WPS5904 Policy Research Working Paper 5904 Power Tariffs Caught between Cost Recovery and Affordability Cecilia Briceño-Garmendia Maria Shkaratan The World Bank Africa Region Sustainable Development Unit December 2011 Policy Research Working Paper 5904 Abstract This is the first paper to build a comprehensive empirical percent of countries and hence provide efficient pricing picture of power pricing practices across Sub-Saharan signals. Africa, based on a new database of tariff structures in 27 As regards affordability, today’s average effective tariffs countries for the years 2004–2008. are affordable for 90 percent of today’s customers. Using a variety of quantitative indicators, the paper However, they would only be affordable for 25 percent evaluates the performance of electricity tariffs against of households that remain unconnected to the grid. four key policy objectives: recovery of historic power Tariffs consistent with full recovery of economic costs production costs, efficient signaling of future power would be affordable for 70 percent of the population. As production costs, affordability to low income households, regards equity, the highly regressive patterns of access to and distributional equity. power services, ensure that subsidies delivered through As regards cost recovery, 80 percent of the countries electricity tariffs are without exception also highly in the sample fully recover operating costs, while only regressive in distributional incidence. around 30 percent of the countries are practicing full The conclusion is that achieving all four of these policy recovery of capital costs. However, due to the fact that objectives simultaneously is almost impossible in the future power development may be based on a shift context of the high-cost low-income environment that toward more economic technologies than those available characterizes much of SSA today. Hence most countries in the past, existing tariffs look as though they would be find themselves caught between cost recovery and consistent with Long Run Marginal Costs in nearly 40 affordability. This paper is a product of the Sustainable Development Unit, the Africa Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at cbricenogarmendi@worldbank.org and mshkaratan@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team POWER TARIFFS: CAUGHT BETWEEN COST RECOVERY AND AFFORDABILITY CECILIA BRICEÑO-GARMENDIA AND MARIA SHKARATAN Acknowledgments This paper draws on contributions from sector specialists from the Africa Infrastructure Country Diagnostic Team; notably, Daniel Camos, Anton Eberhard, Vivien Foster, Sudeshna Ghosh Banerjee, Jean-Noel Gogua, Fatimata Ouedraogo, Orvika Rosnes, Haakon Vennemo, and Quentin Wodon. It is based on data collected by local consultants and benefited greatly from feedback provided by colleagues in Africa Energy Department: Fabrice Bertholet, Luiz Maurer, and Fanny Missfeldt-Ringius. 2 Power tariffs: Caught between cost recovery and affordability1 The efficient pricing of electricity is central to a well-functioning power sector. Power pricing guides investment decisions and is critical for cost recovery. It also signals to users the cost of marginal consumption and should ideally encourage the optimal utilization of installed capacity. But achieving efficient power pricing is easier said than done. The power sector is characterized by substantive up- front fixed costs, and it takes many years for capacity to be fully utilized. Beyond that, costs vary across times of the day (peak/off-peak), seasons (dry/rainy), users (residential/commercial), and geographic areas (urban/rural), which should be taken into consideration when setting prices that promote efficient use. As if the technical issues behind setting efficient tariffs were not complex enough, power providers and regulators also face a conflict between promoting economic efficiency and societal well-being. For example, if income-challenged groups are to enjoy the benefits of power provision, policy makers must set affordable tariffs below production costs or introduce an explicit subsidy regime (Borenstein 2008). In defining tariff structures, policy makers must balance the financial sustainability of the sector on the one hand and the well-being of various segments of society on the other. Given the importance of power, the ramifications of pricing and bill-collection policies are enormous. For example, as imposed transfers from the producer to the consumer, below-cost tariffs can seriously hamper the financial health of the provider. Another common way of lowering electricity prices is cross-subsidization, which can only be implemented if monopoly rights are granted to the power utility. Cross-subsidization has several undesirable consequences: it discourages use by the overcharged and promotes overconsumption by the subsidized. In some cases, it also opens the door for particular interest groups and communities to influence policy makers, for example, by asking them to reduce tariffs for select customers such as large industrial users. While this may be used as a mechanism to spur the development of select economic sectors, the reduced tariffs, ironically, might not even be made available to the general public with its more limited purchasing power. How to get tariffs right is a critical question for every policy maker, and there is no one answer. In this paper we aim to better understand how African countries are dealing with these pricing issues in practice. Most African countries have made efforts to organize their tariff structures and levels so as to recover utility costs while also providing affordable electricity to poorer consumers. But this goal is challenging and has not been reached in most of the countries examined. Obstacles include costly operational inefficiencies, lack of economies of scale due to geopolitical fragmentation, large populations too poor to afford tariffs set at cost-recovery levels, and the dauntingly limited coverage of distribution networks. 1 Comments should be addressed to Cecilia Briceño-Garmendia (cbricenogarmendi@worldbank.org) and Maria Shkaratan (mshkaratan@worldbank.org). 3 The analysis presented here is based on a database put together as part of the Africa Infrastructure Country Diagnostic (AICD).2 The database includes the basic institutional characteristics of African power systems as well as standard power sector indicators (performance, capacity, and so on). In addition the database documents the power-tariff regimes of 27 Sub-Saharan African countries in detail.3 Together, these nations account for over 85 percent of the population and gross domestic product (GDP) of the region. They were carefully selected to represent the economic, geographic, cultural, and political diversity that characterizes Sub-Saharan Africa. They also represent the four Sub-Saharan power pools; include countries with small, medium, and large-scale generation; and constitute a representative mix of predominantly thermal and/or hydro power systems. As such, the sample can be considered a statistically representative basis for inferring tariff-setting trends in Sub-Saharan Africa. The tariff data set includes one published tariff regime for each country in the sample. Results presented here are based on the latest published tariff regime available for each country during the AICD data collection period (2003–08). For most countries that year was 2006 (see annex 1). Using the AICD database, we seek to characterize African power tariffs by (i) describing prevalent tariff structures, both residential and nonresidential; (ii) analyzing their ability to recover costs; (iii) assessing their economic efficiency against long-run marginal costs; and (iv) exploring their equitability and affordability vis-à-vis country-specific purchasing power. What power tariff structures are prevalent in Sub-Saharan Africa? Most electricity tariffs—and Africa is no exception—are based on block tariff-pricing schemes; that is, the price of power is linked to the level of consumption. Power tariffs are commonly structured around blocks. A block is a pre-determined range of power consumption; with the unit price of each kWh being fixed within the block. The relation between blocks and prices defines three types of tariff structures:  Increasing block tariffs (IBTs) is a regime in which the unit price per kWh follows an increasing step-function linked to sequentially defined blocks  Decreasing block tariffs (DBTs) is a regime in which the unit price per kWh follows a decreasing step-function linked to sequentially defined blocks  Linear tariffs (LTs) are a regime in which all units of power consumed are charged at exactly the same rate. http://www.infrastructureafrica.org. 2 3 Benin, Botswana, Burkina Faso, Cameroon, Cape Verde, Chad, Congo, the Democratic Republic of Congo, Côte d’Ivoire, Ethiopia, Ghana, Kenya, Lesotho, Madagascar, Malawi, Mali, Mozambique, Namibia, Niger, Nigeria, Rwanda, Senegal, South Africa, Tanzania, Uganda, Zambia, and Zimbabwe. 4 Any of the above tariff structures may be complemented by a fixed monthly charge, and are then described as two-part electricity tariffs. The fixed charge can be used to cover the fixed administrative costs associated with serving a customer, and is sometimes also used to discriminate between customers based on other cost-related characteristics such as load served and network location. a. Residential tariffs Two-thirds of the prevailing pricing schemes in Africa are IBTs, and the remaining third are linear (figure 1a). The use of linear rates is more common in countries with prepayment systems (Malawi, Mozambique, and South Africa) (table 1). The prevalence of IBTs is consistent with recent trends in power regulation. IBTs have often been put forward as a good tool for reconciling cost-recovery targets with distributional aims, although their success in doing so is critically dependent of the details of tariff design (Filipovid and Tanid 2009; Borenstein 2008). A more detailed description on residential tariff structures practiced in Africa can be found in Annex 2. Figure 1. Prevalence of specific tariff schemes a. Block tariff pricing b. Fixed charges 70% 60% 60% percentage of prevailing percentage of prevailing 50% 50% 40% 40% schemes schemes 30% 30% 20% 20% 10% 10% 0% 0% Increasing Block Linear Rate Fixed No Fixed Fixed No Fixed Charge Charge Charge Charge (Linear) (Linear) (IBT) (IBT) tariff schemes tariff schemes Source: Africa Infrastructure Country Diagnostic Power Tariff Database. About half of the sample countries have adopted two-part tariffs (figure 1b), combining fixed charges with block tariff pricing. Among countries practicing linear tariffs, the use of two-part tariffs is more common than where IBTs are applied. For African countries the fixed charge tends to be relatively high: between $1.00 and $3.00 per month (table 1). As a reference, the average fixed charge in Latin American countries is $0.70 (Foster and Yepes 2006). 5 Table 1. Residential tariff schedules Tariff Fixed charge Fixed Number Size of Price, first Price, % increase type per month? charge of blocks the first block highest from first Yes/No ($/month) block, block block to kWh highest block Benin IBT No — 3 20 9.6 16.3 70 Botswana Linear Yes 1.63 1 — 5.9 5.9 — Burkina Faso IBT Yes 1.1 3 50 18.4 20.8 13 Cameroon IBT No — 3 50 8.6 12 40 Cape Verde IBT No — 2 40 22.5 28 24 Chad IBT No — 3 30 15.7 38.1 143 Congo, Dem. IBT No 11 100 3.98 8.52 114 Rep. of Congo, Rep. of Linear Yes 5.06 1 — 11 11 — Côte d’Ivoire IBT Yes 0.64 2 40 6.9 14.2 106 Ethiopia IBT Yes 0.16–1.58 7 50 3.2 8 150 Ghana IBT Yes 0.54 3 300 7.6 15.3 101 Kenya IBT Yes 1.74 4 50 4.9 44 798 Lesotho Linear No — 1 — 7.2 7.2 — Madagascar Linear Yes 2.98 1 — 7.6 7.6 — Malawi IBT Yes 0.92 3 30 2 4.1 105 MWI–prepaid Linear No — 1 — 3.1 3.1 — Mali IBT No — 4 200 26.6 31 17 Mozambique IBT Yes 2.79 4 100 4 12.1 203 MOZ–prepaid Linear No — 1 — 11 11 — Namibia Linear No — 1 — 11.7 11.7 — Niger Linear Yes 0.43 1 — 13.6 13.6 — Nigeria IBT Yes 0.15–2,38 5 20 0.9 6.5 622 Rwanda Linear No — 1 — 14.6 14.6 Senegal IBT No — 3 — 23.8 26.2 10 South Africa IBT No — 2 50 0 7.2 — Tanzania Linear Yes 1.93 1 — 10.8 10.8 — TZN–low use IBT No — 2 50 4.1 13 217 Uganda IBT Yes 1.1 2 15 3.4 23.3 585 Zambia IBT Yes 1.3 3 300 1.6 3.7 131 Zimbabwe IBT No — 3 50 0.60 13.5 2,159 Source: Africa Infrastructure Country Diagnostic Power Tariff Database Note: For details see annex 2. MWI = Malawi; MOZ = Mozambique; TZN = Tanzania. — Not available. The structure of blocks—their number, size, and respective price levels—is what ultimately defines how tariffs reflect costs, affect demand, and address equity issues. The sample countries have adopted widely differing approaches to tariff design, with one third of countries practicing linear tariffs (as in Rwanda and Tanzania) and the remaining two thirds practicing IBTs. Looking at the two thirds of sample countries that practice IBTs, most have adopted relatively simple structures (figure 2a) with two or three blocks (as in Benin, Burkina Faso, Cameroon, Cape Verde, Côte d’Ivoire, Ghana, Madagascar, Malawi, Senegal, Tanzania, Uganda, Zambia). A minority of countries have opted for more complicated structures that use four and more blocks (the Democratic Republic of 6 Congo, Ethiopia, Kenya, Mali, Mozambique, and Nigeria). The Democratic Republic of Congo is an extreme case, as it uses an 11-block IBT. Interestingly enough, despite the large number of blocks, the price difference across blocks is very small, suggesting that the regime is complicated without really discriminating very much between large and small consumers. Figure 2. Characterization of residential block pricing a. Number of pricing blocks b. Tariff price differential between first and highest block (%) 12 12 10 10 number of countries number of countries 8 8 6 6 4 4 2 2 0 0 one (linear two three four or more 0% < 100% 100-200 200% > rate) number of blocks Source: Africa Infrastructure Country Diagnostic Power Tariff Database The conventional wisdom is that IBTs are designed so that the first and smallest “lifeline� block covering subsistence consumption is subsidized to promote equity, while the subsequent blocks are priced at a higher level that will ultimately enable cost recovery. This of course assumes that poorer customers will have lower consumption. This assumption is more reasonable in the case of power – where usage is driven by ownership of appliances more prevalent among wealthier households and businesses – than it is for example in the water sector – where usage is more correlated with the size of the household. The first question is whether African countries tend to define the consumption level of the first block at a level low enough to be consistent with subsistence consumption and hence with the “lifeline� principle underlying this tariff design (table 2). The response to this question is quite positive. Two-thirds of the sample countries define the first block at 50 kilowatt-hours (kWh)/month or less; a consumption level that is below the average residential power consumption in Africa (75-100 kWh/month) (Foster and Briceño-Garmendia 2009). For example, Uganda sets the first block at 15 kWh/month; Cape Verde and Côte d’Ivoire at 40 kWh/month; and Burkina Faso, Cameroon, Ethiopia, Kenya, Tanzania at 50 kWh/month. But the sizes of the first blocks in Mozambique and the Democratic Republic of Congo (100 kWh), Mali (200 kWh) and in Zambia and Ghana (300 kWh) seem too high to meet the needs of low- consumption, low-income residential customers. 7 Table 2. Size of the first pricing block, and differential between that and second 1st block monthly Countries Price differential between first and consumption threshold second block (%) (kWh) 300 Ghana, Zambia Between 30 and 100 200 Mali Less than 30 100 Mozambique Over 100 Congo, Dem. Rep of Less than 10 50 or less Uganda, Madagascar, Kenya, Nigeria, Tanzania, Over 100 Chad, Cote d’Ivoire, South Africa, Zimbabwe Ethiopia, Benin, Malawi Between 30 and 100 Cameroon, Cape Verde, Burkina Faso Less than 30 Source: Africa Infrastructure Country Diagnostic Power Tariff Database The second question is whether tariff levels on successive blocks rise steeply enough to ensure that costs can be fully recovered on higher volumes of consumption. This principle holds in many cases. Most countries with a first-block threshold set at subsistence consumption levels (50 kWh or less), have a price jump of over 100 percent to the second block (Uganda, Madagascar, Kenya, Nigeria, Tanzania, Chad, Côte d’Ivoire, South Africa, Zimbabwe), signaling a clear intention to differentiate among customers so that larger consumers contribute to cost recovery. But the tariff structures of a few countries—for example, the Democratic Republic of Congo, Cameroon, Cape Verde, and Burkina Faso— are not highly differentiated by consumption level. This makes the implementation of targeted subsidies difficult (table 2). Thereafter, prices among higher consumption blocks do not rise as steeply as they do between the first and second block. In Burkina Faso, the second block is priced only 6 percent higher than the first one and the third block only 7 percent above the second one. In Ethiopia the price increase from the first block to the second is 31 percent and from the second to the third, 40 percent. In Malawi, the increase is 47 percent in the first case and 42 percent in the second one. Despite the fact that IBTs already incorporate a “lifeline� principle, a number of countries have felt the need to introduce parallel “social tariff� that provide an even larger discount to qualifying customers. For example, the Democratic Republic of Congo, Madagascar, Mali, and Benin (whose block pricing structure does not differentiate low-consumption, low-income customers) provide such a parallel “social tariff� (table 3). Most of these social tariffs are linear in nature and include fixed charges; albeit modest ones. The criteria for determining social tariff eligibility are usually based either on total consumption or technical characteristics of service (voltage, load). The very existence of such parallel “social tariffs� – most of which have eligibility criteria based on low consumption – in itself may be a signal that the main IBT is not managing to perform its intended social function. 8 Table 3. Tariff schedules for low-income, low-consumption customers Fixed charge Price per block, Type of tariff ($)/month Block border cents/kWh Benin social tranche n.a. 9.6 Botswana n.a. n.a. n.a. Burkina Faso block 1, residential 0.18 14.3 Cameroon* block 1 residential 12.90 8.6 Cape Verde block 1, residential — 22.5 Chad block 1 residential n.a. 15.7 Congo, Dem. Rep. of social tariff 0.01 4.0 Congo, Rep. of n.a. n.a. n.a. Côte d’Ivoire block 1 residential 0.64 6.9 Ethiopia block 1 residential 0.16 3.2 Ghana block 1 residential 0.54 7.6 Kenya block 1 residential 1.74 4.9 Lesotho — — — Madagascar economic tariff 0.30 25 6.0 > 25 27.6 Malawi block 1 residential 0.92 2.0 Mali social tariff n.a. 50 13.2 100 20.3 200 23.9 >200 27.7 Mozambique block 1 residential n.a. 4.0 Namibia n.a. n.a. n.a. Niger — — — Nigeria pensioners’ tariff 0.23 3.0 Rwanda — — — Senegal tranche 1 residential n.a. 150 0.24 South Africa block 1 residential n.a. — Sudan — — — Tanzania n.a. n.a. 3.0 Uganda block 1 residential 1.09 3.4 Zambia block 1 residential 1.31 1.6 Zimbabwe tranche 1 residential n.a. 0.6 Source: Africa Infrastructure Country Diagnostic Power Tariff Database Note: For details see annex 13. n.a. Not applicable. — Not available. Another important aspect of social policy in the power sector has been the growing use of prepayment meters. These have dual advantages. On the one hand, they help low income customers to control their expenditures on electricity by spreading them out into small frequent payments that they can more readily afford. On the other hand, they eliminate the commercial risk to the utility of serving low income 9 customers. Although growing in popularity worldwide, providing a prepaid option is a relatively new practice in Africa, and its impact is still not well understood. But for those countries for which data are available, roughly 60 percent have introduced the use of prepayment meters (figure 3). Indeed, Lesotho, Namibia, Mozambique, and South Africa already offer this option to a majority of residential customers. Figure 3. Prepaid services a. Frequency distribution of countries adopting b. Time trend in percentage of customers using prepaid meters prepaid meters 10 2000 2001 2002 2003 2004 2005 9 Tanzania 4 4 number of countries 8 7 Benin 0 0 6 6 6 7 6 Burkina Faso 10 10 10 10 10 10 5 Malawi 2 4 6 8 10 14 4 3 Ghana 5 4 4 10 28 2 Rwanda 48 65 1 0 Mozambique 80 73 0 1-20 21-50 51-90 91-100 no data Namibia 93 98 Lesotho 100 100 100 100 percentage of customers with prepaid meters South Africa 100 Source: Africa Infrastructure Country Diagnostic Power Indicators Database b. Nonresidential tariffs Nonresidential tariffs can be classified in two groups: commercial and industrial (Table 4). Linear tariffs are the most common regime for nonresidential customers in Africa (figure 4). About 60 percent of commercial customers and over 50 percent of industrial customers are billed based on linear tariffs. A more detailed description on commercial and industrial tariff structures practiced in Africa can be found in Annexes 3 and 4. Pricing for nonresidential customers is typically more complex than for residential customers. It is usually structured as a three-part tariff including a monthly fixed charge (defined by characteristics of the network), a demand charge (defined by the level of peak demand served in kilovolts or kilowatts), and a volume charge (defined by the energy served and reflected in the definition of the blocks). In addition, volume charges may be differentiated by time-of-use (TOU). In fact, only a handful of countries (Benin, Cape Verde, Rwanda, and Mali) apply simple linear tariffs to their nonresidential customers without making use of any of these additional features. Fixed charges are somewhat more prevalent among non-residential customers, than was the case for residential customers. They are practiced by 18-19 countries in the sample. 10 Demand charges are widespread for non-residential customers, but are almost twice as frequent for industrial customers as for commercial customers (figure 4b). This suggests that most countries find it important to reflect load considerations in designing tariffs for nonresidential customers (table 3). Peak demand is a critical cost driver in the power sector, because it defines the amount of installed capacity needed to provide a given volume of electricity. Figure 4. Prevalence of specific tariff schemes: Commercial and industrial users a. Tariff type b. Fixed charges 20 20 number of countries number of countries 15 15 10 10 5 5 0 0 FR TOU IBT DBT None Only fixed Only Demand demand and fixed commercial industrial commercial industrial Source: Africa Infrastructure Country Diagnostic Power Tariff Database Figure 5. Characterization of nonresidential block pricing a. Numbers of pricing blocks b. Nonlinear tariff price differential between first and last block (%) 20 6 number of countries number of countries 5 15 4 10 3 2 5 1 0 0 One Two Three 4 or more negative no change 0-100% greater than 100% Commercial Industrial Commercial Industrial Source: Africa Infrastructure Country Diagnostic Power Tariff Database Volume charges for nonresidential customers are typically linear for at least 15 of the countries in the sample. In the cases where block tariff structures are used, they are more frequently decreasing blocks (DBT) rather than increasing blocks (IBT) and these are relatively simple with not more than three blocks (figure 5b). The reason for preferring decreasing blocks is to capture the strong scale economies associated with power generation and transmission. The volumetric charge tends to be higher for commercial than for industrial customers. Most commercial tariffs start at over 12 cents/kWh, while most industrial tariffs start at around 8–10 cents/kWh. On average, African commercial tariffs are about 40 percent higher than industrial ones. Only the Democratic Republic of Congo, the Republic of Congo, and Chad have commercial tariffs set at a lower level than residential ones (table 5). 11 Table 4. Nonresidential tariff regimes Commercial Industrial Fixed Demand Fixed Demand charge/month? charge? charge/month? charge? Tariff type yes/no yes/no Tariff type yes/no yes/no Benin Linear No No Linear No No Botswana Linear No No Linear Yes Yes Burkina Faso TOU Yes Yes TOU Yes Yes Cameroon DBT No Yes TOU No Yes Cape Verde Linear No No Linear No No Chad IBT No Yes TOU No Yes Congo, Dem. DBT No No DBT No No Rep. of Congo, Rep. of Linear Yes No Linear Yes Yes Côte d’Ivoire DBT Yes No TOU Yes No Ethiopia TOU Yes No TOU Yes No Ghana IBT Yes No Linear Yes Yes Kenya Linear Yes No DBT Yes Yes Lesotho Linear No Yes Linear No Yes Madagascar Linear Yes Yes Linear Yes Yes Malawi Linear Yes Yes Linear Yes Yes Mali Linear No No Mozambique Linear Yes Yes Linear Yes Yes Namibia Linear Yes Yes Linear Yes Yes Niger Linear Yes Yes Linear Yes Yes Nigeria IBT Yes No IBT Yes Yes Rwanda Linear No No Linear No No Senegal TOU Yes No TOU Yes No South Africa IBT/Linear Yes No TOU Yes Yes Tanzania Linear Yes Yes Linear Yes Yes Uganda TOU Yes No TOU Yes Yes Zambia Linear Yes No DBT Yes Yes Zimbabwe Linear Yes Yes Linear Yes Yes Source: Africa Infrastructure Country Diagnostic Power Tariff Database Note: IBT – increasing block tariff; DBT – decreasing block tariffs; TOU – time of use tariff. Source: Africa Infrastructure Country Diagnostic Power Tariff Database Time-of-use tariffs (TOUs) are only practiced in a minority of cases, and are twice as prevalent among industrial tariff structures as among commercial tariff structures. TOUs allow power consumption to be associated with hours (peak/off-peak/night) and/or seasons (summer/winter, dry/regular), creating incentives for more efficient use of the power network. One-third of industrial and one-fifth of commercial tariff regimes in Africa incorporate TOUs. 12 Table 5. Nonresidential blocks: Number, size, and price level Commercial Industrial First block First block Price Price differential differential Number between first Number between first of Price and last block of Price and last block blocks Size (cents) (%) blocks Size (cents) (%) Benin 1 n.a. 15.1 — 1 n.a. 10.7 — Botswana 1 n.a. 6.7 — 1 n.a. 3.1 — Burkina Faso 2 TOU 31.6 –47 2 TOU 22.6 –54 Cameroon 2 180kWh/kVA 11.3 –12 2 TOU 8.7 –2 Cape Verde 1 n.a. 21.8 — 1 n.a. 17.7 — Chad 3 30 kWh 15.9 152 3 TOU 20.5 85 Congo, Dem. 200 kWh 11.1 –4 15.2 Rep. of 5 5 200 -4 Congo, Rep. 1 n.a. 9.7 — 1 n.a. 11.2 — 18/kVA 18.6 –15 10.7 Côte d’Ivoire 2 bimonthly 3 TOU –18 Ethiopia 3 TOU 6.7 –6 3 TOU 4.7 26 Ghana 3 300 11.1 44 1 n.a. 5.4 — n.a. 21.4 — 20 16.4 Kenya 1 3 kWA –15 Lesotho 1 n.a. 1.2 — 1 n.a. 1.1 — Madagascar 1 n.a. 16.9 — 1 n.a. 9.9 — Malawi 1 n.a. 3 — 1 n.a. 2.4 — Mali 1 n.a. 23.2 — 0 n.a. — Mozambique 1 n.a. 5.4 — 1 n.a. 4.5 — Namibia 1 n.a. 8.4 — 1 n.a. 7.7 — Niger 1 n.a. 12.2 — 1 n.a. 8.8 — 15 kVA 5 30 15 5 Nigeria 4 5 kVA 30 Rwanda 1 n.a. 17.2 — 1 n.a. 17.2 — Senegal 2 TOU 14.4 44 2 TOU 11.8 45 South Africa 3/1 25 kVA 4 138 2 TOU 2.6 –31 Tanzania 1 n.a. 5.3 — 1 n.a. 4.9 — Uganda 1 n.a. 21.8 — 1 n.a. 16.7 — Zambia 1 n.a. 3.7 — 4 1200 2.2 –45 Zimbabwe 1 n.a. — 1 n.a. — Source: Africa Infrastructure Country Diagnostic Power Tariff Database Notes: For details see Annexes 3 and 4. n.a. Not applicable; — Not available; kVa = kilovolt-ampere. 13 Do power tariffs recover costs? a. What is the level of average effective residential tariffs? The effective tariff is the price per kilowatt-hour of electricity consumed at a specific consumption level when all charges—variable and fixed—are taken into account. In a multipart tariff system with a block pricing scheme, the difference between the effective tariff for a kilowatt-hour at a low level of consumption and a kilowatt-hour at a high level of consumption can be significant. The effective tariff is calculated by dividing the total bill based on the current tariff schedule (in currency) by consumption (in kilowatt-hours). It can also be referred to as the unit price at a particular consumption level. Based on the two-part tariffs described in the preceding section, effective residential electricity tariffs can be estimated following the formula: T  ( ai 1  xi 1  b) / x , where a is the n n metered consumption unit price (per kilowatt-hour), x is the volume consumed (metered), i is the block number (in the case of block tariffs), and b is the fixed charge.4 Effective tariffs allow for analyzing pricing patterns at different consumption levels as well as comparing price levels with cost recovery and affordability benchmarks. For the purpose of this analysis, one residential schedule per country was selected. The selection of a specific tariff schedule for the calculation of the effective tariff was done to capture the largest share of residential consumers based on the most commonly used tariff schedule, or the one that most closely corresponds to the monthly average consumption for that country. For example, in South Africa, the residential schedule selected was the “Home Light 1� prepayment option because this is the one that would be most attractive to a South African household with the average residential power consumption level. While South Africa has an admittedly complicated tariff system, other African countries also offer two or three residential schedules. For effective residential tariff calculations, the lower-usage residential tariff was used and the “social tariff� (where relevant) was analyzed separately. Annex 5 lists the representative tariff schedule used for each country. Table 6 showcases the calculation of effective residential tariffs for select African countries. The variation in effective tariffs for the first consumption tranche (up to 50 kWh/month) is enormous, going from zero in South Africa to about 24 cents/kWh in Cape Verde (table 6). But over 60 percent of African countries establish prices below 10 cents/kWh for their smallest consumers. For average consumption levels of 100 kWh/month, Africa has effective tariffs ranging from 3 cents to 30 cents, which is undoubtedly a wide range. 4 Generally there is a third component in the formula that captures maximum demand level and its price. But the demand part of the formula is not applicable to residential customers 14 Table 6. Effective residential tariffs by level of consumption (cents) Level of consumption (kWh/month) 50 75 100 150 200 300 400 450 500 900 Benin 12.6 13.3 13.6 14.0 14.1 19.7 22.5 23.5 24.2 27.2 Botswana 9.1 8.0 7.5 6.9 6.7 6.4 6.3 6.2 6.2 6.0 Burkina Faso 20.6 20.2 20.0 19.9 19.8 20.1 20.3 20.4 20.4 20.6 Cameroon 8.6 10.9 10.9 10.9 10.9 12.0 12.0 12.0 12.0 12.0 Cape Verde 23.6 25.1 25.8 26.5 26.9 27.3 27.4 27.5 27.5 27.7 Chad 22.9 27.3 30.0 32.7 34.1 35.4 36.1 36.3 36.5 37.2 Congo, Dem. 4.0 4.0 4.0 4.0 4.0 3.9 3.9 3.9 3.9 5.5 Congo, Rep. of Rep. of 21.1 17.7 16.0 14.3 13.5 12.6 12.2 12.1 12.0 11.5 Côte d’Ivoire 9.6 11.1 11.9 12.6 13.0 13.4 13.6 13.6 13.7 13.9 Ethiopia 3.9 4.1 4.1 5.3 5.6 6.1 6.2 6.4 6.6 7.2 Ghana 8.7 8.4 8.2 8.0 7.9 7.8 9.1 9.6 9.9 11.8 Kenya 8.4 12.7 14.8 16.9 18.0 19.1 19.9 20.1 20.4 21.2 Lesotho 7.2 7.2 7.2 7.2 7.2 7.2 7.2 7.2 7.2 7.2 Madagascar 6.0 4.0 3.0 2.0 1.5 1.0 0.7 0.7 0.6 0.3 Malawi 4.8 4.3 4.0 3.8 3.7 3.6 3.5 3.5 3.5 3.4 Mali 26.6 26.6 26.6 26.6 26.6 28.1 28.8 29.1 29.3 30.0 Mozambique 9.6 7.7 6.8 7.4 7.7 9.0 9.6 9.8 10.0 10.9 Namibia 11.7 11.7 11.7 11.7 11.7 11.7 11.7 11.7 11.7 11.7 Niger 14.5 14.2 14.1 13.9 13.9 13.8 13.7 13.7 13.7 13.7 Nigeria 2.5 3.8 3.4 3.8 4.2 4.9 5.3 5.4 5.6 6.0 Rwanda 14.6 14.6 14.6 14.6 14.6 14.6 14.6 14.6 14.6 14.6 Senegal 23.8 23.8 23.8 23.8 24.2 24.8 25.1 25.2 25.3 25.7 South Africa — 2.4 3.6 4.8 5.4 6.0 6.3 6.4 6.5 6.8 Sudan — — — — — — — — — — Tanzania 3.2 5.5 6.7 7.9 8.5 9.0 8.8 8.8 8.8 8.6 Uganda 19.5 20.7 21.4 22.0 22.3 22.6 22.8 22.8 22.9 23.1 Zambia 4.2 3.3 2.9 2.4 2.2 2.0 2.1 2.1 2.1 2.5 Source: Africa Infrastructure Country Diagnostic Power Tariff Database Notes: Average residential consumption level is highlighted in bold — Not available. 15 Figure 6. Effective residential tariff for 100 kWh (cents/kWh) 35.0 30.0 25.0 20.0 15.0 10.0 5.0 - Tanzania Ethiopia Chad Uganda Kenya Côte d'Ivoire Namibia Niger Benin Ghana Lesotho Congo, Rep. Mozambique South Africa Mali Cape Verde Zimbabwe Zambia Senegal Rwanda Cameroon Botswana Nigeria Congo Madagascar Burkina Faso Malawi Source: Africa Infrastructure Country Diagnostic Power Tariff Database On average, residential electricity in Africa is among the most expensive in the world. For the average consumer, residential tariffs are over 12 cents/kWh in about 60 percent of the sample countries and over 20 cents/kWh in close to 25 percent of the sample (Burkina Faso, Cape Verde, Uganda, Madagascar, Mali, and Chad) (figure 6). To compare it against other world regions, Africa’s residential electricity price averages between 50 and 150 percent higher than the 8 cents/kWh in Latin-America, Eastern Europe, and East Asia, and up to 400 percent higher than average residential tariffs in South Asia. What is the pattern of average effective residential tariffs across levels of consumption? Tariffs that increase with consumption effectively impose a penalty on higher consumption and vice versa. About half of the sample countries have increasing average effective residential tariffs (for example, Cape Verde, Chad, Côte d’Ivoire, Ethiopia, Ghana, Kenya, Madagascar, Mozambique, Nigeria, Tanzania, and Uganda [figure 7a and b]). One-third of countries have decreasing average effective residential tariffs, not only because they have explicitly adopted DBT (for example, Senegal) but also because the size of their first blocks is large meaning that the weight of the fixed charge is spread across a larger tranche of consumption. Countries with decreasing residential tariffs include Cameroon, Malawi, Zambia, Senegal, Burkina Faso, Cameroon, the Democratic Republic of Congo, Ghana, and Niger (figure 7c and d). 16 Figure 7. Effective tariffs across various consumption levels a. Highly increasing b. Moderately increasing 40 40 US cents per kWh US cents per kWh 30 30 20 20 10 10 0 0 50 75 100 150 200 300 400 450 500 900 50 75 100 150 200 300 400 450 500 900 monthly consumption, kWh monthly consumption, kWh Cape Verde Côte d'Ivoire Madagascar Chad Kenya Mali Tanzania Uganda Zimbabwe c. Neutral d. Decreasing 40 40 US cents per kWh US cents per kWh 30 30 20 20 10 10 0 0 50 75 100 150 200 300 400 450 500 900 50 75 100 150 200 300 400 450 500 900 monthly consumption, kWh monthly consumption, kWh Cameroon Malawi Senegal Burkina Faso Congo, Dem. Rep. Niger Zambia Congo, Rep. Botswana Source: Africa Infrastructure Country Diagnostic Power Tariff Database b. What is the level of average effective nonresidential tariffs? As for residential tariffs, one commercial and one industrial schedule per country were selected for our analysis of nonresidential tariffs. Representative or typical commercial customers are defined as small to medium business users with an average consumption level of 900 kWh/month. Representative or typical industrial customers are medium to large business users usually associated with high-voltage, high- usage tariffs. But in order to exclude very large industries with preferential tariffs from our analysis, we did not use the highest voltage bracket included in tariff schedules (box 1). In cases where tariff schedules for commercial and industrial users are not differentiated or where several tariffs apply to commercial users, tariff schedules were selected based on both voltage level and load. For example, commercial customers were associated with a tariff for medium-voltage and medium-consumption users, and industrial customers were associated with a tariff for medium- to high-end users. 17 Box 1. Special tariffs for large electricity users: The case of Zambia The average effective power tariff in Zambia, at 3 cents/kWh, is one of the lowest in Africa. This current level does not even allow for the recovery of operating costs, yet alone total costs—even though Zambia has one of the lowest average costs in the region (due to a combination of hydropower technologies and excess generation capacity). Such inefficient pricing policies are compounded by the exceptionally favorable prices that the power utility ZESCO gives to mining companies, in particular the Copperbelt Energy Corporation (CEC). A long-term agreement set mining tariffs at 2 cents/kWh, not only below cost recovery but also one-third lower than the effective tariff for an average residential customer (100 kWh/month). As the mining sector is the recipient of 50 percent of total ZESCO sales, this translates into a conservative estimate of $30 million in annual subsidies with a projected cumulative deficit of $926 million over the next 10 years. Zambia’s case is not unique in the region. Until 2003 Ghana’s power distribution company, VRA, was engaged in a long-term agreement with Volta Aluminum Company, which was VRA’s most important customer, consuming one- third of its power generation and benefiting from a preferential electricity price estimated to be half of the cost- recovery level. Source: Zambia Electricity Regulator Board 2008; World Bank 2008; Chivakul and York 2006 Table 7 lists the effective nonresidential tariffs for select African countries. Commercial effective tariffs are higher overall than industrial ones at similar levels of consumption. Two-thirds of the sample countries have commercial tariffs that are, on average, 20–30 percent higher than the industrial ones (table 7). These countries include Burkina Faso, Cape Verde, Chad, Madagascar, and Ethiopia. Another handful of countries have even higher price differentials—Benin, Côte d’Ivoire, and South Africa have commercial effective tariffs between 40 and 77 percent higher than industrial tariffs. This pattern is not unusual, as the genuine production costs of the high-voltage power consumed by industrial users are lower and exposed to fewer transmission and distribution losses (figure 8). The Democratic Republic of Congo and Mozambique are notable exceptions to this pattern: commercial customers pay on average 30–40 percent more than industrial customers. 18 Table 7. Effective nonresidential tariffs by level of consumption (cents) Commercial Industrial Level of consumption (kWh)/month Level of demand 100 450 900 2,500 5,000 10 kVA kVA Benin 15.1 15.1 15.1 15.1 10.7 10.7 Burkina Faso 27.0 26.7 26.5 26.5 14.1 15.0 Cameroon 11.7 11.4 11.3 11.2 8.8 9.2 Cape Verde 21.8 21.8 21.8 21.8 17.7 17.7 Chad 43.7 44.7 45.3 45.5 45.6 38.8 Congo, Dem. 11.0 11.0 10.8 10.8 14.6 14.6 Rep. of Côte d’Ivoire 17.8 16.9 16.3 16.1 10.7 10.7 Ethiopia 9.8 8.3 7.3 7.0 4.8 4.7 Ghana 12.6 13.9 15.2 15.6 5.7 6.4 Kenya 22.1 21.7 21.5 21.4 16.6 15.1 Lesotho 9.3 9.3 9.3 9.3 1.5 3.3 Madagascar 30.9 25.3 21.7 20.7 11.1 10.5 Malawi 8.1 6.9 6.1 5.8 2.6 3.1 Mozambique 9.0 8.0 7.5 7.3 4.7 5.1 Namibia 15.2 14.0 13.2 13.0 12.7 13.6 Niger 13.4 13.2 13.0 12.9 9.0 9.3 Nigeria 5.1 5.0 5.0 5.5 5.0 5.1 Rwanda 17.2 17.2 17.2 17.2 17.2 17.2 Senegal 23.8 22.8 26.2 26.0 15.8 15.8 South Africa 11.4 7.7 5.3 4.7 2.7 2.7 Tanzania 8.6 8.0 7.6 7.5 5.0 5.4 Uganda 22.0 21.9 21.8 21.8 16.8 17.0 Zambia 5.1 4.4 3.9 3.8 2.3 2.5 Congo, Rep. 11.7 10.7 10.1 9.9 11.2 11.2 Mali of 23.2 23.2 23.2 23.2 Botswana 7.7 7.2 6.9 6.8 3.3 4.0 Source: Africa Infrastructure Country Diagnostic Power Tariff Database Notes: Average commercial consumption level is highlighted in bold — Not available. Figure 8. Increasing, decreasing, and neutral tariff structures 20 number of countries 15 10 5 0 domestic commercial industrial type of effective tariff structure increasing neutral decreasing Source: Africa Infrastructure Country Diagnostic Power Tariff Database 19 c. What have been historic costs of power production? As we have discussed, Africa faces effective tariffs that are up to twice as high as in other developing regions. This reflects the use of costly technologies as well as the small scale of most African power generation systems. But what have been the historic costs across the various power systems? The first step toward answering this question is to attribute to each kilowatt-hour produced both the capital and operating (O&M) costs of the three power production segments: generation, transmission, and distribution. O&M unit costs are calculated by prorating total operational costs (as reported in utilities’ income statements) by the total generated electricity over a given year. Data are derived from the AICD’s fiscal spending database,5 and individual results have been verified by country power experts. Operational costs include salaries associated with system operations, fuel charges, the cost of parts needed for daily operations, and so on. Unit capital costs are calculated using the levelized power methodology (commonly used by the International Energy Agency). This requires allocating the value of an asset over its lifetime capacity. In essence the unit capital cost is a ratio of the net present value of total lifetime investment to the total electricity produced.6 For this purpose an annualization factor is applied to the value of generation, transmission, and distribution assets. The overnight investment or capital needed to replace existing assets is the proxy for asset value, to allow for cross-country comparison. The annualization factor is a reversed version of a standard formula for the net present value of a periodic investment of equal amount. It takes into account both depreciation and interest rates: r A C O  C I 1 (1 r)T ; where ACC is annualized capital cost, OI is overnight investment, r is annual discount rate, and T is plant life (in years). In most African countries, the cost of each kilowatt-hour is 10–20 cents (figure 9). In a handful of countries—such as South Africa, Zambia, the Democratic Republic of Congo, Ethiopia, Mozambique, and Malawi (all of them significant hydro-power producers, with the exception of South Africa)—the power cost is below 10 cents/kWh. Niger and Mali, at the upper extreme, have unit costs over 30 cents/kWh. The balance between capital and operating costs also varies from country to country, largely determined by generation technology. By way of example, capital costs as a percentage of total costs range from 14 percent in Botswana to 77 percent in Nigeria. Full details of these calculations are provided in Annexes 6-8. 5 http://www.infrastructureafrica.org. 6 Lifetime capital costs are estimated using: (i) the discounted net present value of lifelong capital costs (for example, the sum of the annual investment expenditure throughout the life of a plant), (ii) the discounted historical cost of existing assets, and (iii) the relative value of a similar asset (replacement cost). We are using method (i), with the overnight construction costs as a proxy for the net present value of the lifelong investment expenses. 20 Figure 9. Historic power costs in select African countries (cents/kWh) 40 35 30 25 20 15 10 5 0 Tanzania Ethiopia Mozambique* Zambia Malawi Uganda Namibia Kenya Rwanda Cape Verde * Benin * Madagascar Lesotho South Africa Botswana Congo, Dem. Rep. of Mali Congo Nigeria Ghana Chad Niger Senegal Cameroon Cote d'Ivoire Burkina Faso Operating Costs Capital Costs Source: Africa Infrastructure Country Diagnostic Power Tariff Database Note: For details see annex 8. But what drives costs? In general, hydro-based systems tend to be more skewed towards capital cost and thermal-based systems more skewed toward operating cost In terms of geography. Africa’s landlocked and island nations seem to bear a significant power cost disadvantage vis-à-vis the coastal countries. High power costs are also driven by the size of markets and their associated scales of production (figure 10). Figure 10. Differences in average costs (US cents per kWh) a. According to type of power system b. According to geography c. According to system size 20 20 20 15 15 10 10 10 5 5 0 0 0 Total Capital Operational Total Capital Operational Total Capital Operational costs costs costs costs costs costs costs costs costs Hydro Thermal Coastal Landlocked Island < 700 MW 700-1,500 MW > 1,500 MW Source: Africa Infrastructure Country Diagnostic Power Tariff Database Note: For details see annex 7. d. Do average effective tariffs cover historic costs? Existing effective tariffs allow for the recovery of operational costs for close to 80 percent of African countries (figure 11b). But when capital investments are considered, the picture looks less rosy. On average, sample countries recover only two-thirds of their capital costs. Indeed, barely one-third of the 21 sample countries are practicing full capital cost recovery (figure 11a); Among the countries that should be able to cover total costs under existing tariff regimes (contingent on their ability to collect bills) are Burkina Faso, Côte d’Ivoire, Kenya, Namibia, and Senegal. Figure 11. Cost-recovery capabilities of effective tariff regimes a. Weighted effective tariffs and total costs b. Weighted effective tariffs and OPEX 30 30 Weighted Effective Tariffs Weighted Effective Tariffs 25 25 (cents/kWh) (cents/kWh) 20 20 15 15 10 10 5 5 0 0 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Historical costs, US cents/kWh Historical opex costs, US cents/kWh Source: Africa Infrastructure Country Diagnostic Power Tariff Database Note: OPEX = operational expenditure. Lack of full cost recovery can be partly explained by the large weight of residential customers in utilities’ revenue structure, whose charges – for political and social reasons – often fail to reflect costs. Excluding South Africa, roughly 50 percent of the total power billing in Africa is associated with households, for just over 40 percent of the total power supplied (figure 13). This is a very high share by global standards. In South Africa, by contrast, the total share of residential billing stands much lower at 17 percent versus only 8 percent of total power supplied. Only in 30 percent of the sample countries do residential tariffs allow for 100 percent cost recovery (figure 12a). While residential tariffs do better at covering operating costs (figure 12b), this is not enough to guarantee financial sustainability in the medium to long term. It is important to underscore that setting tariffs at cost-recovery levels is one thing and actually recovering costs is another. African utilities are characterized by bill-collection rates of well below 100 percent. This translates to financial losses that can amount to more than the entire turnover of a utility (see annex 14 for details). Have African utilities been improving their cost-recovery rates over time? Power tariffs increased substantially over the period 2001 to 2005, but not fast enough to keep pace with rising costs (figure 14). Indeed, by 2005, average revenues had only just caught-up with the average costs in 2001. 22 Figure 12. Cost-recovery capabilities of residential effective tariffs at 100 kWh/month consumption a. Total cost recovery b. Operational cost recovery 80% 80% 70% 70% percentage of countries percentage of countries 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% 30%-50% 50%-80% 80%-99% recovery 30%-50% 50%-80% 80%-99% recovery 0%- 30% 0%- 30% cost cost level of total cost recovery level of operational cost recovery Source: Africa Infrastructure Country Diagnostic Power Tariff Database Figure 13. Household billing as a share of total power billing a. Composition of annual billing per type of customer b. Residential share of billing and consumption 100% 80 Share of residential supply 90% % of amount billed 80% 70% 60% 60 50% 40% 30% 20% 40 10% 0% Cape Verde Senegal Zimbabwe Mali Burkina Faso Cameroon Mozambique Kenya Niger Chad Ghana Nigeria Tanzania Rwanda Cote d'Ivoire Ethiopia South Africa Congo DRC Congo, Rep. 20 0 High voltage industrial 0 20 40 60 80 Medium voltage commercial Share of residential bill Low voltage residential Source: Africa Infrastructure Country Diagnostic Power Tariff Database Note: See annex 9 for details. 23 Figure 14. Costs and revenues over time ($/kWh) a. Unit costs b. Average revenues 0.30 0.30 Predominantly diesel 0.25 0.25 Predominantly hydro Overall 0.20 Predominantly diesel 0.20 Predominantly hydro 0.15 Overall 0.15 0.10 0.10 0.05 0.05 0.00 0.00 2001 2002 2003 2004 2005 2001 2002 2003 2004 2005 Source: Africa Infrastructure Country Diagnostic Power Tariff Database e. Do IBT structures allow for costs to be recovered? As noted above, IBTs are premised on the notion that surcharges on higher volumes of consumption will compensate for discounts on lower volumes of consumption, so that the utility breaks even overall. However, this outcome is contingent on the block sizes and associated price levels being correctly calibrated. It is often the case that the surcharges apply (if at all) to very high levels of consumption that are rarely reached in practice. In the case of Africa, the answer is more promising than might be anticipated. A minority of countries have residential tariffs that recover costs independent of consumption levels, meaning that even the lowest priced block is priced high enough to recover costs (category I in table 8). (This is ideal from a cost recovery perspective, but raises questions as to whether these IBTs are also performing the intended social function of providing subsidized power to small consumers.) A second batch of countries has IBTs that succeed in recovering operational costs (category II), but will not—or are not likely to—attain total cost recovery based on their historical average household consumption. The third group of countries (category III in table 8) is recovering costs within a consumption range that is close to the historic average. Finally, a fourth group of countries (category IV)—given current tariffs and the historical consumption patterns—will never attain operational (let alone total) cost recovery. 24 Table 8. Break-even consumption levels ID Countries Operational cost Total cost I Burkina Faso Cape Verde Chad At any level At any level Namibia Senegal Uganda II Côte d’Ivoire 90 Mozambique At any level 300 Ghana 1,070 Ethiopia Lesotho At any level Never Rwanda Mali Nigeria 24 Never Tanzania 155 III Botswana 27 20 Zambia 65 27 Congo, Rep. of 20 29 Kenya 50 91 Benin 12 110 South Africa 94 290 IV Congo, Dem. Rep. of 600 600 Cameroon Never Never Madagascar Malawi Niger Source: Africa Infrastructure Country Diagnostic Power Tariff Database Are power tariffs efficient from an economic standpoint? We have shown that average effective tariffs are not all that successful at recovering historic costs, and this is important from a financial perspective. However, historic costs are not necessarily a good guide to future power development costs. What is important from an economic perspective is whether average effective tariffs cover long-run marginal costs (LRMC) of system development. In this section we assess whether power tariffs provide this correct economic signal, and thus do not lead to the over- or under- consumption of power from an economic standpoint. a. What do LRMCs look like in Africa? So far, we have seen that tariffs in Africa are high compared with those in other developing regions, but not high enough to allow for historic cost recovery. We have also seen that high tariff levels are a direct consequence of high costs, which are driven by the use of sub-optimal primary energy sources (small scale diesel versus large hydro), difficult geography (with higher costs faced by landlocked countries and 25 islands), and diseconomies of scale (due to the prevalence of small national systems). However, in principle, these costs could come down in the future as countries harness more cost-effective sources of energy and exploit regional power trade to expand the scale of production. LRMCs are calculated using a dynamic model that estimates the needs of African power systems based on economic growth projections and electrification targets. The model simulates optimal (least-cost) strategies for generating, transmitting, and distributing electricity in response to demand increases. It also estimates the cost of meeting power demand under a range of alternative scenarios, including cross-border trade (Vennemo and Rosnes 2008).7 With few exceptions, a more efficient selection of technologies (with and without greater regional trade) would render LRMCs below 10 cents/kWh (figure 15). Only four countries—Burkina Faso, Mali, Niger, and Senegal—would face LRMCs over 20 cents/kWh. Developing African power systems with a view toward expanding regional trade clearly reduces LRMCs across most countries. In fact, if regional trade were fully pursued countries would see their LRMCs reduced by about 10 percent on average, and as high as 40 percent in some cases (figure 16a). These would translate into LRMC reductions of as high as 7 cents/kWh, but more typically in the 1–5 cents range (figure 16b). Figure 15. Long-run marginal costs (US cents per kWh) 50 45 40 35 30 25 20 15 10 5 0 Chad Gambia Ghana Sudan Angola Niger Zimbabwe Tanzania Namibia Rwanda Togo Senegal Congo Burundi Uganda Malawi Lesotho Guinea Botswana Cameroon Zambia Sierra Leone Mozambique Benin Ethiopia Mali Kenya Gabon Guinea-Bissau Nigeria Burkina Faso Equatorial Guinea Liberia Congo (DRC) Central African Rep. Cote d'Ivoire South Africa Current patterns of trade Trade expansion Source: Adapted from Vennemo and Rosnes 2008 7 For a brief description of the model, see annex 10. 26 Figure 16. How trade expansion would effect long-run marginal costs a. Percent reduction on LRMC b. Monetary reduction on LRMC 14 14 12 12 number of countries percentage of countries 10 10 8 8 6 6 4 4 2 2 - - Increase Change 0-20% 20-40% 40% < Increase 5 cents < Change 0-3 cents 3-5 cents No No Source: Adapted from Vennemo and Rosnes 2008 Note: See details in annex 11. b. Do existing average effective tariffs cover LRMC? It is relevant to ask whether existing average effective tariffs are high enough to cover LRMC even if they may not be high enough to cover average historic costs as was demonstrated above. The analysis shows that 38 percent of the sample countries have already achieved average effective tariffs that are high enough for full capital cost recovery (figure 17a). Compare this with only 30 percent of countries that can fully recover historic costs. Figure 17. Cost-recovery capabilities of residential effective tariffs vis-à-vis LRMCs a. LRMC recovery, trade stagnation b. LRMC recovery, trade expansion 80% 30 Effective Residential Tariffs US 70% 25 percentage of countries 60% 20 cents/kWh 50% 40% 15 30% 20% 10 10% 5 0% recovery 30%-50% 50%-80% 80%-99% 0%- 30% 0 cost 0 5 10 15 20 25 30 LRMC, trade expansion, US cents/kWh level of LRMC recovery Source: Adapted from Vennemo and Rosnes 2008 and Africa Infrastructure Country Diagnostic Power Tariff Database. 27 Are power tariffs equitable and affordable? c. Can African households afford power services? Based on information reported in household surveys, on average power bills absorb almost 6 percent of total household budgets. For most countries the share falls below 3 percent; however in a few cases (such as Malawi and Mozambique) it can be as high as 10 or even 20 percent (figure 18). This share is relatively stable across quintiles. Figure 18. Existing household spending on electricity 70% 60% 50% 40% 30% 20% 10% 0% National Poorest Q2 Q3 Q4 Richest quintile Quintile Below 3% 3-5% 5-10% Over 10% Source: Adapted from Banerjee and others (2008). In order to gauge whether power is affordable, two types of evidence can be considered. One possible measure of affordability is non-payment of services. Based on household surveys, we can compare across quintiles the percentage of households that report paying for power against the percentage of households that report using service. Those using without paying include both clandestine users who steal power from the network and formal customers who fail to pay their bills. Overall, about 40 percent of people connected to electricity do not pay for it (figure 19). Nonpayment rates range from about 20 percent in the richest quintile to about 60 percent in the poorest quintile. A significant nonpayment rate, even among the richest quintiles, suggests that a culture of nonpayment exists in addition to affordability issues. 28 Figure 19. Percentage of the population with service connections who do not pay for service 60% 50% 40% 30% 20% 10% 0% Poorest quintile 2nd quintile 3rd quintile 4th quintile quintile Source: Banerjee and others 2008. Another possible measure of affordability is whether the full economic cost of a subsistence level of consumption falls above a normative affordability threshold. Economic cost is defined as the tariff that would fully cover both operating and capital costs and is country specific; reaching an average level of US$0.18/kW for SSA as a whole. Subsistence consumption is defined as 50 kWh per month, which is enough to cover very minimal usage for lighting (roughly one light bulb for two hours per day). The affordability threshold is typically defined as spending on subsistence power needs of between 3 and 5 percent of the total household budget. These values are normative, and are informed by power spending patterns by low income households that have been observed across a wide range of household surveys (recall figure 18 above). By looking at the distribution of household budgets, one can calculate the percentage of households for which subsistence consumption priced at full economic cost would absorb more than 5 percent of their budgets and thus prove unaffordable. For example, looking across the distribution of household budgets for all of SSA, monthly bills of $2 would be affordable for almost the entire population, whereas monthly bills of $10 would only remain affordable for the entire population of middle-income African countries. Based on existing average effective tariffs, the bill for subsistence consumption levels looks very affordable for those that are already connected to the grid (Figure 20a). With a 3 percent affordability threshold, the subsistence consumption of 50kWh/month priced at the current average effective tariff is affordable in 60 percent of the sample countries. If the affordability threshold is further raised to 5 percent, the subsistence consumption is affordable in over 90 percent of the sample countries. However, the picture looks very different for those that are not currently connected to the grid. In these cases, the subsistence consumption level priced at the current average effective tariff would only be affordable in about 25 percent of the countries in the sample (Figure 20b). In conclusion, a significant 29 majority of those connected can afford power at existing prices, while a significant majority of those unconnected cannot do so. This raises questions of circularity: either existing tariffs determine who is connected or tariffs are designed to be affordable to those who are connected. Figure 20. Monthly electricity expenditure as a percentage of total household budget a. Connected households b. Unconnected households 70% 60% Percentage of countries (%) Percentage of countries (%) 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% Below 3% 3-5% 5-10% Over 10% Below 3% 3-5% 5-10% Over 10% Percentage of Household Budget Percentage of Household Budget 50 kWh/mo 75 kWh/mo 100 kWh/mo 50 kWh/mo 75 kWh/mo 100 kWh/mo Source: Adapted from Banerjee and others, 2008 An equally important question is whether tariffs would remain affordable if today’s tariffs were adjusted to allow for the recovery of full economic costs. For this purpose, we use two cost recovery benchmarks: the average historic cost and the Long Run Marginal Cost (table 9). Under historic cost recovery, a subsistence level of consumption of 50 kWh per month would range in cost from $3 to $16 a month. These monthly bills would on average be affordable for 72 percent of households across the sample. There are only a handful of countries where less than half of the population could afford these bills, notably: Niger (7 percent), Ethiopia (12 percent), Malawi (43 percent). If instead, a forward-looking Long Run Marginal Cost is used for cost recovery purposes, the results are very similar on average. Monthly bills would on average be affordable for 73 percent of households across the sample. However, the position of individual countries looks quite different. For one group of countries (DRC, Malawi, Tanzania and to a lesser extent Benin and Kenya), LRMC based tariffs are significantly more affordable than historic cost recovery tariffs. For a second group of countries (Ethiopia, Uganda), LRMC based tariffs are significantly less affordable than historic cost recovery tariffs. 30 Table 9. Monthly power bill for subsistence consumption (50 kWh) Monthly bill Share of households that ($) can afford the monthly bill (%) (*) Historic Effective tariff Historic cost LRMC Effective tariff cost LRMC Benin 6.31 9.92 9.50 95 68 72 Burkina Faso 10.29 7.53 12.50 51 69 36 Cameroon 4.30 8.56 3.50 100 100 100 Cape Verde 11.81 8.95 100 100 — Congo, Dem. Rep. of 1.99 3.38 2.00 100 63 91 Côte d’Ivoire 4.81 5.47 7.50 100 99 98 Ethiopia 1.97 4.23 9.50 60 12 1 Ghana 4.36 6.18 5.00 97 93 96 Kenya 4.21 7.10 6.00 99 87 95 Madagascar 2.98 7.49 92 55 — Malawi 2.39 4.54 2.50 92 43 91 Niger 7.25 16.07 12.50 55 7 19 Nigeria 1.25 4.84 6.50 97 84 74 Senegal 9.31 5.77 21.50 100 100 — South Africa — 2.98 3.00 100 100 100 Tanzania 1.60 7.04 5.00 99 59 84 Uganda 9.74 5.19 6.00 20 66 55 Zambia 2.09 3.26 4.00 100 97 96 Sub-Saharan African average 4.81 6.58 7.28 86.51 72.39 73.73 Source: Adapted from Vennemo and Rosnes 2008 and Africa Infrastructure Country Diagnostic Power Tariff Database. Notes: See Annex 12 for further details (*) it is assumed that a bill is affordable if it is below 5 percent of the household budget. — Not available. A frequent argument used for not raising tariffs to full cost recovery levels is the potential impact on poverty. However, empirical evidence suggests that the immediate poverty-related effect of raising tariffs to cost-recovery levels is generally quite small, although it may have second-order effects. Detailed analysis of the effect of significant tariff increases of the order of 40 percent for power and water services in Senegal and power services in Mali confirms that the immediate poverty-related effect on consumers is small, essentially because very few poor consumers are connected to the service (Boccanfuso, Estache, and Savard 2008a; 2008b; 2008c). As the consequences of higher power or water prices work their way through the economy, however, broader second-order effects on wages and prices of goods in the economy as a whole can have a more substantial impact on poverty (Boccanfuso, Estache, and Savard 2008a; 2008b; 2008c). d. Are power tariffs equitable? Notwithstanding these findings, most African countries subsidize tariffs for power. On average, power tariffs recover only 80 percent of costs. The resulting implicit subsidies amount to as much as $2.3 billion a year on aggregate (or 0.4 percent of Africa’s GDP) (Foster and Briceño-Garmendia 2009). The aggregate burden of underpricing can be as much as 1–1.5 percent of a country’s GDP (as in Cameroon, the Democratic Republic of Congo, Ghana, Mali, Malawi, Nigeria, South Africa, Tanzania, Zambia, 31 Uganda) and even higher (Botswana and Niger). From the utility’s perspective, underpricing can amount to losses valued as much (and even more than) 100 percent of the utility’s turnover (see annex 14 for details). Because electricity subsidies are typically justified by the need to make services affordable to low- income households, a key question is whether subsidies reach such households. Results across a wide range of African countries show that the share of subsidies going to the poor is less than half their share of the population, indicating a very pro-rich distribution (figure 21). This result simply reflects the fact that connections to power are already highly skewed toward more affluent households. In SSA as a whole, access to power among the bottom three quintiles of the budget distribution is no more than 12 percent on average compared to 72 percent in the top budget quintile. To put these results in perspective, one must compare them with the aims achieved by other forms of social policy. Estimates for Cameroon, Gabon, and Guinea indicate that expenditures on primary education and basic health care reach the poor better than do power and water subsidies (Wodon 2007) Figure 21. Extent to which electricity reaches the poor Nigeria Gabon Congo Côte d'ivoire Cape Verde Togo Sao Tome Senegal Cameroon Mozambique Ghana CAR Guinea Burundi Burkina Chad Malawi Uganda Rwanda 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Measure of distributional incidence Sources: Banerjee and others 2008; Wodon 2007. Note: A measure of distributional incidence captures the share of subsidies received by the poor, divided by the proportion of the population in poverty. A value greater than 1 implies that the subsidy distribution is progressive (pro-poor), because the share of benefits allocated to the poor is larger than their share of the total population. A value less than 1 implies that the distribution is regressive (pro-rich). A key message is that power subsidies will always be highly regressive as long as access is highly regressive. The distributional score presented above (figure 21) can be decomposed into access and subsidy design factors (figure 22). The access factor is related to the availability of electricity in the area where the household lives and to the household’s choice to connect to the network if service is available. The subsidy design factor relates to who is targeted to receive the subsidies, rates of subsidization, and consumption levels. As for the overall distributional score, values higher (lower) than one for access and subsidy design factors are indications that those factors are progressive (regressive). 32 Figure 22. Access factors and subsidy design factors affecting targeting performance 1.60 Congo 1.40 1.20 Tariff Structure and Subsidy Design Gabon 1.00 Togo Nigeria Guinea Mozambique Cape Verde d'ivoire Côte 0.80 Ghana CAR Cameroon Sao Tome Burkina Senegal 0.60 Chad Rwanda Burundi 0.40 Uganda 0.20 0.00 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 Access Factors Source: Banerjee and others 2008. Note: Access factors capture the rates of connection among the poor to the network divided by the rates of connection to the population as a whole. Subsidy design factors are the ratio of the average benefit from the subsidy among all poor households connected to the network divided by the average benefit among all households connected to the network. A value greater than 1 implies that the factor distribution is progressive (pro-poor), because the share of benefits allocated to the poor is larger than their share in the total population. A value less than 1 implies that the distribution is regressive (pro-rich). In general the findings are that the subsidy design factor exceeds the access factor. As was to be expected, given the pattern of connections to power, the access factor is always less than one meaning highly regressive. The tariff factor, on the other hand, ranges from 0.4 (indicating a highly regressive tariff structure) and 1.6 (indicating a moderately progressive tariff structure). The most progressive tariff structures are found in DRC, Gabon and Togo. However, for the bulk of countries analyzed are marginally below unity, suggesting that the tariff structures are at best distributionally neutral. However, this is not much of an achievement given that the intention behind the predominantly IBT tariff structures is to favor the poor. This finding is explained by the fact that the traditional IBTs that prevail in Africa tend to be poorly targeted because tariff structures subsidize consumption in the first blocks even for customers whose aggregate consumption is high. On top of that, the consumption threshold for the lower blocks tends to be too high to single out the poor, the price difference between blocks is not very large, and fixed charges are too high.8 8 Also discussed in Wodon (2007) and Banerjee and others (2008). 33 How much are poor customers penalized by fixed charges and block structures? As noted above, over 50 percent of the sample countries have incorporated fixed charges in their tariff schemes. Fixed charges in African countries range from US$0.43 cents to US$5.00 per month (figure 23a). Furthermore, these charges constitute a large portion of the aggregate monthly bill particularly at subsistence levels of consumption (figure 23b). At 50 kWh/month—subsistence consumption—the fixed charge represents more than 40 percent of the monthly bill in over 30 percent of the countries. At higher levels of consumption—75kWh/month and 100 kWh/month—the weight of the fixed charge is less prominent. This indicates the disproportionate weight of the fixed charge in the bills of households consuming at the subsistence level. Figure 23. Fixed portion of residential tariffs a. Residential fixed charge per month ($) b. Fixed charge as a share of monthly bill (%) Niger Nigeria 70% Ghana 60% Percentage of countries Côte d'Ivoire Ethiopia 50% Malawi Uganda 40% Burkina Faso 30% Zambia Tanzania 20% Botswana Kenya 10% Mozambique 0% Madagascar Cameroon 50 kWh/mo 75 kWh/mo 100 kWh/mo Congo, Rep. 1%-20% 21%-40% >40% - 1 2 3 4 5 Source: Africa Infrastructure Country Diagnostic Power Tariff Database. e. Are there any other kinds of hidden subsidies? Besides tariffs, there are also other less explicit mechanisms by which policy makers subsidize consumption. For instance, when policy makers overlook, tolerate, or even promote certain operational inefficiencies, they are in practice transferring resources from one sector of the economy to another, from the producer to the consumer, from future taxpayers to current customers, and so on. Tolerance of nonpayment is an implicit tax on utilities (and/or a transfer to consumer). Tolerance of pilferage—one of the main causes of transmission and distribution losses—is an implicit subsidy to customers and an implicit burden on future taxpayers. Acceptance and promotion of over-employment represents an untargeted transfer of resources from the utility to the society. These inefficiencies can be empirically quantified and prove to be substantial relative to GDP (Briceño-Garmendia, Foster, and Smits 2008). For most countries, over-manning as well as collection inefficiencies amount to less than 0.2 percent of GDP, whereas transmission and distribution losses tend to be much larger in value amounting to around 0.2-0.6 percent of GDP (figure 24a). These operational inefficiencies also look very large in 34 comparison with utility revenues; amounting to between 20 and 60 percent of utility revenues in most cases, and exceeding 100 percent of utility revenues in a few cases (figure 24b). In countries such as the Democratic Republic of Congo, the Republic of Congo, Nigeria, Namibia, and Cameroon—to name only a few—transmission and distribution losses (technical and nontechnical) are the lead cause of hidden losses (figure 24b). In Côte d’Ivoire, Niger, Ghana, Uganda, and Botswana—to cite some examples—under-collection of bills is the main offender, though it is observed that unpaid bills are from the government or other public enterprises. Finally, over-manning seems to be an issue for countries such as Cape Verde and Chad. Figure 24. Monetary costs of overlooked inefficiencies a. % of GDP b. % of revenue 90% Congo DR Cote d'Ivoire Percentage of Countries (%) 80% Nigeria 70% Congo, Rep. Niger 60% Ghana 50% Uganda Mali 40% Malawi 30% Botswana Cape Verde 20% Namibia 2006 Cameroon 10% Tanzania 0% Mozambique Chad Ethiopia Benin Burkina Faso Lesotho Rwanda Senegal Kenya Transmission and distribution Losses Zambia Madagascar South Africa Collection Inefficiencies 0 20 40 60 80 100 Over-manning Transmission and Distribution Losses Collection Inefficiencies Over-manning Source: Briceño-Garmendia, Foster, and Smits 2008 In fact it is not infrequent that the financial burden of the myriad operational inefficiencies is higher than the cost of subsidizing via under-pricing. In about half of the countries in our sample, the magnitude of these operational inefficiencies is higher than the magnitude of price subsidies (figure 25). Figure 25. Weight of underpricing vis-à-vis operational inefficiencies 35 200 Operational Inefficiencies 180 160 (% Revenues) 140 120 100 80 60 40 20 0 0 50 100 150 200 Price subsidy (% Revenues) Source: Africa Infrastructure Country Diagnostic Power Tariff Database. Overall, how would we rate Sub-Saharan African power tariffs? A scorecard combining four of the key goals in the design of power tariffs—cost recovery, efficiency, equity and affordability—illustrates the challenges of simultaneously achieving these sometimes conflicting objectives. For each of these objectives a quantitative indicator is used based on the foregoing analysis. Cost recovery is measured as the ratio of the current average effective tariff to the average historic cost of power production. Efficiency is measured as the ratio of the current average effective tariff to the Long Run Marginal Cost of power production. Affordability is measured as the percentage of households that are able to purchase a subsistence level of consumption of 50 kWh/month at the prevailing average effective tariff without spending more than 5 percent of their household budgets. Equity is measured as the share of the subsidy captured by households living under the poverty line divided by the percentage of households in the population that live under the poverty line. The analysis shows that the average scores for the sample are 78 percent for cost recovery, 82 percent for efficiency, and 87 percent for affordability. And for equity the average score is 0.29 indicating a highly regressive distributional incidence, relative to a score of 1.00 for a tariff that is neutral in distributional terms (table 10). 36 Table 10: Scorecard for performance of country’s power tariffs against four key policy objectives Objective Cost Recovery Efficiency Affordability Equity Indicator Ratio of Ratio of average Share of population that Percentage of subsidy average effective effective tariff to can afford subsistence captured by poor as a tariff to average LRMC consumption priced at ration of percentage of historic cost average effective tariff poor in the population Benin 0.72 0.75 0.95 Botswana 0.54 1.00 0.06 Burkina Faso 1.00 0.87 0.51 Cameroon 0.63 1.00 1.00 0.36 Cape Verde 1.00 1.00 0.48 Chad 1.00 1.00 0.06 Congo, Dem. Rep. of 0.59 1.00 1.00 0.62 Congo 0.80 1.00 Cote d'Ivoire 1.00 0.91 1.00 0.51 Ethiopia 0.76 0.40 0.60 Ghana 0.81 1.00 0.97 0.31 Kenya 1.00 1.00 0.99 Lesotho 0.79 1.00 Madagascar 0.93 0.92 Malawi 0.62 1.00 0.92 Mali 0.79 0.95 Mozambique 0.87 1.00 0.31 Namibia 1.00 0.97 Niger 0.44 0.47 0.55 Nigeria 0.44 0.32 0.97 Rwanda 0.88 1.00 0.01 Senegal 1.00 0.35 1.00 South Africa 0.84 0.72 1.00 0.41 Sudan Tanzania 0.52 0.91 0.99 Uganda 1.00 1.00 0.20 0.02 Zambia 0.44 0.36 1.00 Zimbabwe 0.47 Average 0.78 0.82 0.87 0.29 Source: Derived from Africa Infrastructure Country Diagnostic Power Tariff Database. What becomes immediately clear is that some countries rank very high for cost recovery but do very poorly for equity and affordability and vice versa (table 11). Countries such as Chad, Mozambique, Rwanda, and Uganda tend to rank well for cost recovery but poorly for affordability and equity. On the other end, countries such as South Africa, the Democratic Republic of Congo, Tanzania, and Zambia fare relatively well in terms of equity and affordability but have not been able to achieve cost recovery. What is striking is that achieving all four objective simultaneously is almost impossible in the context of the high-cost low-income environment that characterizes much of SSA today. Hence most countries are caught between cost recovery and affordability. 37 Table 11. Overview of scorecard results Target fully achieved Performs above the median Cost Burkina Faso, Cape Verde, Chad, Côte Ghana, Madagascar, Mozambique, Rwanda, recovery d’Ivoire, Kenya, Namibia, Senegal, Uganda South Africa, Efficiency Chad, Kenya, Uganda, Ghana, Mozambique, Namibia, Congo, Dem. Rep. of Rwanda, Botswana, Cameroon, Lesotho, Malawi, Congo, Rep. of Affordability Senegal, Cameroon, Cape Verde, South Africa, Kenya, Tanzania, Zambia, Côte d’Ivoire Congo, Dem. Rep. of Equity None Cameroon, Cape Verde, South Africa, Côte d’Ivoire, Congo, Dem. Rep. of Source: Derived from Africa Infrastructure Country Diagnostic Power Tariff Database. 38 References Banerjee, Sudeshna, Quentin Wodon, Amadou Diallo, Taras Pushak, Hellal Uddin, Clarence Tsimpo, and Vivien Foster. 2008. “Access, Affordability, and Alternatives: Modern Infrastructure Services in Africa.� Background Paper 2, Africa Infrastructure Sector Diagnostic, World Bank, Washington, DC. Boccanfuso, Dorothée, Antonio Estache, and Luc Savard. 2008a. “Electricity Reforms in Mali: A Micro- Macro Analysis of the Effects on Poverty and Distribution.� Working Paper 4, Africa Infrastructure Country Diagnostic, World Bank, Washington, DC. ———. 2008b. “Electricity Reforms in Senegal: A Micro-Macro Analysis of the Effects on Poverty and Distribution.� Working Paper 5, Africa Infrastructure Country Diagnostic, World Bank, Washington, DC. ———. 2008c. “Water Reforms in Senegal: A Micro-Macro Analysis of the Effects on Poverty and Distribution.� Working Paper 16, Africa Infrastructure Country Diagnostic, World Bank, Washington, DC. Borenstein, S. 2008. “Equity Effects of Increasing-Block Electricity Pricing.� Paper CSEMWP180, Centre for the Study of Energy Markets, University of California, Energy Institute. Briceno-Garmendia, C., K. Smits, and V. Foster. 2008. “Fiscal Costs of Infrastructure in Sub-Saharan Africa.� Africa Infrastructure Country Diagnostic, World Bank, Washington, DC. Chivakul, M., and R. M. York. 2006. “Implications of Quasi-Fiscal Activities in Ghana.� IMF Working Paper, International Monetary Fund, Washington, DC. Eberhard, Anton, Vivien Foster, Cecilia Briceño-Garmendia, Fatimata Ouedraogo, Daniel Camos, and Maria Shkaratan. 2008. “Underpowered: The State of the Power Sector in Sub-Saharan Africa.� Background Paper 6, Africa Infrastructure Sector Diagnostic, World Bank. Filipovid, S., and G. Tanid. 2009. “The Policy of Consumer Protection in the Electricity Market.� Economic Annals, Faculty of Economics, University of Belgrade. http://ea.ekof.bg.ac.yu/pdf/178- 179/6.%20Filipovic_Tanic.pdf. Foster, V., and C. Briceño-Garmendia. 2009. “Africa’s Infrastructure: A Time For Transformation,� chapter 1. World Bank, Washington, DC. Foster, V. and Yepes, T. 2006. “Is Cost Recovery a Feasible Objective for Water and Electricity? The Latin American Experience.� Policy Research Working Paper 3943, World Bank, Washington, DC. Vennemo, Haakon, and Ornica Rosnes. 2008. “Powering-Up: Costing Power Infrastructure Investment Needs in Africa.� Background Paper 5, Africa Infrastructure Country Diagnostic, World Bank, Washington, DC. Wodon, Quentin. 2007. “Electricity Tariffs and the Poor: Case Studies from Sub-Saharan Africa.� Working Paper 11, Africa Infrastructure Country Diagnostic, World Bank, Washington, DC. World Bank. 2008. Zambia Growth Infrastructure and Investments: A Role for Public Private Partnership. Washington, DC: World Bank. 39 Zambia Electricity Regulator Board. 2008. Press Statement on the ERB Decision on ZESCO Application to Revise Electricity Tariffs, other Charges, Fees and Penalties. 40 Annexes Annex 1. Country coverage, country classification, and year of tariff schedule data set Economic/CPIA Capacity Generation classification Power pools level type Reference Year of the LIC fragile nonfragile Resource Thermal Medium Hydro WAPP CAPP EAPP SAPP High tariff MIC Low rich LIC Benin 1 1 1 1 2003 Botswana 1 1 1 1 2008 Burkina Faso 1 1 1 1 2006 Cameroon 1 1 1 1 2003 Cape Verde 1 1 1 1 2006 Chad 1 1 1 1 2005 Congo, Rep. of 1 1 1 1 2007 Congo, Dem. Rep. of 1 1 1 1 2005 Côte d’Ivoire 1 1 1 1 2006 Ethiopia 1 1 1 1 2004 Ghana 1 1 1 1 2006 Kenya 1 1 1 1 2006 Lesotho 1 1 1 1 2006 Madagascar 1 1 1 2005 Malawi 1 1 1 1 2006 Mali 1 1 1 1 2008 Mozambique 1 1 1 1 2006 Namibia 1 1 1 1 2006 Niger 1 1 1 2003 Nigeria 1 1 1 1 2005 Rwanda 1 1 1 1 2005 Senegal 1 1 1 1 2006 Seychelles 1 1 1 2006 Sudan 1 1 1 1 2003 Tanzania 1 1 1 1 2006 Uganda 1 1 1 1 2006 Zambia 1 1 1 1 2005 Zimbabwe 1 1 1 1 2008Dec Source: Africa Infrastructure Country Diagnostic Power Tariff Database. Note: LIC = low-income country; MIC = middle-income country; CAPP = Central Africa Power Pool; EAPP = East African Power Pool; SAPP = Southern African Power Pool; WAPP = West African Power Pool. 41 Annex 2. Residential tariff schedules Fixed charge/month, Fixed charge/month, Demand level, kWh Demand level, kVa Block border, kWh Number of blocks Block range, kWh kWh/kVa/month Demand charge, Demand charge, Price per block, Price per block, monthly, $ Tariff type LCU/kWh $/kWh LCU $ Country Benin IBT no no 3 <20 20 56 0.096 20–200 200 85 0.146 >200 95 0.163 Botswana FR 11.11 1.63 no 1 0.4 0.06 Burkina Faso IBT 582 1.11 3 <=50 50 96 0.184 >50–200 200 102 0.195 >200 109 0.208 Cameroon* IBT* no 3 <=50 50 50 0.086 50–200 200 60/67 0.109 >200 65/75 0.120 Cape Verde IBT no no 2 <=40 40 20 0.225 >40 25 0.280 Chad IBT no no 3 <=30 30 83 0.157 ?? 30–60 60 177 0.336 ?? >60 201 0.381 Congo, Dem. Rep. of** IBT no no 11 <=100 100 — 0.040 >100–200 200 — 0.039 >200–300 300 — 0.039 >300–400 400 — 0.039 >400–500 500 — 0.038 >500–600 600 — 0.038 <=600 600 — 0.089 >600–800 800 — 0.088 >800– 1,000 1,000 — 0.087 >1,000– 1,200 1,200 — 0.086 >1,200 — 0.085 Congo, Rep. of FR 2,268 5.06 no 1 49.08 0.11 Côte d’Ivoire IBT 333 0.64 2 <=40 40 36 0.069 >40 74 0.142 Ethiopia*** IBT cons levels for fixed charge 7 <=50 50 0.27 0.032 1.40 0.16 0–25 >50–100 100 0.36 0.041 3.40 0.39 26–50 >50–100 100 0.50 0.058 3.82 0.44 51–105 >100–200 200 0.55 0.064 10.24 1.19 105–300 >200–300 300 0.57 0.066 13.65 1.58 301+ >300–400 400 0.59 0.068 >400 0.69 0.080 Ghana IBT 5,000 0.54 3 <=300 300 700 0.076 >300–700 700 1,200 0.131 >700 1,400 0.153 Kenya 2000 IBT 75 1.04 4 <=50 50 1.6 0.021 >50=300 300 6.7 0.092 >300– 3,000 3,000 7.0 0.097 >3,000– 7,000 7,000 13.8 0.191 Kenya adjusted IBT n.a. 1.74 4 <=50 50 n.a. 0.049 >50=300 300 n.a. 0.212 >300– 3,000 3,000 n.a. 0.223 >3,000– 7,000 7,000 n.a. 0.440 Kenya 2008 120.00 1.74 <=50 50 2 0.029 >50=1,500 1500 8.1 0.117 >1,500 18.57 0.269 Lesotho FR no no 1 0.49 0.072 Madagascar FR 5,962 2.98 1 152 0.076 Malawi billing IBT 124.71 0.92 3 <=30 30 2.7 0.020 >30–750 750 3.9 0.029 >750 5.6 0.041 prepayment FR no no 1 4.2 0.031 Malawi 2009 billing IBT 124.71 1.05 <=30 30 2.7 0.023 >30–750 750 3.9 0.033 42 Fixed charge/month, Fixed charge/month, Demand level, kWh Demand level, kVa Block border, kWh Number of blocks Block range, kWh kWh/kVa/month Demand charge, Demand charge, Price per block, Price per block, monthly, $ Tariff type LCU/kWh $/kWh LCU $ Country >750 5.6 0.047 prepayment FR no no 4.2481 0.04 Mali IBT no no 4 <=200 200 119 0.27 >200 139 0.31 Mozambique billling IBT 70,799 2.79 4 >=100 100 1,010 0.040 100–200 200 2,198 0.087 >200–500 500 2,929 0.115 >500 3,077 0.121 prepayment FR no no 1 2,802.0 0.110 Namibia FR no no 1 0.79 0.117 Niger FR 250 0.43 1 79.25 0.136 Nigeria*** IBT 20 0.15 <5 1,084.1 5 <20 20 1.2 0.009 30 0.23 >5–15 3,252.2 >20–60 60 4 0.030 120 0.91 >15–45 9,756.5 >60–180 180 6 0.046 >180– 5,000 38.09 >45–500 108,405 2,000 2,000 8.5 0.065 >500– >2,000– 31,250 238.06 20,000 80,000 80,000 8.5 0.065 Rwanda FR no no 1 81.25 0.146 Senegal UDS, special domestic customers (poor) no no 3 >20 95.48 0.183 20–44 106.55 0.204 >44 62 0.119 UDG, general domestic customers VDT no no 3 >20 20 120 0.230 20–44 44 87 0.17 >44 62 0.119 Senegal 2008 IBT no no <150 150 106 0.238 151–250 250 114 0.255 >250 117 0.262 South Africa IBT no no 2 <=50 50 — — >50 0.49 0.072 Tanzania**** IBT low usage 2 50 40 0.032 128 0.102 general usage FR 1,892.00 1.51 >=275 >=275 106 0.085 Tanzania 2008 IBT low usage 2 50 49 0.041 156 0.130 general usage FR 2,303.00 1.93 >=275 >=275 129 0.108 Uganda IBT 2,000 1.09 2 <=15 15 62 0.034 >15 426 0.233 Zambia IBT 5,845 1.31 3 >=300 300 70 0.016 <300–700 700 100 0.022 >700 163 0.037 Zimbabwe IBT 3 <50 50 29,289.15 0.01 51–500 500 389,020.02 0.08 >500 661,469.58 0.13 Source: Africa Infrastructure Country Diagnostic Power Tariff Database. Note: kVa = kilowatt-ampere, LCU = [[?]], kWh = kilowatt-hour, IBT = increasing block tariff, FR = [[?]], UDG= [[?]], UDS= [[?]], VDT= [[?]]. * Cameroon: Each price applies to all consumed within the corresponding consumption range, as in tariff 1 = unit price if cons<50; tariff 2 = unit price if 50200; second tariff is dry-season tariff, dry season is from January to June. *** Ethiopia: Consumption levels for fixed charge: 0–25, 26–50, 51–105, 105–300, 301+ kWh. **** Nigeria: Consumption levels for fixed charge: <5, >5–15, >15–45, >45–500, >500–20,000 kVa. ***** Tanzania: General usage fixed charge is applicable if consumption reaches or exceeds 275 kWh/month and is not charged below it. ** Congo, Dem. Rep. of: According to a Project Appraisal Document of 2007, average residential tariff in 2005 was 1.2 cents/kWh; collected revenue was 0.4 cents. Tariff was increased early 2007 by 50 percent to 1.7 cents/kWh. n.a. Not applicable. — Not available. 43 Annex 3. Commercial tariff schedules Demand charge per kWh, Price per block, LCU/kWh Threshold of power that Fixed charge/ month, $ Threshold of power for Price per block, $/kWh can be received (kWh) Fixed charge/ month, demand charge (kVa) Demand charge per Demand charge per Number of blocks KVa/month, LCU KVa/month, $ Type of tariff Block range LCU $ Country Benin FR 1 88.00 0.15 Botswana FR 30 4.36 1 0.46 0.07 Burkina Faso TOU 1,169 2.24 2,882 5.51 0.023 2 (10 am–2 pm and 4 pm–7 pm) 165.00 0.32 (12 am–10 am/2 pm–4 pm/7 pm–12 am) 88.00 0.17 180 kWh/kVa of subscribed Cameroon* DBT 2,000 3.44 0.013 2 load 63/68 0.11 >1,801 kWh/kVa of subscribed load 55/60 0.10 Cape Verde FR 1 19.20 0.22 Chad IBT 8,055 15.27 0.056 3 <=30 84.00 0.16 ???30–60 186.00 0.35 ???>60 211.00 0.40 Congo, Dem. Rep. of DBT 5 200 0.111 500 0.11 1,000 0.109 1,500 0.108 >1,500 0.107 Congo, Rep. of FR 3,972 8.87 1 43.56 0.10 Côte d'Ivoire DBT 1,882 3.60 2 <=180 * kVa bimonthly 97.09 0.19 >180 * kVa bimonthly 83.25 0.16 Ethiopia TOU 122 14.12 3 equivalent flat rate 0.58 0.067 peak hour 0.74 0.086 off-peak hour 0.54 0.063 Ghana IBT 25,000 2.72 3 300 1,020.00 0.11 600 1,250.00 0.14 >600 1,450.00 0.16 Kenya FR 150 2.08 1 >7,000 6.70 0.09 Kenya adjusted n.a. 3.47 1 n.a. 0.21 Lesotho FR 133 19.64 0.081 1 0.08 0.012 Madagascar FR 101,271 50.56 13,370 6.67 0.027 1 338.44 0.169 Malawi FR 1,509 11.10 961 7.07 0.026 1 4.09 0.03 Malawi 2009 1,509 Mali 104.00 0.23 Mozambique FR 207,308 8.16 105,973 4.17 0.017 1 1,378.00 0.05 Namibia FR 75 11.08 80 11.81 0.044 1 0.57 0.08 Niger FR 1,500 2.58 1,000 1.72 0.007 1 70.71 0.12 44 Demand charge per kWh, Price per block, LCU/kWh Threshold of power that Fixed charge/ month, $ Threshold of power for Price per block, $/kWh can be received (kWh) demand charge (kVa) Fixed charge/ month, Demand charge per Demand charge per Number of blocks KVa/month, LCU KVa/month, $ Type of tariff Block range LCU $ Country Nigeria IBT 5–15 3,252 90 0.69 4 6.50 0.05 15–45 9,756 120 0.91 8.50 0.06 55– 500 108,405 240 1.83 8.50 0.06 500– 2,000 80,000 250 1.90 0.007 8.50 0.06 Rwanda FR 1 95.88 0.17 Senegal TOU 4,023 8.98 2 regular hours 88.84 0.20 peak hour 142.15 0.32 South Africa IBT <=25 5,420 227 33.58 3 0.27 0.04 25–50 10,841 276 40.80 0.27 0.04 50– 100 21,681 430 63.53 0.27 0.04 FR <=25 5,420 no 1 0.64 0.09 Tanzania FR 6,615 5.28 7,245 5.79 0.021 1 66.00 0.05 Uganda TOU 2,000 1.09 1 398.80 0.22 Zambia FR 15 29,227 6.55 1 163.00 0.04 Source: Africa Infrastructure Country Diagnostic Power Tariff Database. Note: TOU = time-of-use tariff; DBT = decreasing block tariff. *Cameroon: fixed charge is 2,500 per kilowatt if subscribed load is up to 200 hours and 4,200 per kilowatt if it is above 200 hours; second tariff is dry-season tariff, dry season is from January to June. n.a. Not applicable 45 Annex 4. Industrial tariff schedules Fixed charge/ month, Fixed charge/ month, Threshold of power Demand charge for Demand charge for 100 kVA/mo, $/mo Demand charge per Demand charge per Threshold of power 10 kVA/mo, $/mo Number of Blocks KVA/month, USD KVA/month, LCU demanded (kVA) demanded (V) Price per block, Price per block, Type of tariff Block range LCU/kWh $/kWh LCU $ Country Benin FR 1 62.00 0.107 Botswana 30 4.36 59 8.61 86.11 861 1 0.21 0.03 (10am–2pm & 4pm– Burkina Faso TOU 1,050 2.01 5,962 11.40 114.02 1,140 2 7pm) 118.00 0.226 (12am–10am/2pm– 4pm/7pm–12am) 54.00 0.103 >200 Cameroon TOU hours 108 0.19 2,778 4.78 47.79 478 2 (11pm-6pm) 40/61.25 0.087 (6pm-11pm) 40/50 0.085 Cape Verde FR 1 15.60 0.177 Chad TOU 8,055 15 152.71 1,527 3 regular hours 108.00 0.205 night hours peak hours 200.00 0.379 Congo (DRC) DBT 5 200 0.152 500 0.150 1,000 0.149 1,500 0.148 >1500 0.146 Congo, Rep. FR 1,260 2.81 15 0.03 0.35 3 1 50.16 0.11 (7:30am–7:30pm, Côte d'Ivoire TOU 3,303 6.32 3 11pm–12am) 55.71 0.107 (7:30pm–11pm) 75.95 0.144 (11pm–7:30am) 46.09 0.088 Ethiopia TOU 116 13.39 3 equiv. flat rate 0.41 0.047 peak hour 0.51 0.059 off-peak hour 0.39 0.046 Ghana FR 125,000 13.62 90,000 9.81 98.10 981 1 500.00 0.054 Kenya DBT 240-415 600 8.32 300 4.16 41.61 416 3 5.16 0.072 11,000- 33,000 2,000 27.74 200 2.77 4.60 0.064 66,000- 132,000 7,500 104.02 100 1.39 4.40 0.061 Kenya adjusted nap 13.90 nap 6.95 69.49 463 3 nap 0.164 nap 46.32 nap 4.63 nap 0.147 nap 173.72 nap 2.32 nap 0.140 Lesotho FR 147 21.76 217.58 2,176 1 0.07 0.011 Madagascar FR 1,137,264 567.77 — — — 1 199.00 0.099 Malawi FR 1,455 10.70 899 6.61 66.06 661 1 3.28 0.024 Mali TOU 1,471 3.28 1 75.75 0.169 Mozambique FR 973,079 38.31 131,794 5.19 51.89 519 1 1,144.00 0.045 Namibia (Nampower) FR 324 47.85 75 11.04 110.36 1,104 1 0.84 0.124 46 Fixed charge/ month, Fixed charge/ month, Threshold of power Demand charge for Demand charge for 100 kVA/mo, $/mo Demand charge per Demand charge per Threshold of power 10 kVA/mo, $/mo Number of Blocks KVA/month, USD KVA/month, LCU demanded (kVA) demanded (V) Price per block, Price per block, Type of tariff Block range LCU/kWh $/kWh LCU $ Country Niger FR 15,000 25.81 2,778 4.78 47.79 478 1 51.22 0.088 Nigeria IBT 5-15 90 0.69 175 5 6.50 0.050 15-45 120 0.91 8.50 0.065 55-500 240 1.83 230 1.75 8.50 0.065 500- 2000 250 1.90 8.50 0.065 >2000 270 2.06 8.50 0.065 Rwanda FR 1 95.88 0.172 Senegal TOU 9,855 22.01 2 58.01 0.130 83.54 0.187 South Africa TOU <=100 159 23.48 6.74 1.00 9.95 100 2 June-August 0.18 0.026 100- 500 547 80.72 6.74 1.00 Sept-May 0.12 0.018 500- 1000 3,131 462.30 6.74 1.00 >1000 3,131 462.30 6.74 1.00 Tanzania FR 7,012 5.60 7,123 5.69 56.90 569 1 61.00 0.049 Uganda TOU 20,000 10.92 5000 2.73 27.30 273 1 369.70 0.167 Zambia DBT 16-300 78,002 17.48 6,943 1.56 15.56 291 4 1,200 100.00 0.022 300- 2000 136,003 30.47 12,990 2.91 8,000 85.00 0.019 2000- 7500 272,006 60.94 19,587 4.39 30,000 63.00 0.014 >7500 544,012 121.88 19,696 4.41 52.00 0.012 Source: Africa Infrastructure Country Diagnostic Power Tariff Database. n.a. Not applicable. — Not available. 47 Annex 5. Representative schedule used in calculations Residential Commercial Industrial Social Public lighting Benin Electricite basse Professionnel (BT2) Electricite moyenne The first 20 kWh Electricite basse tension, Domestique (client categorie: tension, Tarif 2 (client within tarif tension, Eclaraige (BT1) (client commercial) categorie: moyenne domestique are called publique (BT3) (client categorie: menages industries, force "tranche sociale") categorie: lumieres et motrice brasseries) municipalites) climatisation) Botswana TOU 4 domestic TOU 6 small business TOU 8 large business n.a. — Burkina Faso Basse tension, Basse tension, double Moyenne tension, No tariff named Tariff type F: Eclairage monophase 2 fils, tarif, categorie: Tarifs categorie: Tarifs “social.� Used the public categorie: Usage horaires particuliers horaires particuliers following tariff as domestique et administration, et administration, social: Basse tension, particuliers et tarif type D1 tarif type E2 monophase 2 fils, administration, tarif (nonindustrial) (industrial) categorie: Usage type B (monophase) domestique particuliers et administration, tarif type A (monophase) Cameroon LV Domestic LV Business MV tariffs No tariff titled No such tariff in the subscribers subscribers "social." Used first schedule tranche residential as social. Cape Verde BT: low voltage BT Especial: low MT: medium voltage No tariff titled Iluminaçao Publica: voltage special "social." Used first public lighting tranche residential as social. Chad Basse tension, Usage Basse tension, Gros Moyenne tension, No tariff titled Basse tension, domestique (I.1.a). clients (I.1.b). Tarif preferentiel "social." Used first Eclairage public. (I.2.b). tranche residential as social. Congo, Dem. Rep. of Clients avec Clients avec Clients avec Tariff titled "social": — compteur, clients compteur, clients compteur, clients Clients avec basse tension, basse tension, basse tension, force compteur, clients residentialle 1 and commerciale motrice basse tension, residentielle 2 sociale. Congo, Rep. of Tarifs en basse Tarifs en basse Tarifs en moyenne n.a. — tension, T1, Mono, tension, T7-1, Tri- tension et haute puisance souscrite phase, puisance tension, T13, (kW): 1,2. souscrite (kW): 12. 32.9 5 amperes et (puissance fils 5 amperes) compteurs 4 fils) souscrite<25 kW) Mozambique Domestic, customers LV large customers MV customers Used "tarifa sociale," — with conventional which coincides with meters the first tranche of the domestic tariff. Namibia Nored, prepaid Nored, business Nored, business three n.a. Nored, streetlights single phase phase Niger Basse tension, Basse tension, Moyenne tension, — Basse tension, Unique: Electricite Unique: K32 Longue utilisation: Unique: K34 usage domestique, average of K22 and K33.1 to K33.5 K23 Nigeria Residential category, Commercial Industrial No tariff titled Street lighting single phase "social." Used tariff titled "pensioners" instead. South Africa Homelight (for low- Business rate (for Miniflex (TOU for No tariff titled Public lighting usage residential small businesses in urban customers "social." Used first customers in urban urban areas, up to from 25 kVa to 5 MVa tranche residential areas), Homelight 1, 100 kVa), Business (free of charge) as prepaid (first 50 kWh rate 1 social. free) Senegal Tarif UP2 Tarif moyenne Tarif moyenne n.a. Eclairage Public BT tension, Tarif General tension, Tarif Longue (TG) Utilization (TLU) Tanzania Domestic low-usage Low-voltage High-voltage Domestic low-usage — tariff (D1) (for up to maximum-demand maximum-demand tariff 50 kWh) and general- tariff (T2) tariff (T3) usage tariff (T1) (for above 50 kWh) Uganda Code 10.2/10.3: Low Code 10.2/10.3: Low Code 20: Low voltage No tariff titled Code 50: Street voltage supply for voltage supply for supply for medium "social." Used first lighting small general services small general services scale industries tranche residential as (domestic) (commercial) social. Zambia Metered residential Commercial tariffs Maximum demand No tariff titled Street lighting tariffs (capacity 15 (capacity 15 kVa) tariffs, MD2— "social." Used first kVa) capacity 301 to 2000 tranche residential as kVa social. Source: Africa Infrastructure Country Diagnostic Power Tariff Database. Note: LV = low voltage, MV = medium voltage, BT = basse tension o baixa tensao, MT = moyen tension, KVa= kilovolt, HV = high voltage, MVa = megavolt. n.a. Not applicable. — Not available. 49 Annex 6. Methodological notes and inputs for historic cost calculation Calculating the historical capital costs of generation Step 1. Calculations are based on generation unit overnight investment costs per kilowatt. For oil-, coal-, and gas-based production, internationally accepted unit overnight investment costs ($/kW)9 are assumed. For hydroproduction, country-specific unit costs calculated as weighted averages10 of unit costs for hydropower projects in each country are applied.11 Step 2. Unit costs are discounted using the annualization factor assuming a 10 percent discount rate and a standard expected lifetime of the power plant, which differs depending on the type of generation.12 Step 3. Unit generation investment cost per kilowatt for each country is calculated considering the country-specific generation mix (percentage of each type of generation in total) and discounted unit costs produced at step 2. Step 4. Capital costs of generation per kilowatt-hour are calculated by multiplying country-specific discounted unit costs (step 3) by the country’s generation capacity and dividing by the country’s power generation. Calculation of historical capital costs of transmission and distribution Step 1. A proxy used for total lifetime investment is the overnight transmission and distribution (T&D) investment calculated under the assumption of constant 2005 access to power.13,14 Step 2. Since the scenario of constant 2005 access was run for the “trade expansion� option only, an adjustment is made to exclude the cost of the new cross-border transmission lines. This was done using the annualized cross-border investment as a share of the total. Step 3. Since constant 2005 access rates in the investment needs model is applied to the population in 2015, we adjusted the denominator (generation 2005) using—as a proxy for the generation increase— growth in the number of households between 2005 and 2015. Then we applied the annualization factor to come out with the present value of required annual future T&D investment per kilowatt-hour. 9 Source of the unit costs: AICD, BP5 (Investment Needs Paper); original sources: International Energy Agency, Energy Information Administration (United States), Royal Academy of Engineering (United Kingdom). 10 Weighted by plant capacity. 11 Source of unit costs for hydropower projects: AICD, BP5 (Investment Needs Paper). 12 For hydroplants, the assumed lifetime is 50 years; for coal plants, 25 years; and for oil and gas plants, 30 years. 13 With population growth, the number of people with access to power is increasing under this scenario, although the percentage of population with access is constant. 14 Source: AICD, BP5 (Investment Needs Paper). 50 Annex 7. Inputs for calculating historical costs Unit costs Economic ($/MW) lifetime (years) Discount rate 10% Generation Country Hydro specific 50 Coal 1,100 25 Gas 670 30 Oil 810 30 Country Transmission specific 40 Distribution 40 Sources: Rosnes and Vennemo, 2008 Note: MW = megawatt. 51 Generation investment unit Cost, $/kW Country T&D investment unit cost, cross-border, discounted, T&D overnight excluding Generation investment Generation, GWh/year Installed capacity, MW unit cost, cents/kWh Total cost per unit, Capex, cents/kWh Opex, cents/kWh cents/KWh cents/kWh $ million Oil Gas Coal Hydro Angola 830 3,722 126 4 810 670 1,100 1,966 5 9 Benin * 60 124 23 3 810 670 1,100 4,671 5 8 12 20 Botswana 132 631 15 1 810 670 1,100 1,496 1 2 12 14 Burkina Faso 236 516 15 2 810 670 1,100 4,767 8 11 4 15 Burundi 37 92 6 3 810 670 1,100 3,476 11 14 Cameroon 875 4,004 53 1 810 670 1,100 1,428 3 4 13 17 Cape Verde * 80 250 1 1 810 670 1,100 3,356 3 4 14 18 Central African Republic 40 115 1 1 810 670 1,100 1,500 4 5 Chad 29 117 2 2 810 670 1,100 1,568 3 4 9 14 Congo, Dem. Rep. of 2,443 7,193 46 1 810 670 1,100 644 2 3 4 7 Congo, Rep. of 121 400 10 2 810 670 1,100 1,775 5 7 13 20 Côte d'Ivoire 1,084 5,524 61 2 810 670 1,100 2,283 3 4 7 11 Equatorial Guinea 13 28 1 2 810 670 1,100 2,292 6 7 Ethiopia 814 2,589 111 4 810 670 1,100 1,016 3 6 2 8 Gabon 415 1,774 6 0 810 670 1,100 3,356 5 6 Gambia 30 160 5 3 810 670 1,100 3,356 2 4 Ghana 1,490 6,750 103 2 810 670 1,100 2,098 3 5 8 12 Guinea 274 850 14 1 810 670 1,100 1,547 4 6 Guinea-Bissau 21 65 1 1 810 670 1,100 4,100 3 4 Kenya 1,312 5,347 96 2 810 670 1,100 2,889 4 6 8 14 Lesotho 76 410 5 1 810 670 1,100 1,938 3 4 6 11 Liberia 188 350 4 1 810 670 1,100 4,158 5 6 Madagascar 227 973 20 2 810 670 1,100 1,496 3 4 11 15 Malawi 285 1,368 10 1 810 670 1,100 1,488 3 3 6 9 Mali 280 515 29 5 810 670 1,100 3,225 12 17 16 34 Mauritius 688 2,321 13 1 810 670 1,100 1,496 2 3 Mozambique* 2,383 15,914 25 0 810 670 1,100 1,432 3 3 6 9 Namibia 264 1,580 57 2 810 670 1,100 1,778 2 4 7 11 Niger 145 202 7 1 810 670 1,100 3,356 8 9 23 32 Nigeria 5,898 24,079 1,132 5 810 670 1,100 1,222 2 7 2 10 Rwanda 39 116 13 4 810 670 1,100 1,930 5 10 7 17 Senegal 509 2,105 231 2 810 670 1,100 3,356 3 6 19 25 Sierra Leone 50 80 15 2 810 670 1,100 3,089 8 11 South Africa 41,904 228,071 981 0 810 670 1,100 1,496 2 3 3 6 Sudan 961 4,341 92 3 810 670 1,100 2,509 3 6 Tanzania 919 1,880 44 2 810 670 1,100 1,957 4 6 8 14 Togo 85 230 8 1 810 670 1,100 2,387 10 11 Uganda 303 1,893 35 2 810 670 1,100 2,377 4 5 5 10 Zambia 1,700 8,850 43 0 810 670 1,100 1,336 2 3 4 7 Zimbabwe 2,099 8,890 61 0 810 670 1,100 1,386 3 3 Sources: Rosnes and Vennemo, 2008 Note: GWh = gigawatt-hour 52 Annex 8. Historic unit cost of power cents/kWh Total T&D capital Generation capital Operating Country cost capital cost cost cost Total cost Angola 4.4 4.8 9.2 Benin* 3.1 5.1 8.2 11.6 19.8 Botswana 0.6 1.4 2.0 11.9 13.9 Burkina Faso 2.3 8.4 10.7 4.4 15.1 Burundi 3.2 11.1 14.3 Cameroon 1.3 3.1 4.4 12.7 17.1 Cape Verde* 0.5 3.1 3.6 14.3 17.9 Central African Republic 1.0 4.3 5.3 Chad 1.6 2.7 4.2 9.4 13.7 Congo, Dem. Rep. of 0.7 2.2 2.9 3.9 6.8 Congo, Rep. of 1.8 4.9 6.7 13.4 20.1 Côte d’Ivoire 1.6 2.8 4.4 6.6 10.9 Equatorial Guinea 1.7 5.7 7.4 Ethiopia 3.5 2.9 6.4 2.1 8.5 Gabon 0.4 5.4 5.7 Gambia 2.5 1.7 4.2 Ghana 1.5 3.3 4.8 7.5 12.4 Guinea 1.5 4.3 5.8 Guinea-Bissau 1.0 3.0 4.0 Kenya 1.5 4.3 5.8 8.4 14.2 Lesotho 1.2 3.2 4.5 6.4 10.8 Liberia 0.9 5.1 6.0 Madagascar 1.7 2.8 4.5 10.5 15.0 Malawi 0.6 2.6 3.2 5.9 9.1 Mali 5.1 12.2 17.3 16.3 33.6 Mauritius 0.6 2.0 2.6 Mozambique* 0.2 2.5 2.8 6.3 9.0 Namibia 1.7 2.4 4.0 7.3 11.3 Niger 1.1 7.7 8.8 23.4 32.1 Nigeria 5.3 2.2 7.5 2.2 9.7 Rwanda 4.4 5.5 9.8 6.8 16.6 Senegal 2.2 3.4 5.6 19.4 25.0 Sierra Leone 2.2 8.5 10.7 South Africa 0.5 2.1 2.6 3.4 6.0 Sudan 2.6 2.9 5.5 Tanzania 1.8 4.3 6.1 8.0 14.1 Togo 1.0 9.7 10.7 Uganda 1.6 3.5 5.1 5.3 10.4 Zambia 0.4 2.5 2.9 3.6 6.5 Zimbabwe 0.5 2.9 3.4 Sources: Rosnes and Vennemo, 2008 53 Annex 9. Value and volume of sales to residential customers as percentage of total Share of residential Share of residential supply sales (LCU) in total (GWh) in total Benin 48.7 Burkina Faso 63 63.1 Cameroon 60 32.8 Cape Verde 56.2 49.7 Chad 67 63.5 Congo, Dem. Rep. of 47.3 Congo, Rep. of 52.9 Côte d’Ivoire 46.9 34.5 Ethiopia 26.6 44.3 Ghana 64.8 42.8 Kenya 37.4 35.7 Lesotho 100 35.2 Madagascar 60 Malawi 36 Mali 64.9 Mozambique 42.8 47.4 Niger 58.7 99.9 Nigeria 39.1 51 Rwanda 5.5 Senegal 62.7 58.6 South Africa 17.2 7.5 Tanzania 47.6 43.6 Uganda 33.2 Zimbabwe 30.5 Source: Africa Infrastructure Country Diagnostic Power Tariff Database. 54 Annex 10. A calculation of long-run marginal costs The long-run marginal cost (LRMC) of power was calculated using the investment needs model developed under the umbrella of the Africa Infrastructure Country Diagnostic (Rossines and Vennemo 2008). The model is based on estimates of future increase in demand and cost of corresponding supply. It minimizes the total annualized cost of system expansion and operation. This includes the operation and maintenance (O&M) cost of producing and distributing electricity according to expanded demand, as well as the capital cost of refurbishing old capacity and constructing new capacity, including generation plants, cross-border transmission, and distribution and connection. The model is run under two trade scenarios (trade expansion, under which all economically viable cross- border transmission capacity is developed, and trade stagnation, under which no further cross-border transmission capacity is built) and three future access-rate assumptions (current access level, 35 percent access, and national access targets). As model outcomes, two sets of country-level LRMCs are produced: (i) LRMC under trade expansion, national access targets and (ii) LRMC under trade stagnation, national access targets. Some details of demand and cost of meeting demand estimations: Projecting power demand over 2005–15. Demand consists of (i) market demand associated with different levels of economic growth, structural change, and population growth; (ii) suppressed demand created by blackouts and practice of power rationing; and (iii) social demand, as expressed in political targets for increasing popular access to electricity. Based on historic trends, demand is projected to grow at 5 percent per year in Sub-Saharan Africa to reach levels of 680 terawatt-hours (TWh), including: at 4–5 percent per year in SAPP and EAPP, at 7 percent per year in CAPP, 9 percent per year in the island states, and 12 percent per year in WAPP. Cost of supply needed to meet the projected demand comprises cost of refurbishment, new construction, and O&M. The analysis covers thermal generation—natural gas, coal, heavy fuel oil, and diesel—and renewable generation technologies—large hydropower, mini-hydro, solar photovoltaic, and geothermal. Operation of current nuclear power is considered, but not as new investment.  Cost of refurbishment of existing capacity is estimated based on refurbishment needs of each country in megawatts (plant-specific data) and unit cost of refurbishment for thermo and hydro generation. For hydro generation, unit costs are based on estimated costs of actual planned hydropower projects in each region. Thermal power plant technology is generic and the unit costs are therefore the same across countries. The refurbishment requirements for T&D are based on asset age.  Cost of construction of new capacity for cross-border electricity transmission is estimated. As in case of refurbishment, unit cost of construction is standard for thermal plants and country specific for hydro plants. Cost of T&D construction equals line length times unit cost. Unit costs of lines to be built—per km and per megawatt—are country specific. For lines between countries, average unit costs of two countries are used. 55  O&M 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 10-year period. 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 LRMC calculation. 56 Annex 11. Effects on long-run marginal costs of trade expansion Installed capacity, MW Thermal capacity as % LRMC decrease (%) Reduction in LRMC Net exports, trade expansion, TWh of total (cents) Angola 45 5 (6.0) 41 843 Guinea-Bissau 44 7 (0.2) 100 65 Liberia 43 6 (1.7) 100 350 Chad 36 4 (1.3) 100 29 Mozambique 33 2 5.9 9 2,383 Burundi 27 4 (0.7) 3 92 Countries with LRMC reductions Congo 25 2 (4.4) 24 400 Equatorial Guinea 20 2 (0.1) 77 28 Niger 17 5 (1.5) 100 105 Lesotho 14 1 (0.7) 0 76 South Africa 14 1 (36.4) 91 40,481 Zimbabwe 11 1 (3.5) 64 8,890 Mali 11 3 (1.9) 45 515 Sierra Leone 10 1 (0.9) 92 80 Togo 9 1 (0.9) 21 230 Senegal 9 4 (1.4) 100 300 Namibia 8 1 (3.8) 6 393 Kenya 8 1 (2.8) 39 1,211 Burkina Faso 4 1 (1.0) 87 180 Benin 0 0 (0.9) 98 60 Botswana 0 0 (4.3) 100 132 Countries with no change on LRMC Central African Republic 0 0 — 53 115 Congo, Dem. Rep. of 0 0 51.9 1 2,443 Côte d’Ivoire 0 0 0.9 44 1,084 Gabon 0 0 (1.0) 59 1,774 Ghana 0 0 (9.6) 26 1,622 Malawi 0 0 (1.5) 8 285 Nigeria 0 0 2.1 67 5,898 Rwanda 0 0 10 31 Sudan 0 0 13.1 68 4,341 Zambia 0 0 (1.8) 0 1,778 Countries with LRMC Uganda –9 –1 2.8 26 321 Gambia –14 –1 0.1 100 160 increases Cameroon –17 –1 6.7 8 902 Guinea –17 –1 17.4 54 850 Ethiopia –19 –3 26.2 17 755 Tanzania –25 –2 2.4 39 881 Sources: Adapted from Rosnes and Vennemo, 2008 and AICD Power Tariffs Database 57 Annex 12. Average monthly electricity tab based on subsistence consumption Monthly electricity tab as a percentage Monthly electricity tab as a percentage of household budget (%): of household budget (%): Monthly electricity tab ($) connected households unconnected households Consumption 100 Consumption 100 Consumption 100 Consumption 50 Consumption 75 Consumption 50 Consumption 75 Consumption 50 Consumption 75 kWh/month kWh/month kWh/month kWh/month kWh/month kWh/month kWh/month kWh/month kWh/month Benin 12.6 13.3 13.6 5.0 8.0 10.9 9.4 14.9 20.3 Burkina Faso 20.6 20.2 20.0 4.4 6.5 8.5 13.4 19.8 26.2 Cameroon 8.6 10.9 10.9 3.1 5.8 7.8 6.0 11.4 15.2 Cape Verde 23.6 25.1 25.8 Chad 22.9 27.3 30.0 1.6 2.9 4.3 4.0 7.2 10.6 Congo, Dem. Rep. of 4.0 4.0 4.0 Côte d’Ivoire 9.6 11.1 11.9 1.5 2.7 3.8 3.3 5.7 8.1 Ethiopia 3.9 4.1 4.1 2.2 3.5 4.6 3.8 5.9 7.9 Ghana 8.7 8.4 8.2 2.1 3.0 3.9 3.3 4.8 6.2 Kenya 8.4 12.7 14.8 1.7 3.9 6.1 3.6 8.1 12.6 Lesotho 7.2 7.2 7.2 Madagascar 6.0 4.0 3.0 0.5 0.5 0.5 1.5 1.5 1.5 Malawi 4.8 4.3 4.0 1.9 2.6 3.2 3.7 4.9 6.2 Mozambique 9.6 7.7 6.8 3.2 3.9 4.5 8.7 10.5 12.3 Namibia 11.7 11.7 11.7 Niger 14.5 14.2 14.1 3.3 4.8 6.4 7.1 10.4 13.7 Nigeria 2.5 3.8 3.4 1.2 2.7 3.3 2.1 4.9 5.8 Rwanda 14.6 14.6 14.6 3.0 4.5 6.0 7.7 11.6 15.5 Senegal 18.6 16.4 15.2 3.1 4.0 5.0 5.9 7.8 9.7 South Africa — 2.4 3.6 — 0.3 0.6 — 1.4 2.7 Sudan Tanzania 3.2 5.5 6.7 1.6 4.2 6.7 2.9 7.5 12.1 Uganda 19.5 20.7 21.4 4.1 6.5 8.9 10.7 17.0 23.4 Zambia 4.2 3.3 2.9 1.2 1.4 1.6 2.8 3.4 3.9 Source: AICD Power Tariffs Database — Not available. 58 Annex 13. Social tariff schedules Price per Fixed charge Fixed charge block, Price per Type of tariff (LCU)/month ($)/month Block border LCU/kWh block, $/kWh Benin social tranche n.a. n.a. 56 0.10 Burkina Faso 1 to 3 A, tranche 1 94 0.18 75 0.14 Cameroon* tranche 1 residential 7,500 12.90 50 0.09 Cape Verde tranche 1 residential — — 20 0.23 Chad tranche 1 residential n.a. n.a. 83 0.16 Congo, Dem. Rep. of social tariff 2.65 0.01 n.a. 0.04 Côte d’Ivoire tranche 1 residential 333 0.64 36 0.07 Ethiopia tranche 1 residential 1.4 0.16 0.27 0.03 Ghana tranche 1 residential 5,000 0.54 700 0.08 Kenya tranche 1 residential — 1.74 n.a. 0.05 Lesotho — — — — — Madagascar economic tariff 600 0.30 25 120 0.06 553 0.28 Malawi tranche 1 residential 125 0.92 2.67 0.02 Mozambique tranche 1 residential n.a. n.a. 1,010 0.04 Namibia n.a. n.a. n.a. n.a. n.a. Niger — — — — — Nigeria "pensioners" 30 0.23 4 0.03 Rwanda — — — — — Senegal tranche 1 residential n.a. n.a. 150 106 0.24 South Africa tranche 1 residential n.a. n.a. 0 0 Sudan — — — — — Tanzania n.a. n.a. n.a. 38 0.03 Uganda tranche 1 residential 2000 1.09 62 0.03 Zambia tranche 1 residential 5,845 1.31 70 0.02 Congo, Rep. of n.a. n.a. n.a. n.a. n.a. Mali social tariff n.a. n.a. 50 59 0.13 100 91 0.20 200 107 0.24 >200 124 0.28 Botswana n.a. n.a. n.a. n.a. n.a. Zimbabwe tranche 1 residential n.a. n.a. 29,289 0.01 Source: AICD Power Tariffs Database n.a. Not applicable. — Not available. * Cameroon: fixed residential charge is 2,500 per kW if subscribed load is up to 200 hours and 4,200 per kW if it is above 200 hours. 59 Annex 14. Operational inefficiencies % of revenues % of GDP Undercollection of bills Undercollection of bills distribution losses distribution losses Transmission and Transmission and Underpricing Underpricing Overmanning Overmanning Benin 12.8 39.1 0.5 13.8 0.2 0.7 0.0 0.2 Botswana 0.7 138.7 61.1 0.0 1.8 0.8 — Burkina Faso 12.5 0.0 14.7 9.5 0.2 0.0 0.3 0.2 Cameroon 36.3 57.9 0.0 8.3 0.8 1.2 0.0 0.2 Cape Verde 8.1 0.0 29.6 20.8 0.2 0.0 0.9 0.6 Chad 11.0 0.0 9.1 23.6 0.0 0.0 0.0 0.1 Congo, Rep. of 63.1 30.9 21.0 30.9 0.6 0.3 0.2 0.3 Côte d’Ivoire — 0.0 417.1 24.2 — 0.0 4.4 0.3 Congo, Dem. Rep. of 163.6 201.6 0.0 — 1.3 1.6 0.0 — Ethiopia 18.6 33.5 6.3 — 0.2 0.3 0.1 — Ghana 26.5 52.4 2.1 — 0.7 1.5 0.1 — Kenya 9.1 0.0 34.6 5.1 0.3 0.0 1.1 0.2 Lesotho 16.9 32.5 19.5 — 0.3 0.6 0.3 — Madagascar 5.0 2.3 0.0 — 0.3 0.2 0.0 — Malawi 40.5 105.3 75.1 — 0.5 1.3 0.9 — Mali 23.4 36.8 39.1 6.4 0.6 1.0 1.0 0.2 Mozambique 19.9 15.0 4.6 17.7 0.3 0.2 0.1 0.3 Namibia 51.6 0.0 — — 0.1 0.0 — — Niger 39.1 116.5 0.0 12.5 0.6 1.8 0.0 0.2 Nigeria 76.8 195.1 50.3 — 0.4 1.0 0.3 — Rwanda 10.8 9.3 0.0 6.8 0.2 0.1 0.0 0.1 Senegal 9.6 0.0 10.8 5.4 0.3 0.0 0.3 0.2 South Africa 0.0 5.9 0.0 — 0.0 1.0 0.0 — Tanzania 33.5 90.9 0.0 6.1 0.5 1.3 0.0 0.1 Uganda 34.6 0.0 39.4 5.2 0.6 0.0 0.7 0.1 Zambia 2.9 72.9 2.3 — 0.0 1.2 0.0 — Source: Briceño-Garmendia, Smits, and Foster, 2008 — Not available. 60 About AICD and its country reports 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 provides a baseline against which future improvements in infrastructure services can be measured, making it possible to monitor the results achieved from donor support. It also offers a solid empirical foundation for prioritizing investments and designing policy reforms in Africa’s infrastructure sectors. the AICD is based on an unprecedented effort to collect detailed economic and technical data on African infrastructure. The project has produced a series of original reports on public expenditure, spending needs, and sector performance in each of the main infrastructure sectors, including energy, information and communication technologies, irrigation, transport, and water and sanitation. Africa’s Infrastructure—A Time for Transformation, published by the World Bank and the Agence Française de Développement in November 2009, synthesized the most significant findings of those reports. The focus of the AICD country reports is on benchmarking sector performance and quantifying the main financing and efficiency gaps at the country level. These reports are particularly relevant to national policy makers and development partners working on specific countries. The AICD was commissioned by the Infrastructure Consortium for Africa following the 2005 G8 (Group of Eight) summit at Gleneagles, Scotland, which flagged the importance of scaling up donor finance for infrastructure in support of Africa’s development. The AICD’s first phase 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 was expanded to include as many of the remaining 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� is used throughout this report as a shorthand for “Sub-Saharan Africa.� 61 The World Bank has implemented the 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 (AfDB), the Development Bank of Southern Africa (DBSA), and major infrastructure donors. Financing for the AICD is provided by a multidonor trust fund to which the main contributors are the United Kingdom’s Department for International Development (DFID), the Public Private Infrastructure Advisory Facility (PPIAF), Agence Française de Développement (AFD), the European Commission, and Germany’s Entwicklungsbank (KfW). 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 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. The data underlying the 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. Many AICD outputs will appear in the World Bank’s Policy Research Working Papers series. Inquiries concerning the availability of data sets should be directed to the volume editors at the World Bank in Washington, DC. 62