporated in a national or regional plan. Determination of potential production capacity of new projects As mentioned above, the production capacity of a hydropower site is basically a stochastic variable parameter, namely the run-off from the catchment area. This variable can to a certain degree be manipulated through storage reservoirs, through head variations created by selection of dam heights and by defining the project boundaries. In order to predict the production capacity for a potential site one has to find the best methodology to simulate the run-off from the catchment area. Long series of historical data are necessary for that (series of at least 30 years are normally created). The best accuracy is achieved if one has actual readings of water flow at the specific site over a long period, but if not, artificial series are created by statistical methods using regression analysis by comparison with similar rivers or catchments areas. The accuracy of the predictions is often amplified by taking into account rainfall recordings, evaporation measurements and other meterological information. The resulting flow series are then used for simulations at various production scenarios taking into account various plant options (head to be utilised, storage possibilities etc.). The potential income from 119 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA the production scenarios is compared with the investment predictions for the same scenarios to establish the most economic development. Firm production capacity is usually given most emphasis but peak production capacity and non-firm production capacity also play a role. For existing plants, the simulations are corrected by records of actual flow and productions, but the future production will also here be basically stochastic. Operation, refurbishment and reconstruction Hydropower stations have very low operation and maintenance cost; this is normally taken into account as 1.5–2 percent of the investment cost per year. In rivers with heavy silt loads, high wear on the turbine runners can increase the operation and maintenance cost. The major civil structures in a hydropower project (dams, tunnels and powerhouses) have a life span in the range of 100 years and more, provided normal maintenance. Machinery (gates, hydraulic steelworks, turbines, generators and transformers) have a life span of 30–50 years whereas smaller equipment (breakers, control equipment, protection etc.) have a life span of 10–20 years. However, major overhaul and upgrading are normally performed every 20 years. The nature of refurbishment work is such that it normally requires the replacement of runners, rewinding of generators, partial replacement of transformers and replacement and upgrade of switchgear, control and protection equipment. At the same time increase in production capacity can be achieved by using more modern equipment. Heavy siltation can reduce the lifespan of dams considerably if not sufficient desiltation arrangement is provided. Some of the countries incorporated in the study (e.g., Angola and Mozambique) have gone through civil wars and hydropower stations have been more or less destroyed, requiring additional funds for reconstruction. Other countries have been so strapped for funds that normal maintenance has not taken place. Such negligence leads to rapid deterioration of the installation which in turn increases the repair cost tenfold. Sources of data As explained above, the hydropower data are never generic, but site-specific. That's why some sort of study must be available in order to estimate the potential and cost of a project. The normal study cycle is screening of potential, prefeasibility study of interesting sites and feasibility study of the best sites. As the costs for hydropower projects are site specific, a generation master plan is normally developed for a country or a pool area where specific hydropower projects are mixed with other generation project forming a least cost development plan. A request for data and information from the utilities with a cover letter from the World Bank and follow-up by telephone failed to yield response. It was therefore necessary to base the study on information available in the public domain. 120 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Data for the EAPP, SAPP and WAPP regions is mostly derived from regional master plans. These are prepared by the WAPP and SAPP power pools, the East African Union and the Nile Basin Initiative through NELSAP (Nile Equatorial Lakes Subsidiary Action Program) and ENSAP (Eastern Nile Subsidiary Action Program). The data for the CAPP region is mostly derived from various internet sources and is therefore not as complete and coherent as for the other regions. Data sets In order to make the database manageable and not to overwhelm the respondents, only power stations larger than 10 MW are included in the study. Where available, predicted firm production capacity is used. Where firm production capacity is not available, average production capacity is used instead, and if this is not available, 50% firm load factor is assumed (this means that mean firm energy production is estimated to be 50% of peak power potential). Two economic key figures are calculated based on anticipated construction cost: • USD per kWh annual production • USD per kW peak power capacity Our unit investment costs are based on estimated investment costs for actual planned hydropower projects for each country. Where there are several planned projects, the unit investment costs used in the least cost expansion model is the weighted average of investment costs of the planned projects in the country, where weights reflect the plants’ planned capacity. Where riparian countries share water rights to a project, the cost and potential of the project are allocated equally between the partners. We want to emphasize that this is not an exhaustive list of planned hydropower projects in any region, but a subset for which good estimates of investment costs could be obtained. It should also be borne in mind that the economic key indicators only illustrate different merits of the project. As the input data are of very variable quality, the key indicators can be used for overall generic planning but not for detailed investment planning. The refurbishment needs are estimated based on year of construction, period of time from last known refurbishment and size of installation, if actual refurbishment information is not available. The tables below give an overview over • Potential and costs for new hydropower projects (tables A3.1–A3.5) • Refurbishment needs and costs of existing hydropower plants (tables A3.6–A3.10) 121 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Investment plans and costs for hydropower projects CAPP region Table A3.1 Investment Costs for Hydropower Projects, Central Africa Power Pool region Plant factor = Planned Planned Firm Planned energy Cost/ Country/ installed construction production/ production peak power Reference Project capacity cost full utilization (GWh) (USD/kW) (MW) (mill. USD) (percent) Cameroon Lom Pangar 56 50* 245.3 115 2054 4 Bini a Warak 75 50* 328.5 N/A N/A 5 Memve'Ele 202 50* 884.8 160 793 1 Nachtigal 267 45 1060 318 1191 2,4 Njock 150 72 950 371 2473 2 Sonmbengue 900 50* 3942 N/A N/A 3 Song Ndong 280 50* 1226.4 N/A N/A 3 Kikot 500 50* 2190 N/A N/A 3 Central African Republic Palambo 300 50* 1314 153 1500** Boali III 10 50* 43.8 N/A N/A 5 Republic of Congo Sounda II 160 50 700 230 1438 1,5 Sounda III 750 50* 3285 N/A N/A 5 Imboulou 100 74 645 200 2000 1,5 Kouembali 150 50* 657 N/A N/A Nambili 28 50* 122.6 N/A N/A Gamomba 14 50* 61.3 N/A N/A Mpe 20 50* 87.6 N/A N/A Kouli 180 50* 788.4 N/A N/A Dijoue ext. 15 67 88 58 3867 1 Chollet 600 50* 2628 N/A N/A Equatorial Guinea Name unknown 120 50 525.6 275 2292 Gabon Two projects 90 50* 394.2 N/A N/A 5 * Estimated 50 percent plant factor ** Assumed 1,500 USD/kW in the model simulations References: 1 UNDP/World Bank - Energy Sector Management Assistance Programme (ESMAP); 2001; "Africa Gas Initiative"; ESM 240 2 UN Framework Convention on Climate Change 3 WB Project Appraisal Document for an Energy Sector Development Project, 2008-05-29 4 Direct communication WB Task Manager 5 Various internet sources 122 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA EAPP – Nile Basin region Table A3.2 Investment Costs for Hydropower Projects, East Africa Power Pool region Plant factor = Planned Planned Planned Firm Cost/ Country/ installed energy construction production/ peak power Reference Project capacity production cost full utilisation (USD/kW) (MW) (GWh) (mill. USD) (percent) Burundi Ruzizi III (**DCR & Rwanda) 27 58 139.3 32 1171 1,2,6 Rusumo Falls (**Tanz. & Rwanda) 21 74 134.3 38 1839 1,2,6 Kabu 16 20 38 67 44 2200 1 Kaganuzi Complex 39 50 171 162 4154 1 Muele 17 26 39 38 2235 1 Jiji 03 16 24 33 49 3063 1 Mpanda 10 46 40 51 5100 1 Siguyaye 90 62 486 420 4667 1 Egypt Damietta Barrage 13 58 66 10 769 6 New Nagi Hammidi Barrage 64 83 464 386 6033 6 New Assuit Barrage 32 50* 140.2 42 1314 6 Ethiopia Beles 220 56 1086 294 1338 4,5,6 Chemoga Yeda (1&2) 281 57 1415 403 1434 4,5,6,7 Halele Wrabesa (1 & 2) 422 61 2245 474 1123 4,5,6,7 Karadobi 1600 61 8600 2040 1275 4,5,6,7 Baro 1 200 52 904 505 2525 4,5,6,7 Baro 2 + Genji 700 57 3505 502 717 4,5,6,7 Geba (1 & 2) 372 55 1788 361 418 4,5,6,7 Gojeb 153 39 520 287 1876 4,5,6,7 Neshe 43 58 220 69 1605 4,5,6 Alethu East I 186 49 800 438 2355 4,5,6,7 Alethu West 265 45 1050 561 2117 4,5,6,7 Genale III 258 53 1200 304 1178 7 Genale IV 256 45 1000 383 1496 7 Border 1200 57 6000 1626 1355 7 Mandaya 2000 69 12100 247 124 7 Kenya Ewaso Ngiro 220 32 609 493 2241 1,2,6 Low Grand Falls 140 58 715 455 3250 1,2,6 Magwagwa 120 64 669 383 3192 1,2,6 Mutonga 60 62 328 229 3817 1,2,6 Rwanda Nyabarongo 28 58 142 111 3964 1,2,6 Panzi (**DCR) 19 53 87.5 69 3632 1,2,6 Rusumo Falls (**Tanz. & Burundi) 31 74 201.5 57 1839 1,2,6 Ruzizi Div (**DCR) 135 55 650 160 1181 1,2,6 Ruzizi III (**DCR & Burundi) 27 58 139.3 32 1171 1,2,6 Sici 3 (**DCR) 87 58 441.5 203 2333 1,2,6 Sudan 123 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Plant factor = Planned Planned Planned Firm Cost/ Country/ installed energy construction production/ peak power Reference Project capacity production cost full utilisation (USD/kW) (MW) (GWh) (mill. USD) (percent) Dal Low 340 65 1944 1096 3224 3,5,6,7 Dagash 285 59 1476 1048 3677 3,5,6,7 Shereiq 315 56 1546 1190 3778 3,5,6,7 Sabaloka 120 49 520 396 3300 3,5,6 Fula I 720 65 4119 1319 1832 3,5,6,7 Shukoli 210 77 1422 420 2000 3,5,6,7 Lakki 210 77 1415 429 2043 3,5,6,7 Bedden 400 79 2761 880 2200 3,5,6,7 Kabjar 300 53 1400 1125 3750 3,5,6,7 Morgat 240 65 1360 595 2479 3,5,6 Shirri 400 58 2020 549 1373 3,5,6 Roseires ext. + Dinder 135 29 341 173 1281 3,5,6 Tanzania Upper Kihansi 120 12 126 81 677 2,3,4 Stiegler Gorge I 344 86 2598 1899 5520 2,3,4 Stiegler Gorge II 600 13 693 388 647 2,3,5 Stiegler Gorge III 300 67 1754 637 2123 2,3,4 Ruhudji 358 57 1800 611 1707 2,3,6,7 Rumakali 222 60 1170 351 1581 2,3,4 Mpanga 144 68 863 191 1326 2,3,7 Masigira 118 51 528 157 1331 2,3,4 Mandera 21 59 109 42 2005 2,3 Ruzomo (**Rwanda & Burundi) 62 57 308 2,3,4 Songwe (**Malawi) 330 47 1352 2,3 Uganda Bujagali 200 97 1703 526 2630 1,2,5,6 Bujagali # 5 50 51 222 31 620 1,2,5,6 Karuma 180 81 1271 429 2383 1,2,5,6 Kalagala tot 450 64 2525 512 1137 1,2,5,6 Ayago North 304 99 2624 730 2401 1,2,5,6 Murchison Base 2 222 91 1773 397 1788 1,2,5,6 Bygoe 13 72 82 25 1923 1,2,5,6 Masindi I 360 83 2615 1347 3742 1,2,5,6 Masindi II 360 57 1798 1088 3022 1,2,5,6 * Estimated 50 percent plant factor ** Riparian country References: 1 Power Development Options in the Nile Equatorial Lakes Region; SNC-Lavalin/Hydro Quebec; 2005 2 East African Master Plan; Acers; 2004 3 Long term Power System Planning Study, Sudan; ACRES, 1993 4 Ethiopian Power System Expansion Plan; ACRES, 2002 5 Various feasibility studies and master plans 6 Various internet sources 7 Odegard, World Bank, e-mail communication 124 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA SAPP region Table A3.3 Investment Costs for Hydropower Projects, Southern Africa Power Pool region Plant factor Planned = Firm Planned Planned Cost/ Country/ installed production/ energy construction peak power Reference Project capacity full production cost (USD/kW) (MW) utilisation (GWh) (mill. USD) (percent) Angola Capanda II 260 66 1500 344 1323 1,2,3,6 Cambambe II 260 40 900 772 2969 1,2,3,6 Medio Kwanza 6700 50* 29346 1,2,3 Baine (**Namibia) 180 55 862 320 1778 1,2,3,6 Gove 60 50* 263 58 967 6 Dem. Rep of Congo Inga III 3500 75 22995 1730 494 2,3,6 Grand Inga 1 6000 50* 26280 4025 671 2,3 Grand Inga 2 6000 50* 26280 4025 671 2,3 Busanga 240 60 1261 300 1250 2,3,6 Ruzizi III (**Rwanda & Burundi) 82 58 418 105 1280 2,3 Zongo 2 150 50 657 2,3 Nzilo 2 120 50 526 2,3 Lesotho Muela II 110 9 90 2,3,6 Oxbow 80 59 410 155 1938 2,3,6 Montosa Pumped Stor. 1000 n/a 2666 6 Malawi Kaphichira Phase II 58 92 469 50 862 2,3,4,6 Fufu 100 61 530 141 1410 2,3,4,6 Kholombizo 240 71 1500 391 1629 2,3,4,6 Mpatamanga 260 55 1250 397 1527 2,3,4,6 Songwe (**Tanzania) 343 45 1352 6 Mozambique Mepanda Uncua Ph I 1300 80 9070 2000 1538 2,3,5,6 Cahora Bassa North 850 38 2835 771 907 2,3,5 Boroma 160 50* 701 150 938 2,3,5 Lupata 654 50* 2865 1215 1858 2,3,5 Mazingir 40 35 123 55 1375 2,3,5,6 Lurio 183 50* 802 339 1852 2,3,5,6 Mawuzi III 60 50* 263 119 1983 2,3,5 Namibia Baine (**Angola) 180 55 862 320 1778 2,3,6 Popa 23 50* 101 - 2,3 South Africa Bramhoek Pumped Storage. 1332 n/a 1400 2,6 Steelport Pumped Storage. 1000 n/a 976 2,6 Zambia Kafue Lower 600 57 3000 600 1000 2,3,6 Itezhi Tezi 120 64 676 142 1183 2,3,6 Batoka (**Zimbabwe) 800 65 4550 1250 1563 2,3,6 125 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Plant factor Planned = Firm Planned Planned Cost/ Country/ installed production/ energy construction peak power Reference Project capacity full production cost (USD/kW) (MW) utilisation (GWh) (mill. USD) (percent) Lulapula 950 50* 4161 1500 1579 2,3 Kalungwhishi 163 62 880 210 1288 2,3,6 Kariba North ext. (**Zimbabwe) 360 16 500 192 533 2,3,6 Mupata Gorge (**Zimbabwe) 300 50* 1314 454 1512 2,3 Devils Gorge (**Zimbabwe) 300 50* 1314 454 1512 2,3 Zimbabwe Gairezi 35 50* 153 35 1000 2,3 Condo 100 50* 438 100 1000 2,3 Victoria Falls (**Zambia) 390 50* 1708 590 1513 2,3 Batoka (** Zambia) 800 65 4550 1250 1563 2,3,6 Mupata Gorge (**Zambia) 300 50* 1314 454 1512 2,3 Devils Gorge (**Zambia) 300 50* 1314 454 1512 2,3 Kariba South Extension 300 19 500 200 667 2,3,6 * Estimated 50 percent plant factor ** Riparian country References: 1 Planjemento Integrado do Setor Elétrico, Sondotécnia - Odebrecht, 2003 2 SAPP Priority Projects for Investor Consideration and Funding” presented on SAPP Executive Committee 21 November 2005 3 Various internet sources 4 Power System Development Study and Operation Study, Malawi; Lahmeyer, Knight Pieshold; 1998 5 EDM annual reports and information leaflets 6 Southern Africa Power Pool Study, Progress Report 4, Nexant, 2007 126 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA WAPP region Table A3.4 Investment Costs for Hydropower Projects, Western Africa Power Pool region Plant factor = Planned Firm Planned Planned Cost/ Country/ installed production/ energy construction peak power Reference Project capacity Full production cost (USD/kW) (MW) utilisation (GWh) (mill. USD) (percent) Benin Adjaralla (**Togo, 50%) 49 39 166.5 1,2 Adjaralla ext. (**Togo, 50%) 38.5 53 178 76.5 1987 1,2 Ketou 72 50* 315.4 302 4194 2,5 Olougbe 42 50* 184 210 5000 2,5 Assante 36 50* 57.7 287 7972 2 Batchanga 15 50* 65.7 75 5000 2 Burkina Faso Nounibiel (**Ghana, 80%) 48 39 162.4 228.8 4767 1,4 Diebuugo 12 50* 52.6 N/A N/A 4 Cote d'Ivore Soubre 316 54 1500 471 1491 1,4,5 Louga 280 50* 1226.4 678 2421 1,2,5 Gibo-Popoli 112 50* 490.6 N/A N/A 1 Boloubre 156 50* 683.3 N/A N/A 1 Tayaboui 100 50* 438 N/A N/A 1 Missouli 21 50* 92 N/A N/A 1 Kokumbo 78 50* 341.6 N/A N/A 1 Singrobo 80 50* 350.4 184 2300 1,2,5 Daboite 91 50* 398.6 269 2956 1,2,5 Malamasso 90 50* 394.2 245 2722 1,2 Abisso-Come 90 50* 394.2 N/A N/A 1 Ndieliesso 90 50* 394.2 315 3500 1,2,5 Gambia Yelitenda N/A N/A N/A N/A N/A 4 Ghana Bui 400 27 963 622 1555 1,2 Nounibiel (**Burkina Faso, 20%) 12 39 40.6 57.2 4767 1 Juale 193 50* 845.3 528 2736 2,5 Awisam 88 50* 385.4 174 1977 2,5 Abatumesu 63 50* 275.9 162 2571 2,5 Asouso 51 50* 223.4 150 2941 2,5 Kojokrom 36 50* 157.7 N/A N/A 3 Hemang 54 50* 236.5 N/A N/A 5 Sedukrom 26 50* 113.9 N/A N/A 5 Jomuro 26 50* 113.9 N/A N/A 5 Tanoso 30 50* 131.4 N/A N/A 5 Guinea Kaléta I 105 98 900 143 1362 1,4 Sambangalou (**Senegal) 60 38 200 139 2317 1,2 Kaléta II 210 71 1300 209 995 1,2 127 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Plant factor = Planned Firm Planned Planned Cost/ Country/ installed production/ energy construction peak power Reference Project capacity Full production cost (USD/kW) (MW) utilisation (GWh) (mill. USD) (percent) Kassa 118 45 463 93 788 1 Amaria I 300 55 1435 309 1030 1 Souapiti II 500 64 2825 605 1210 1,2,5 Amaria II 665 57 3325 835 1256 1,2,5 Balassa 181 30 470 171 945 1,3 Gozoguezia 48 62 259 110 2292 1 Morisanako 100 60 523 246 2460 1,2 Tiopo 120 56 590 296 2467 1,2 Bouréa 161 51 717 373 2317 1,5 Souapiti III 750 56 3667 1008 1344 1 Koukoutamba 281 35 858 440 1566 1,5 Diaoya 149 45 581 332 2228 1 Téné 76 30 199 122 1605 1 Fello Sounga 82 46 333 263 3207 1 Fomi 90 47 374 300 3333 1,2,5 Grand Kikon 291 26 656 637 2189 1 Guinea Bissau Salthino 20 50 88 82 4100 1 Liberia Manu River 45 47 187 N/A N/A 1 Mount Coffee expansion 60 50* 262.8 N/A N/A 5 Via Storage 142 50* 622 850 5986 2 SP2 214 50* 937.3 813 3799 2 SP-B1 120 50* 525.6 437 3642 2 Mano 1 180 50* 788.4 605 3361 2 Mano 2 74 50* 324.1 330 4459 2 Mali Kenie 56 36 178 100 1786 1,2 Feloue 62 61 330 84 1350 1,2,4 Gouina 95 54 450 160 1684 1,2 Tossaye 30 50* 131.4 330 11000 1,2 Baoule 3 24 50* 105.1 119 4958 2 Bagoe 2 36 50* 157.7 182 5056 2 Gaoulgo 300 50* 1314 814 2713 2 Labeanga 17 50* 74.5 211 12412 2 Niger Kandadji 125 52 564 359 2872 1,2,4,5 Gambou I 50 50* 219 N/A N/A 1 Dyondyonga 26 36 82 69 2654 1,5 Nigeria Lakoja 1950 50* 8541 4463 2289 1,2,5 Makurdi 1062 50* 4651.6 1983 1867 1,2,5 Zungeru 950 50* 4161 N/A N/A 1,4 Mambila 3900 50* 17082 2000 513 1 Onitsha 1050 50* 4599 N/A N/A 1 128 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Plant factor = Planned Firm Planned Planned Cost/ Country/ installed production/ energy construction peak power Reference Project capacity Full production cost (USD/kW) (MW) utilisation (GWh) (mill. USD) (percent) Ikom 736 50* 3223.7 N/A N/A 1,4 Guara 300 50* 1314 N/A N/A 1 Dadin Kova 34 50* 148.9 N/A N/A 4 Katsina Ali 310 50* 1357.8 N/A N/A 4 Senegal Sambangalou (**Guinea) 60 38 200 139 2317 1,2 Sierra Leone Bumbuna Ph II-III 275 58 1400 957 3480 1,2 Benkongor I-III 200 66 1164 511 2551 1,2 Kuse II 92 84 680 N/A N/A 1 Kambatibo 53 58 269 N/A N/A 1 Bitmai I – II 89 66 518 N/A N/A 1 Manu River 45 47 187 N/A N/A 1 Togo Adjaralla (**Benin, 50%) 49 39 166.5 N/A N/A 1,2 Adjaralla ext. (**Benin, 50%) 38.5 53 178 77 1987 1 Tetetou 20 50* 87.6 132 6600 2 Tchala 90 50* 394.2 146 1622 2 * Estimated 50 percent plant factor ** Riparian country/shared power plant References: 1 Nexant; 2004; "West Africa Regional Transmission Stability Study, Vol 2, Master Plan" 2 Sparrow, F.T., W.A. Masters, B.H. Bowen, Purdue University & Jeffrey C. Metzel AIRD; 2000; "Electricity Trade and Capacity Expansion Options in West Africa, Revised Draft" 3 UNDP/World Bank - Energy Sector Management Assistance Programme (ESMAP); 2001; "Africa Gas Initiative"; ESM 240 4 Various internet sources 5 Direct communication with World Bank Task Manager 129 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Island states Table A3.5 Investment Costs for Hydropower Projects, Island states Plant factor = Planned Planned Firm Planned energy Cost/ Country/ installed construction production/ Full production peak power Reference Project capacity cost utilisation (GWh) (USD/kW) (MW) (mill. USD) (percent) Cape Verde No candidates Madagascar No candidates Mauritius No candidates 130 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Refurbishment needs and costs of existing hydropower projects CAPP region Table A3.6 Assumptions regarding refurbishment of hydropower projects, CAPP area Country/ Refurbishment Capacity to be Refurbishment Refurbishment Reference Station year refurbished cost cost (MW) (mill. USD/MW) (mill. USD) Cameroon Edea I, II and III 2009 263 0.3 61 1,3 Song Loulou 2010 387 0.3 96 1,3 Lagdo 2013 72 0.3 22 1,3 Central African Republic Boali I-II 2009 18 0.5 9 Republic of Congo Djouel 2010 15 0.3 5 2,3 Maukaukoule 2012 74 0.3 22 2,3 Sounda 2014 10 0.3 12 3 Gabon Kinguélé 2010 58 0.3 17 3 Tchimbélé 2011 69 0.3 21 3 Poubara I & II 2015 36 0.3 11 3 References: 1 Nexant (2004): "West Africa Regional Transmission Stability Study, Vol 2, Master Plan" 2 UNDP/World Bank - Energy Sector Management Assistance Programme (ESMAP) (2001): "Africa Gas Initiative"; ESM 240 3 Various internet sources EAPP – Nile Basin region Table A3.7 Assumptions regarding refurbishment of hydropower projects, EAPP/NB area Country/ Refurbishment Installed Refurbishment Refurbishment Reference Station year capacity cost cost (MW) (mill. USD/MW) (mill. USD) Burundi Rwegura 2015 18 0.3 5 Egypt Aswan High Dam 2100 6 Old Aswan HPP 550 6 Esna 90 6 Naga Hamadi 64 6 Ethiopia Tis Abay I 2008 11 0.3 3 4,5,6 Tis Abay II 2009 73 0.3 22 4,5,6 Finchaa (I-IV) 134 4,5,6 Koka 43 4,5,6 Awash II 2010 32 0.3 10 4,5,6 Awash III 2011 32 0.3 10 4,5,6 Awash IV 2012 34 0.3 10 4,5,6 Melka Wakena 2007 153 0.3 46 4,5,6 131 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Country/ Refurbishment Installed Refurbishment Refurbishment Reference Station year capacity cost cost (MW) (mill. USD/MW) (mill. USD) Gilbet Gibe I 183 4,5,6 Gilbet Gibe II 420 4,5,6 Kenya Turkwel Gorge 2009 106 0.3 32 1,2,6 Tana 14 1,2,6 Masinga 2010 40 0.3 12 1,2,6 Kamburu 2013 92 0.3 28 1,2,6 Gitaru 2010 145 0.3 44 1,2,6 Kiandurama 40 - 1,2,6 Kiambere 144 - 1,2,6 Sondu Miru 60 1,2,6 Rwanda Ntaruka 2010 11.5 0.3 3 1,2,6 Mukungwa 2012 12.4 0.3 4 1,2,6 ** Ruzizi I&II ( DCR & Burundi) 8.1 0.3 2 1,2,6 Sudan Roseires 280 3,6 Merowe 1250 3,6 Sennar 50 3,6 Jebel Aulia 30.4 3,6 Kashm Algirba 12.6 3,6 Tanzania Hale 2009 21 0.3 6 2,6,7 Pangani 2020 68 0.3 20 Kidatu 2015 204 0.3 61 Mtera 2010 80 0.3 24 Lower Kihansi 2025 180 0.3 54 Uganda Mboku III 2010 11 0.3 3 1,2,5,6 Nabuaale 2015 180 0.1 18 1,2,5,6 Kiira 250 1,2,5,6 ** Riparian country References: 1 Power Development Options in the Nile Equatorial Lakes Region; SNC-Lavalin/Hydro Quebec; 2005 2 East African Master Plan; Acers; 2004 3 Long term Power System Planning Study, Sudan; ACRES, 1993 4 Ethiopian Power System Expansion Plan; ACRES, 2002 5 Various feasibility studies and master plans 6 Various internet sources 7 SAPP Priority Projects for Investor Consideration and Funding 132 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA SAPP region Table A3.8 Assumptions regarding refurbishment of hydropower projects, SAPP area Country/ Connected Refurbish- Capacity to Refurbish- Refurbish- Reference Station to ment year be ment cost ment cost refurbished (USD/MW) (mill. USD) (MW) Angola Mabubas nat. grid 2010 18 0.3 5 1,3 Cambambe nat. grid 2015 180 0.3 54 1,3 Capanda nat. grid 520 0.3 156 1,3 Biopio island 2010 23 0.3 7 1,3 Lomaum island 2008 35 0.3 11 1,3 Luachimo island 2008 10 0.3 3 1,3 Matala island 2008 41 0.3 12 1,3 Democratic Republic of Congo Inga I SAPP 2007 341 0.23 78 2,3,4 Inga II SAPP 2007 1424 0.32 452 2,3,4 Zongo Songa SAPP 2010 89 0.22 20 2,3,4 Nseki SAPP 2009 262 0.21 56 2,3,4 Nzilo SAPP 2009 108 0.26 28 2,3,4 Koni island 2008 42 0.36 15 2,3,4 Mwadingusha island 2010 70 0.43 30 2,3,4 Tshopo island 2010 19 0.3 6 2,3 Mobay island 2010 11 0.3 3 2,3 Ruzizi I (**Rwanda & Burundi) Gr. Lack 2010 28 0.3 8 2,3 Ruzizi II (**Rwanda & Burundi) Gr. Lack 2015 40 0.3 12 2,3 Solen II island 2009 11 0.3 3 2,3 Budane island 2009 10 0.3 3 2,3 Lesotho Muela SAPP 2020 72 0.3 22 3 Malawi Tedzani nat. grid 2008 90 0.13 12 Kaphichira nat. grid 2030 128 0.3 38 Nkula Falls nat. grid 2015 10 0.3 3 Mozambique Corumana SAPP 2020 14 0.3 4 Mavuzi SAPP 2009 62 0.32 20 Chicambha SAPP 2009 38 0.26 10 Cahora Bassa SAPP 2075 0.3 623 Namibia Ruacana SAPP 2020 249 0.3 75 3 South Africa Gariep SAPP 2010 360 0.3 108 3 Van der Kloof SAPP 2010 240 0.3 72 3 Zambia Kariba North (**Zimbabwe) SAPP 2009 720 0.03 20 2,3 Kafue Gorge, Upper SAPP 2009 900 0.02 20 2,3 Mulungushi SAPP 2008 20 0.3 6 2,3 Lunsemfa SAPP 2009 18 0.3 5 2,3 133 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Country/ Connected Refurbish- Capacity to Refurbish- Refurbish- Reference Station to ment year be ment cost ment cost refurbished (USD/MW) (mill. USD) (MW) Lusiwasi SAPP 2010 12 0.3 4 2,3 Victoria Falls SAPP 2025 108 0.3 32 2,3 Zimbabwe Kariba South (**Zambia) SAPP 2009 760 0.3 228 2,3 ** Riparian country References: 1 Planjemento Integrado do Setor Elétrico, Sondotécnia - Odebrecht, 2003 2 SAPP Priority Projects for Investor Consideration and Funding 3 Various internet sources 4 Southern Africa Power Pool Study, Progress Report 4, Nexant, 2007 134 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA WAPP region Table A3.9 Assumptions regarding refurbishment of hydropower projects, WAPP area Country/ Refurbishment Capacity to be Refurbishment Refurbishment Reference Station year refurbished cost cost (MW) (mill. USD/MW) (mill. USD) Benin Nangbeto East (*Togo) 2010 31 0.3 9 1,2 Burkina Faso Bagre 2013 18 0.3 5 1,2 Kompienga 2008 15 0.3 5 1,2 Nounibiel 2010 30 0.3 9 1,2 Cote d'Ivore Ayame 2009 52 0.3 16 1,2 Kossou 2010 174 0.3 52 1,2 Taabo 2011 210 0.3 63 1,2 Buyo 2012 165 0.3 50 1,2 Ghana Kpong 2009 160 0.3 48 1,2 Guinea Donkea 2010 14 0.3 4 1 Grandes Chutes 2012 14 0.3 4 1 Liberia Mount Coffee 2011 64 0.5 32 1,2 Mali Selingue 46 0.3 14 1,2 Nigeria Kainji 2009 760 0.3 228 1,2 Jebba 2010 540 0.3 162 1,2 Shiroro 2011 600 0.3 180 1,2 Zungeru 2012 950 0.3 285 1,2 Togo Nangbeto West (*Benin) 2010 31 0.3 9 1,2 *Riparian country/shared power plant References: 1 Nexant (2004): "West Africa Regional Transmission Stability Study, Vol 2, Master Plan" 2 Various internet sources 135 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Island states Table A3.10 Assumptions regarding refurbishment of hydropower projects, island states Country/ Refurbishment Capacity to be Refurbishment Refurbishment Reference Station year refurbished cost cost (MW) (mill. USD/MW) (mill. USD) Cape Verde No hydropower Madagascar Mandraka 2011 24 0.3 7 1 Andekaleka 2012 58 0.3 17 1 Mauritius No refurbishment References: 1 Various internet sources 136 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Appendix 3 Documentation of asset stock and refurbishment needs of transmission and distribution The objective of this section is to detail the assumptions, approximations and methodology adopted for deriving a simple and robust methodology for generic cost estimation at country level of the transmission and distribution networks. The starting point for the analysis was to examine data which is available for those countries where the Consultant has had a direct involvement in transmission and distribution investment analysis, through other projects and activities. A request for data and information from the utilities with a covering letter from the World Bank and telephone follow-ups failed to yield responses. It was therefore necessary to progress the study using information available in the public domain, some country specific projects and regional data. It was observed that in each country there are various voltages, substation configurations, line constructions and ratings at transmission, subtransmission and distribution level, with the exception of a single voltage level for domestic supplies. The observations for the individual countries in the region are tabulated below. The countries can be grouped as follows: • Group I: Two voltage levels at transmission and one or two voltage levels at sub-transmission. • Group II: One voltage level at transmission and one or two at subtransmission. It was also observed that there were a small number of unusual line voltages for a few lines (including interconnection lines) and double circuit line construction in some countries. The total route lengths of these lines were not significant. Total line lengths and ratings at the higher voltages were noticeably higher than those at the lower voltages as expected because these lines are designed to transfer large amounts of power. Table A4.1 Central Africa Group II-Utility Voltage Levels Low Voltage Country Utility Transmission (kV) Subtransmission (kV) Distribution (kV) (kV) Cameroon AES n/a 225 90 n/a 30 11 0.4 CAR Enerca Chad STEE n/a 161 69 n/a 33 11 0.4 Republic of SNE 220 n/a n/a n/a 20 & 10 6.6 & 5.5 0.38 Congo Equatorial SEGESA 0.38 Guinea Gabon SEEG 20 5.5 0.38 Note: n/a denotes no other voltage at the level. 137 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Table A4.2 Eastern Africa Group II-Utility Voltage Levels Low Voltage Country Utility Transmission (kV) Subtransmission (kV) Distribution (kV) (kV) Burundi REGIDESO n/a 110 n/a 70 35 & 30 10 0.4 Djibouti EdD n/a 63 n/a n/a 20 n/a n/a Egypt EGELEC 220 n/a 132 66 33 11 0.4 Ethiopia EEPCo 230 132 66 n/a 33 11 0.4 Kenya KPLC n/a 220 132 66 33 11 0.4 Rwanda ELECTROGAZ n/a 110 70 n/a 30 15 0.4 Sudan NEC n/a 220 110 66 33 11 0.4 Tanzania TANESCO 220 132 66 n/a 33 11 0.4 Uganda UMEME n/a 132 66 66 33 11 0.4 Note: n/a denotes no other voltage at the level. Table A4.3 SAPP Group I-Utility Voltage Levels Low Voltage Country Utility Transmission (kV) Subtransmission (kV) Distribution (kV) (kV) DRC SNEL 220 120 70 15 6.6 0.4 Mozambique EdM 220 110 66 n/a 33 11 0.4 Namibia NAMPOWE 220 132 66 33 22 11 0.4 R South Africa ESKOM 400 275 132 88 22 & 33 11 0.4 Zambia ZESCO 330 220 88 66 33 11 0.4 Note: n/a denotes no other voltage at the level. Table A4.4 SAPP Group II-Utility Voltage Levels Low Voltage Country Utility Transmission (kV) Subtransmission (kV) Distribution (kV) (kV) Botswana BPC n/a 220 132 66 33 11 0.4 Lesotho LEC n/a 132 66 n/a 33 11 0.4 Malawi ESCOM n/a 132 66 n/a 33 11 0.4 Zimbabwe ZESA 330 n/a 88 66 33 11 0.4 Note: n/a denotes no other voltage at the level. Table A4.5 Western Africa Group II-Utility Voltage Levels Low Voltage Country Utility Transmission (kV) Subtransmission (kV) Distribution (kV) (kV) Benin CEB n/a 161 69 n/a 33 11 0.4 Burkina Faso SNEB n/a 161 69 n/a 33 11 0.4 Cote d’Ivoire CIE n/a 225 90 n/a 30 6.6 0.4 Gambia NAWEC 33 11 0.38 Ghana VRA n/a 161 69 n/a 33 11 0.4 Guinea EDG n/a 110 60 n/a 30 11, 6.3, 0.38 5.5 138 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Guinea Electricidade 10 & 6 0.38 Bissau et Aguas de Guinee Bissau Liberia LEC n/a 69 n/a 12.6 7.2 0.24 Mali EDM n/a 225 90 n/a 30 11 0.4 Mauritania SOMELEC 225 n/a 90 n/a 33 15 0.4 Niger NIGELEC n/a 132 n/a n/a 33 11 0.4 Nigeria NEPA 330 n/a 132 n/a 33 11 0.4 Senegal SENELEC n/a 225 90 n/a 30 6.6 0.4 Sierra Leone NPA n/a 110 66 n/a 33 11 0.4 Togo CEET n/a 132 66 n/a 33 11 0.4 Note: n/a denotes no other voltage at the level. Table A4.6 Island States Group II-Utility Voltage Levels Low Voltage Transmission (kV) Subtransmission (kV) Distribution (kV) (kV) Cape Verde Electra 20 15,13, 6, 0.38 6.3 Madagascar LA JIRAMA n/a 138 63 n/a 35 15 0.4 Mauritius CEB n/a 132 66 n/a 22 6.6 0.4 Note: n/a denotes no other voltage at the level. Assumptions and approximations In order to develop a simple and robust methodology for generic estimation of the cost of power system equipment, it is necessary to make assumptions and approximations. These assumptions and approximations will either be of a general or a specific nature at transmission, sub-transmission, distribution or low voltage level for all the countries, some which are country specific because of the diverse nature of the systems. Whilst these assumptions and approximations, which are made from knowledge of the region and various sources of data, will lead to a reduced order of accuracy, the resulting equipment costs will be adequate for the purpose of the investment study. General assumptions and approximations for all network subsystems The assumptions and approximations of a general nature applied at each voltage level are detailed below: The line rating will always be greater than the demand to be supplied. The resulting excess line capacity and interconnection is the cost for the security of that supply. Substation costs based on assumptions as to typical configurations and ratios of substations per unit length of transmission lines were factored into the transmission line costs. At each system level, except for low voltage, the design is (N-1) security which includes interconnection. Where there are more than two transmission or sub-transmission voltage levels in a country, the higher of the two voltages would be used for line investment as this is likely to be more cost effective. 139 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Where several voltage levels are present in a country voltage levels are lumped to enable the estimation of the maintenance and refurbishment costs within the transmission and distribution categorization. Interconnecting voltages will only be considered as a transmission voltage if there are other lines in the country under consideration. Transmission level assumptions The assumptions that are applicable at the transmission level are detailed below. Generator transformer and substation costs were considered to be part of the generation costs. For cost comparison, if there are two transmission voltages, then the higher voltage line will have the cost of a duplicated HV line bay (extra line bay for interconnection), transformers of adequate rating including HV and LV bays, stepping down to the lower transmission voltage as shown in below. This is applicable to Group I Country Utilities. Figure A4.1 Group I – Higher level voltage configuration For Group I-Country Utilities, the lower transmission voltage line cost will comprise the line construction cost, two line bays plus the cost of an extra bay for the interconnection as shown in figure below. Figure A4.2 Group I – Lower level voltage configuration. 140 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA For Group II-Country Utilities, the unit cost for the transmission voltage was determined separately for the two different sub-transmission voltages as shown in below. The subsequent sub-transmission line costs excluded the transmission substation costs but included two line bays on either end plus the actual line costs. Figure A4.3 Group II – Higher and Lower voltage configuration With the data available, the average transmission line length in each country differed from approximately 100 km to 200 km. This was calculated by ignoring very short lines which would have distorted the average. An average transmission length of line of 150 km was then calculated for all the countries. This line length is reasonable when considering investment costs as the countries have some infrastructure already in place and it should be borne in mind that the overall objective is to derive generic investment costing. If an investment were to be made which was say, twice the average country line length, this would mean duplicating the terminal equipment in the costing at the transmission voltage. This is acceptable as over the planning horizon there may be a requirement for an intermediate substation as the load increases with increased access to electricity. The nature of investments on power systems is discrete in nature; this implies that the unit costs in kUSD/km/MW can only be used for the calculated average line rating for new equipment. The decision to use the higher or lower voltage unit investment cost is based on the length of line, the power to be transferred and other investment parameters as well as any country standard voltage. Some odd voltage levels were ignored as they were considered to be of special design (for example 765 kV in ESKOM.). Because of the very high line ratings inclusion of these circuits would lead to very low unit costs (kUSD/km/MW), which would distort the average unit investment costs. Subtransmission level assumptions The assumptions that are applicable at the sub-transmission level, i.e. the second voltage transformation level after the generator transformer, are detailed below With the data available and knowledge of the region, the average sub-transmission line length in each country was estimated to be 50 km. 141 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA For both groups of countries, the unit costs will comprise the line cost, two line bays, plus half the cost of the transmission substation. Half the cost of the substation is assumed as it would be expected that there will be at least two out going circuits. Figure A4.4 Group I and Group II – Subtransmission voltage configuration Distribution level assumptions The assumptions that are applicable to the distribution voltage levels are detailed below. With the data available, the average distribution line length in each country was estimated to be 30 km for the 33 kV and 5 km for 11 kV. For 33 kV it was assumed that for the 33/11 kV substation there will be two bays at the 11 kV side. The cost of the distribution line was calculated as the total of half the cost of the subtransmission substation and the interconnection bay at the high voltage side plus the actual distribution line cost and the two LV bays. The exception was for SNEL in DRC where the use of 15 kV necessitates the use of a higher number of bays. This is applicable to both Group I and Group II Country Utilities. Figure A4.5 Group I and Group II – Distribution voltage configuration For 33/11 kV substations it was assumed that the 11kV side will be equipped with metalclad switchgear with on average six bays except for SNEL in DRC where there is use of 6.6 kV which will necessitate a higher number of bays. For MV line material and transformers, there are small differences in price between lower MV levels of 6 kV or 11 kV and the higher 22 kV equipment 142 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Domestic low voltage assumptions/approximations An After Diversity Maximum Demand of 2.0 kVA was judged to be good average for each domestic supply.42 The average transformer size was taken as 200 kVA because from the data available the transformer sizes vary from 25 kVA to 630 kVA. From knowledge of the region there are numerous conductor sizes and types for LV networks. In this study it only one size was assumed, which is 95mm2 aerial bundled conductor. It is assumed that the LV transformer will be 50% loaded to allow for load growth. The LV line length was approximated to be on average no more than 500m with two lines per transformer loaded to about 50% to allow for load growth. The total cost of the LV line included the pole mounted transformer and the cost of the 0.4 kV line including all tee-offs. In the absence of country specific data, the same distribution data has been used for all the countries except for SNEL-DRC which has different system voltages. This was judged to be a fairly good assumption as the major source of power system equipment and materials for African countries is South Africa and there are several initiatives to standardize in the region. CAPP: Country-specific assumptions and unit costs There is very limited data and information on CAPP utility networks on the public domain or from specific projects in the region. Central Africa is the least electrified sub-region in Africa. CAPP has proposals for the interconnection of the CAPP countries and external interconnection with WAPP, SAPP and EAPP. Group II – CAPP Country Utilities Cameroon – AES AES has one voltage at each level, transmission at 225 kV and subtransmission at 90 kV. Distribution is at 30 kV, 6.6 kV and 0.4 kV. No data has become available regarding interconnections, however, we believe that there are no major inter connection with Chad or Nigeria. Central African Republic – Enerca Very little information is available about the electricity system in the Central African Republic, which is likely to be limited in its extensiveness as only 3% of the population have access to electricity and the 42 However, in many developing regions, notably Africa, supplies are at a much lower capacity, typically ranging between 5 A, 1.1 kW and 10 A, 2.2 kW. In South Africa a 550 W supply has been piloted, and is now the default option, with a design after-diversity-maximum-demand of 0.4 kVA (R. Stephen, personal communication, 1998). These limited-current supply options offer the potential to reduce the distribution infrastructure costs significantly in comparison with conventional 20 A or greater supply capacities. There are also significant advantages in terms of load factor, as the magnitude of domestic peaks is effectively reduced. 143 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA country has had an unstable past. There are no interconnections. CAPP have plans to interconnect the Central African republic with the Democratic Republic of Congo. Information from the World Bank details a short 120 kV transmission line within the country. The lengths of the distribution system and lv system have been estimated from other similar countries based on the distribution length / net consumption of the country. Chad – STEE There is no significant network in Chad mainly due to the prolonged civil strife in the country. The network is concentrated in N'Djamena at 20, 15 and 0.4 kV. Peak electricity consumption in Chad is less than 25 MW, of which about 20 MW is generated in N'Djamena, Chad’s capital, the remainder being in three smaller urban centers (Moundou, Sahr and Abeche). There are about 7,500 electricity customers in a country of 7 million people. In other words, only 1.5% of Chad’s population benefits from electricity. This electricity density rate is on of the lowest in the world.43 Republic of Congo – SNE The power infrastructure was severely damaged during the conflicts which ended in 2003. Prior to the conflict there was transmission at 225 kV and 110 kV. Distribution was at 35 kV. There is a 225 kV transmission interconnection with DRC. CAPP have plans to connect the Republic of Congo with the Democratic Republic of Congo and Gabon. The length information provided by the World Bank from their 1996 source has been used in the tables as it is likely the system will be being reconstructed. Equatorial Guinea – SEGESA There are two networks, one on the mainland and one on Bioko Island, which are likely to be at distribution voltage level. There is no interconnection to other countries. CAPP have plans to connect Equatorial Guinea with Gabon and Cameroon. No information has been obtained about the electrical system, therefore the lengths of the distribution system and LV system have been estimated from other similar countries based on the distribution length / net consumption of the country. Gabon – SEEG Information about the transmission and distribution systems in Gabon has been obtained from the World Bank. There is transmission at 225 kV and distribution at 90 kV, 63 kV, 30 kV and 20 kV. There is no interconnection to other countries. CAPP have plans to interconnect Gabon with Cameroon, Equatorial Guinea and the Congo. 43 World Bank: Chad—Critical Electricity and Water Services Rehabilitation Project. Report No. PID10962 February 25, 2002 144 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA EAPP – Nile Basin: Country-specific assumptions and unit costs Group I – Eastern Africa/Nile Basin Country Utilities There is very limited data and information on Eastern African utility networks on the public domain or from specific projects in the region, mainly because most of them have been privatized, with KPLC in Kenya being an exception. The only country with two levels of transmission voltages is Tanzania. This may be due to low load densities in most countries except Egypt. Tanzania – TANESCO TANESCO has two transmission voltages 220 kV and 132 kV, 66 kV at sub-transmission with 33 kV, 11 kV and 0.4 kV on the distribution system. Tanzania and Kenya have completed a feasibility study for linking Arusha to Nairobi by a 220 kV transmission line. Group II – Eastern Africa/Nile Basin Country Utilities Burundi – REGIDESO There is only one voltage at each level in Burundi, transmission is at 110 kV and subtransmission is at 70 kV. Distribution is at 35 kV, 30 kV, 10 kV and 0.4 kV. Burundi and Eastern DRC are connected at 110 kV. Djibouti – EdD In Djibouti there is only 5 km of 63 kV transmission lines and 245 km of distribution lines at 20 kV. There is a current project to connect Djibouti to Ethiopia at 230 kV. Egypt – EGELEC Egypt’s utility EGELEC is the second largest in Africa after ESKOM, South Africa. The system has two transmission voltages 220 kV and 132 kV, sub-transmission is at 66 kV with 33 kV, 11 kV and 0.4 kV on the distribution system. Egypt is connected to other countries through the 220 kV Libya-Tunisia interconnection which will soon be upgraded to 400 kV. Egypt is also connected to the Middle East through 500 kV/400 kV Egypt- Jordan interconnection and to the European electricity system through the Morocco-Spain interconnection. Ethiopia – EEPCo In EEPCo there is only one voltage at each level, transmission at 230 kV and subtransmission at 132 kV. Distribution is at 33 kV, 11 kV and 0.4 kV. There are plans to interconnect: • Ethiopia with Sudan by a 220 kV power line; • Ethiopia with Djibouti with a 220 kV power line; • Sudan with Egypt by linking Merowe hydropower station to the Aswan dam. 145 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Kenya – KPLC KPLC has one transmission voltage at 220 kV with subtransmission at 132 kV and 66 kV. Distribution is at 33 kV, 11 kV and 0.4 kV. Kenya and Tanzania have completed a feasibility study for linking Nairobi to Arusha by a 220 kV transmission line. Rwanda – ELECTROGAZ In Rwanda there is only one voltage at each level, transmission is at 110 kV and subtransmission is at 70 kV. Distribution is at 15 kV, 6.6 kV and 0.4 kV. Rwanda and Uganda are connected at the level of border towns in 30 kV, and there are plans to link the two countries by a 132/110 kV the transmission line. Sudan – NEC In NEC there is only one voltage at each level, transmission is at 220 kV and subtransmission is at 110 kV. Distribution is at 66 kV, and 33 kV, 11 kV and 0.4 kV. There seems to be no interconnection with neighbouring countries. Uganda – UMEME In UETCL &UMEME (the Ugandan Transmission and Distribution Company, respectively) there is only one voltage at each level, transmission at 132 kV and subtransmission at 66 kV. Distribution is at 33 kV, 11 kV and 0.4 kV. There is an interconnection with Kenya at 132 kV and Rwanda at 30 kV which provide adhoc cross border supplies. SAPP: Country-specific assumptions and unit costs This section deals with country specific data because of different voltages and transformation configuration at transmission and subtransmission level within the countries. The costs of distribution investments tend to be less installation specific and more generic than is the case for transmission investments. Distribution investments are influenced by the choice of primary distribution voltage levels to a degree, and in the SAPP countries the voltages are fairly standard Group I – SAPP Country Utilities South Africa – ESKOM ESKOM is by far the largest utility by asset base in every category of the system as well as having a very complex system with numerous voltage levels and transformation, which includes 765(DC), 400, 275, 220, 132, 88, 33, 22, 11 and 0.4 kV. From the available data and some knowledge of the system, it was assumed that the transmission voltages would be 400 kV, 275 kV and subtransmission, 132 kV and 88 kV, with distribution at 33 kV and 11 kV with 0.4 kV low voltages. Any line lengths at intermediate voltage (220 kV and 66 kV) are considered at the higher voltages in the estimates. 146 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The data and information available did not include the large number of Regional Electricity Distribution Companies including the municipalities. Democratic Republic of Congo – SNEL There are numerous voltage levels and transformation configuration for the SNEL network which include at transmission, 220 kV and 132 kV, at subtransmission, 110 kV, 70 kV, 50 kV, and 15 kV. Distribution is at 11 kV, 6.6 kV and 0.4kV. With the data and information available it was assumed that the transformation will be 220, 70, 15, and 6.6 down to 0.4 kV. The 70 kV and 15 kV equipment and line cost were approximated to 66 kV and 11 kV respectively. Mozambique – EDM EDM has numerous voltage levels and transformation configuration for the network which included 400, 275, 330, 220, 110, 60, 33, 22, 11 6.6 and 5.5 kV. Some of the voltage levels are mainly for interconnection with other countries. There are two main transmission voltages, 220 kV and 110 kV, as well as 66 kV on the subtransmission and 33 kV, 11 kV and 0.4 kV on the distribution system. There are some 22, 6.6 and 5.5 kV network but the total line lengths are not significant. Namibia – NAMPOWER Nampower has two transmission voltages, 220 kV and 132 kV, subtransmission at 66 kV with 33kV, 22kV, 11 kV and 0.4 kV on the distribution system. Interconnection with ESKOM is at 400 kV and there is an odd 330 kV line. At distribution level there are various configurations, the data available indicated that 66 kV was a subtransmission voltage with significant lines and substations at 33 and 22 kV lines and for this study the higher voltage was used for the estimates. Zambia – ZESCO ZESCO has two transmission and subtransmission voltages 330 kV and 220 kV, 88 kV and 66 kV, with 33 kV, 11 kV and 0.4 kV on the distribution system. There are a few kilometres of line at 132 kV and this was not considered significant for the study. Group II – SAPP Country Utilities Botswana – BPC BPC has one transmission voltage of 220kV and subtransmission at 132 kV and 66 kV, with 33 kV, 11 kV and 0.4 kV on the distribution system. There are some 22 kV cross border supplies with ESKOM but these are very short lines no more than 5km per connection. Lesotho – LEC There is only one voltage at transmission level of 132 kV, and subtransmission level of 66 kV. Distribution is at 33 kV, 11 kV and 0.4 kV. There are a few kilometres of line at 88 kV. Malawi – ESCOM There is only one voltage at each level transmission 132 kV, subtransmission 66 kV, and distribution at 33 kV, 11 kV and 0.4 kV. 147 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Zimbabwe – ZESA From knowledge of the system, ZESA has only 330 kV as the transmission voltage with subtransmission at 132 and 88 kV. There are some lines at 66 kV but the total line length is not significant. Distribution is at 33 kV and 11 kV with the low voltage at 0.4 kV. WAPP: Country-specific assumptions and unit costs Group I – Western Africa Country Utilities There is very limited data and information on Western African Utility networks in the public domain or from specific projects in the region, mainly because most of them have been privatized. There are no countries in WAPP with two levels of transmission voltages. This may be due to low load densities in most countries except Nigeria, Ghana and Cote d’Ivoire. Generally the countries are small or they are on the fringes of the Sahara desert with sparse population. Group II – Western Africa Country Utilities All the countries in WAPP are Group II countries operating at different voltage levels.44 There is an interconnector from Nigeria, Benin, Togo at 161 kV, this continues into Ghana and Cote d’Ivoire at 225 kV. There is a separate 225 kV interconnector from Cote d’Ivoire to Mali and Burkina Faso, from Mali the interconnector goes to Senegal. Benin – CEB In CEB there is one voltage at each level, transmission at 161 kV and subtransmission at 63 kV. Distribution is at 20 kV, 15 kV and 0.4 kV. There is an interconnection with Nigeria and Togo at 161 kV. Burkina Faso – SNEB Burkina Faso is interconnected with Cote d’Ivoire at 225 kV, transmission is at 132 kV and 90 kV. Distribution is at 15 kV. Cote d’Ivoire – CIE In CIE there is one voltage at each level, transmission at 225 kV and subtransmission at 90 kV. Distribution is at 33 kV, 11 kV and 0.4 kV. There is an interconnection with Ghana (VRA) and Burkina Faso at 225 kV. Gambia – NAWEC The Gambia does not have a transmission network. Distribution is at 33 kV, 11 kV and 0.38 kV. The OMVG interconnector project is planned for 2004–2011 and will connect The Gambia to Senegal at 225 kV. 44 West Africa Regional Master Plan – Report June 2004 148 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Ghana – VRA & EGC In VRA & EGC there is one voltage at each level, transmission is at 161 kV and subtransmission is at 69 kV. Distribution is at 33 kV, 11 kV and 0.4 kV. There is an interconnection to Togo at 161 kV and at 225 kV with Cote d’Ivoire. Guinea – EDG Transmission in Guinea is at 110 kV with some Sub-transmission at 66 kV. Distribution is at 33 kV, 20 kV and 11 kV. Guinea is not presently interconnected, however the OMVG interconnector project is planned for 2004–2011 and will connect Guinea to Guinea-Bissau at 225 kV. There is also a planned interconnection as part of the LSG program to connect Guinea to Sierra Leone after 2011. Guinea-Bissau – Electricitie et Aguas Guinea-Bissau does not have a transmission network. Distribution is at 30 kV, 10 kV, 6 kV and 0.38 kV. Guinea-Bissau is not presently interconnected, however the OMVG interconnector project is planned for 2004–2011 and will connect Guinea-Bissau to Guinea and Senegal at 225 kV. Information about the lengths of the 30 kV and 10 kV systems have been provided by the World Bank. Liberia – LEC Liberia lost a large part of its electricity generation, transmission and distribution systems in the recent civil war. Prior to the war there was 69 kV subtransmission and 12.6 kV distribution. The system was 60 Hz, but has been converted to 50 Hz during the re-construction. The LSG program is planning the connection of Liberia to Sierra Leone and Cote d’Ivoire (East) after 2011. The lengths of the distribution system and LV system have been estimated from other similar countries based on the distribution length / net consumption of the country. Mali – EDM In Mali there is one voltage at each level, transmission at 161 kV and subtransmission at 63 kV. Distribution is at 20 kV, 15 and 0.4 kV. There is an interconnection with Nigeria and Togo at 225 kV. Mauritania – SOMELEC Mauritania is connected to Senegal at 225 kV as part of the OMVS project, there is a small amount of 90 kV system connecting generation in the south of the country. Distribution is at 33 kV, 15 kV and 0.4 kV. Niger – NIGELEC NIGELEC has 132 kV as the transmission voltage with subtransmission at 66 kV. Distribution is at 33 and 11 kV with the low voltage at 0.4 kV. NIGELEC has interconnections with NEPA at 132 kV. Nigeria – NEPA NEPA has 330 kV as the transmission voltage with subtransmission at 132 and 66 kV. Distribution is at 33 and 11 kV with the low voltage at 0.4 kV. NEPA has interconnections with NIGELEC at 132 kV and with Benin (CEB) at 161 kV. 149 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Senegal – SENELEC In Senegal there is only one voltage at each level, transmission is at 225 kV and subtransmission is at 90 kV. Distribution is at 30 kV, 6.6 kV and 0.4 kV. There is an interconnection with Mali at 225 kV. Sierra Leone – NPA Sierra Leone has some 161 kV transmission lines. Distribution is at 66 kV, 33 kV, 11 kV, 3.3 kV and 0.4 kV. There is presently no interconnection to other countries, however the LSG program is planning the connection of Sierra Leone to Liberia and Guinea after 2011. Information about the length of the transmission system was obtained from the World Bank. The lengths of the distribution system and LV system have been estimated from other similar countries based on the distribution length / net consumption of the country. Togo – CEET In Togo there is one voltage at each level, transmission at 161 kV and subtransmission at 63 kV. Distribution is at 20 kV, 15 kV and 0.4 kV. There is interconnection with Benin and Ghana at 161 kV. Island states: Country-specific assumptions and unit costs Cape Verde – ELECTRA Cape Verde is a collection of islands in the Atlantic Ocean west of Senegal. Information from the World Bank suggests there is a small amount of transmission system on the islands; this has been assumed to be 110 kV. Distribution is a mix of 20 kV, 15 kV, 13 kV, 6.6 kV, 6 kV and 0.38 kV. For the purposes of this analysis estimates have been made for 20 kV and 6.6 kV based on the application of typical ratios of distribution length to net consumption from other countries. Madagascar – LA JIRAMA The network has one voltage at the transmission and subtransmission level, 138 kV and 63 kV respectively, but at the distribution level there is 35, 20, 15 6.6 and 0.4 kV. Mauritius – CEB Mauritius has a transmission system at 132 kV and sub-transmission at 66 kV. Distribution is at 22 kV and 6.6 kV. There are no interconnections. Unit investment costs outputs and analysis The analysis has been split into the three distinct regions to differentiate between the type of terrain and climate. Another major aspect is the availability of materials and High Voltage contractors within the country. It was assumed that because of the volumes of work on the distribution systems and judging from the knowledge available for SAPP utilities, distribution work could be done in-house or by local contractors. 150 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Analysis of transmission unit costs SAPP countries Country unit investment costs indicate that by investing at the higher voltage, the unit investment costs are cheaper per MW of power transfer. Countries with similar voltages have comparable unit costs, the determining factor being the line rating. Except for ESKOM, investing at a lower voltage is more costly because of the lower ratings. This is the expected result because the structural and construction costs are not significantly reduced for the lower voltage and single conductor line construction. For ESKOM the line ratings for the lower voltage are significantly higher than those of other utilities due to enhanced design standards and this leads to almost the same unit costs for both voltages per MW of power transferred. For all countries except South Africa it is highly likely that foreign contractors will be awarded the construction of the lines because of lack of local expertise and construction equipment. This together with the higher cost of materials results in the unit investment cost for ZESA and ZESCO at 330 kV being higher than unit costs for ESKOM 275 kV lines. Further, ESKOM has an added advantage that most materials are locally manufactured at prices which are competitive compared with the international market; this, coupled with an abundance of local contractors for construction, tends to significantly reduce the costs. Central and Western Africa countries The unit costs follow the same pattern as described for SAPP and with no detailed data and information on the utilities, it was impossible to cite specific issues that would differentiate the individual utilities or compare with other regions. There is no knowledge of the existence of local manufacture of materials and neither equipment nor High Voltage construction contractors. The region has also very varied climatic conditions and terrain ranging from a rain forest and coastal countries (Cameroon and Nigeria) to arid desert (Mali and Chad). Eastern Africa – Nile Basin countries As highlighted for CAPP and WAPP countries, the unit costs follow the same pattern as described for SAPP. There is the same problem of no detailed data and information on the EAPP utilities. This made it impracticable to cite specific issues that would differentiate unit costs for the individual utilities or compare with other regions. There is also no knowledge of the existence of local manufacture of materials and neither equipment nor High Voltage construction contractors. The region also has very varied climatic conditions and terrain ranging from high rainfall region (Uganda and Rwanda) to arid desert (Sudan and Egypt). Analysis of distribution unit costs As highlighted before, the costs of distribution investments tend to be less installation specific and more generic than is the case for transmission networks; they are influenced by the choice of primary distribution voltage levels to a degree, and in the SAPP countries the voltages are fairly standard. 151 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA The major differentiating cost parameters associated with these standardized systems arise from whether networks are predominantly urban cabled systems, or whether they are largely rural systems utilizing overhead distribution lines. Urban networks are generally characterized by shorter distribution distances and higher costs per km of reticulation as compared with rural systems. Distribution network investment cost estimates can be derived from generic system models which are designed to deliver standardized quanta of load using network and equipment configurations which ensure appropriate levels of supply reliability, taking account of standard equipment ratings and circuit configurations. In the absence of specific data for many of the countries it is considered accurate enough for the purpose of this study to assume standard distribution costs for all the countries by extrapolating from the countries for which data is available. However, the distribution costs for South Africa will be slightly lower because of the advantage of some materials being manufactured locally. The margin will not be significant because the other countries have local supplies of poles, labor and other materials which are a major cost component for distribution systems. Refurbishment of existing structures Infrastructure life spans Infrastructure components have a life span ranging from approximately 10 years for mechanical metering (for electronic metering ~5 years) to 50 years for cables and transformers. The long life spans can render equipment obsolete before the end of its life due to changes in technology resulting in the non availability of spare parts. A benchmark planning average of 30 years has been used in this study as the life span which the transmission and distribution infrastructure can be refurbished to full functionality or extend the life by another equal number of years. The nature of refurbishment work is such that it normally requires the replacement and repair of transmission towers, the reconductoring of lines or the replacement of transformers and switchgear. The costs of carrying out these activities can be estimated as a percentage of the full installed costs of modern equivalent assets. From operating experience in the region, economic refurbishment of the infrastructure cost is normally limited not to exceed 60% of the replacement cost. If the refurbishment cost exceeds 60%, the usual recommendation is to replace with either like for like equipment or with new technology with enhanced features. Transmission and distribution infrastructure age data A fair amount of data regarding the age of the infrastructure was available for the transmission system for all the countries, which was used for estimating the refurbishment costs. The amount of equipment that would reach the limit of 30 years within the planning horizon to 2015 was estimated and used for the study. For distribution there was very little infrastructure age data, except for Namibia which was available from a project undertaken by the Consultant. 152 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Refurbishment factors The major influences on whether transmission infrastructure needs to be refurbished are associated with: • the age of the infrastructure; • the rate of deterioration of the assets, • due to normal wear and tear; • due to the effects of a particularly harsh operating environment (e.g. a coastal location or high levels of pollution); or • due to the operating and maintenance regime to which equipment has been exposed, e.g. frequent overloading due to rapid demand growth not being matched by adequate network investment, or the requirement to keep equipment in service in cases where continuity of supply is deemed to be more important than the need for preventative maintenance; and • the possible need for asset replacement with equipment of higher capacity, recognizing the effects of demand growth on issues such as energy losses or the need for expanded substation capacity within the constraints of existing sites. Distribution In order to assess the costs of distribution refurbishment, similar considerations apply to those described above for transmission. Distribution refurbishment costs are less dependent on the costs of specific projects, however, and can be assessed on a broader basis. There are two key options available, depending on the quality of the data which can be obtained from the utilities: • analysis can be performed with reference to the asset register for the distribution network, in order to identify the proportion of assets that are over a given age and which are therefore likely to require refurbishment over the next ten years, see figure below, or • analysis can be undertaken with reference to the overall levels of distribution capacity utilization and investment that can be projected for the future, extrapolating from historical records of investment against demand growth, see figure below. 153 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Figure A4.6 Optional approaches to distribution refurbishment evaluation The choice between these approaches is entirely dependent on the extent to which accurate information about the asset bases of the utilities is available. Neither method could be used for the systems under consideration, however, as there was no historical data on demand growth or age profiles for any of the countries except Namibia. The age data of the transmission system was therefore used to estimate the age of the distribution data. Transmission investments are normally lumped to cover a 5% demand growth over a twenty year horizon. With the growth in demand, there will be more investments in distribution infrastructure lagging the original transmission investment over the planning horizon. This result in a lower percentage of distribution equipment requiring refurbishment within the same planning horizon compared to transmission equipment. With the available transmission data and specific data for Namibia, the percentage equipment that would require refurbishment was estimated. The percentage will be slightly higher than reality because due to economic problems, forecast average load growth has not always been achieved. Further, investments in most of these countries are financed from loans which make the investments very discreet. Maintenance of existing and new infrastructure Overall approach The derivation of maintenance costs per km for transmission and distribution network assets is approached by allowances for maintenance as a percentage of the capital cost of new lines and substation equipment. This is normal practice for the purposes of assessing future network investment needs. A figure of 2% of the investment cost/km line is generally used for planning and this is the same level of O&M expenditure which are allowed for by the SAPP member utilities in planning transmission 154 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA system investments. However, to be conservative we will in the least cost expansion model use 5% of asset value, which is the same percentage that will be used at the distribution level. It was considered unnecessary to differentiate rates which are used for operating and maintaining equipment in harsh environments or in terrain where access is difficult because an average percentage cost has been used. Further for unit investment costs for new equipment it was considered not necessary to differentiate costs depending on the environment or accessibility. At the distribution level, there is much more activity because due to the nature of the distribution system and lack of automation there is more human interaction than for transmission systems. For distribution systems the rule of thumb is to estimate operations and maintenance costs to, on average, 5% of the investment cost/km line. Since this is an average cost for maintenance it is not necessary to distinguish between the maintenance costs of overhead lines that are close to the coast (and therefore subject to corrosion problems), and other lines. It will not be necessary to differentiate levels of maintenance cost/km across a broad range of line sizes and types since it is an average cost. Conclusion It can be concluded that the utilities that have been considered so far, excluding ESKOM, are highly likely to invest in transmission lines at their higher voltage which has a lower unit investment cost. For a particular transmission voltage the line rating can vary between countries, however, the spread of unit costs per MW is small and hence an average can be taken to derive a generic cost per MW for use where data is sparse. The data and costs available do not indicate an adverse cost implication on terrain and accessibility in the different countries; however, this could have become clearer if the country utilities had provided the requested information. For refurbishment, a rule of thumb of not exceeding 60% of the replacement cost is used in the study for transmission and distribution infrastructure. For operations and maintenance, the average cost for transmission and distribution is 5% of the cost/km of unit investment. Finally the assumptions, approximation and methodology adopted in the study yielded the desired outputs which are accurate enough for use to determine the investment needs given that there was very scant data for the infrastructure of the countries. 155 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Appendix 4 Comparison of Econ Pöyry cost estimates with those of Nexant study Here we compare our findings for the SAPP to those of a regional generation and transmission expansion plan drawn up by Nexant45 that assesses power sector investment requirements in SAPP for the period 2006–25.46 There are two fundamental differences between Nexant’s and Econ Pöyry’s analyses: methodological approach and scope of the costs included. In addition, assumptions about fuel prices, unit cost of investments, demand growth etc. are crucial to the numerical analyses and are therefore potential sources for divergent results. Here we compare the results along broad lines. Different methodologies: generic least cost model vs. costs related to existing plans Nexant’s analysis takes the existing generation expansion plans as a starting point and calculates the costs of specific transmission reinforcements that are needed to support the anticipated levels of trading. The Base Case is based on the existing generation and transmission plans for each of the 12 SAPP utilities. The generation plans are generally based on an assumption of moderate trade or self-sufficiency. Econ Pöyry’s least cost expansion model takes a more generic approach by estimating the costs of meeting a certain level of demand from the cheapest sources in the whole SAPP region. In other words, there is no requirement of national self-sufficiency. An implication of the different methodological approach is that the two studies estimate the costs of different investment volumes. Nexant’s study includes 93 projects or generic technologies. Many of the projects are also included in our analysis.47 However, in our least cost approach, existing plans for generation capacity expansion are included as potential investments, among which the model chooses the least cost solution. Plants where the construction has already started or where the investment decision has been made are included as exogenous investment. Projects that are characterized as “very close to financial closure” are included as potentials: based on experience, projects can stay at this stage for a long time. Hence, Econ Pöyry’s least cost approach enables an assessment and comparison of economic profitability of different projects. Differences in the treatment of trade Nexant compares two cases: 45 Nexant (2007): “SAPP Regional Generation and Transmission Expansion Plan Study.” October 46 Nexant’s study includes Tanzania, while Tanzania is included in EAPP in Econ Pöyry’s analysis. 47 However, since our study of SAPP was completed before Nexant’s report was published in October 2007, we were not able to use inputs from the report in our work. 156 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA • A base case based on the existing generation and transmission plans for each of the 12 SAPP utilities. This scenario has a national focus, with most TSOs trying to cover domestic demand from domestic generation sources. • An alternative case that considers various scenarios for the optimization of generation and transmission capacity additions assuming free trade, no constraints on the expansion of the interconnecting lines, and removal of the constraints within the utilities internal networks. Therefore, Nexant’s alternative case is methodologically more comparable to Econ Pöyry’s study, while its base case is more comparable to our trade stagnation scenario. It should be noted, however, that Econ Pöyry’s trade stagnation scenario limits the development of new international transmission lines, but trade within the limits of existing transmission capacity is allowed. Hence, the trade stagnation scenario does not take national self-sufficiency as its starting point (as in Nexant’s base case), but allows for exploiting the cheapest resources in the region within the limits set by existing transmission lines. In some cases, the trade flows in the opposite direction compared to today’s trade flows. Therefore, the resulting trade flows in the trade stagnation scenario do not necessarily coincide with the anticipated trade flows or levels in Nexant’s base case. Different scope: transmission, distribution and connection costs The focus of Nexant’s study is on transmission reinforcements that are required to accommodate the anticipated trade levels. The transmission costs appear to be a small portion of total expansion costs, ”Transmission interconnection costs are a very small component of overall system costs, indicating that utility expansion plans are mostly based on self sufficiency requiring few new transmission interconnection lines” (p. 2-2 of Executive Summary). Lower-level (e.g., distribution) grid costs are not included in Nexant’s study, as far as we can see. These lower-level grid costs make up a substantial part of the estimated system expansion costs in Econ Pöyry’s analysis. Grid costs include O&M costs during the ten-year period, investments in distribution grid that are necessary to meet demand growth due to economic growth and increased electricity access due to new connections. Finally, grid costs also include transmission grid expansion. Since Econ Pöyry’s analysis includes many additional and large grid costs that Nexant does not include, total grid costs in the Econ study make up a large share of total expansion cost. This is in contrast to the small share claimed by Nexant. Comparison of results Generation costs Nexant’s base case estimates the overnight investment costs in generation to be 25.9 billion USD during the 10-year period of 2006–2015. In the alternative case, the overnight generation investment costs are only 13.7 billion, indicating a more efficient exploitation of the region’s resources. 157 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA Econ Pöyry’s cost estimate is higher: 32 billion by 2015 (Trade expansion with national targets for electricity access). In addition, 7.5 billion are required for refurbishment of the existing plants. It is unclear whether the refurbishment requirement is taken into account in Nexant’s analysis. The following differences in assumptions explain the diverging cost estimates. Demand. This large difference in cost estimates is largely explained by different investment volume: 21,600 MW are invested in Nexant’s base case and 19,000 MW in the alternative case, while 33,300 MW are added in Econ Pöyry’s analysis. This difference in investment requirements is partly due to different assumptions about demand: • Nexant (gross demand?): 386 TWh (includes Tanzania, ca. 5 TWh) • Econ Pöyry’s gross demand: 438.9 TWh, net demand 396 TWh (without Tanzania) Our demand assumptions take into account demand growth due to economic growth and social demand (increased electricity access). In addition, the suppressed demand is assumed eliminated. Admittedly the targets for increased electricity access are quite ambitious. We haven’t examined the reasoning behind the Nexant estimate. Unit cost assumptions. As Nexant’s study is based on existing plans, obtained from project developers, it is possible that the unit costs of investments are more precise than Econ Pöyry’s assumptions (that are based both on actual project costs where available and on generic costs of different technologies). Cost estimates can differ due to different assumptions about discount rate and life-time of different assets. It is worth noting that Econ Pöyry’s investment cost figures include costs related to off-grid technologies as well (associated with increased electricity access in remote areas), which are quite expensive. Fuel costs. Econ Pöyry’s variable production costs (fuel and O&M) are higher in total. This is an obvious result, given the higher demand. Another source of difference may be assumptions about fuel costs. We haven’t examined which fuel (coal and natural gas) costs are assumed in the Nexant’s study. Econ finds, however, that there are large savings of variable costs with trade: the fuel and O&M costs are almost 20 percent lower in trade expansion scenario. In contrast, Nexant estimates these cost savings to be less than 4 percent. A possible explanation is that Econ Pöyry’s analysis finds greater switch from thermal into hydro, which in turn is related to the factors mentioned above: different attitude to exogenous investment and different interpretation of what is meant by trade expansion. Network costs: transmission and distribution The big difference in the results is, however, in the costs of network expansion. As explained above, this is due to different scope of the costs included in the analysis: Distribution costs are omitted in Nexant’s analysis. Comparing the transmission costs only, the results are quite similar: the alternative case transmission investments are almost 4 billion. The comparable figure in Econ Pöyry’s analysis (investments in cross- 158 COSTING POWER INFRASTRUCTURE SPENDING NEEDS IN SUB-SAHARAN AFRICA border grid) is slightly more than 3 billion. In Nexant’s base case the transmission investment costs are very low (due to focus on self-sufficiency). This is similar to Econ Pöyry’s Trade stagnation scenario. It is reasonable to assume that Nexant’s study is based on more detailed information about specific transmission lines and reinforcement needs; hence, their estimates may be more accurate for some connections. 159