Vie2wp o int i!rr;if No. 209 May 74 I;II Does Reform of Energy Sector Networks Improve Access for the Poor? Stepbenz Powtell Unless energy can be produced and delivered more cheaply, it will stay beyond the reach of many andrkas of the poor. For energy delivered through networks, the costs that matter are not only the unit Starks energy costs, but the costs of extending the network-into an urban slum, for example, or to a rural town. Extending a network can be very expensive-a major barrier to access for poor households and small or isolated communities. A central goal of the reform of electricity and gas networks, now occurring in an increasing number of developed and developing countries, is to provide incentives to reduce the costs of producing energy and getting it to consumers. New technologies in electricity are drastically reducing costs. But transmission costs are still a major hurdle to expanding networks in isolated or lightly populated areas. As a result it is the urban poor who stand the greatest chance of benefiting from network reform. For the rural poor, alternative solutions, including mini-grid and off-grid services, may be required. Electricity reform is based on the premise that mar- electricity cheaper for the poor who already ket mechanisms supply electricity much more effi- have access to it, and provide it to those who ciently than central planning can.1 But how will do not. It analyzes the fundamental costs of gen- the poor, who have the least purchasing power, erating electricity and distributing it through a fare in a competitive electricity market? Will those grid to rural and poor populations. It describes without access continue to be denied it as elec- how electricity reform and technological devel- tricity supply changes from a public service to a opments have reduced these costs in recent profit-seeking business? And will the poor who years and discusses institutional arrangements have access suddenly find it unaffordable? to ensure that lower costs are passed to cus- tomers. It examines whether reform will One response to these general concerns is that a increase access to electricity for poor house- well-functioning power sector is crucial to macro- holds and comments on policies to further the economic stability and growth. It is precisely interests of the poor in the context of electricity because poorly run, state-owned electric utilities reform. have been such an impediment to growth that so many countries are trying to reform them. For As context for this discussion, map 1 shows the those who believe that the best way to make the extent of power sector reform in developing poor richer is to make everybody richer, that is countries. Network reform is most advanced in how electricity reform helps the poor. Latin America, least advanced in Africa, and is quite mixed in Eastern Europe and South and However, this Note concentrates on the micro- East Asia (also see Izaguirre 2000 for private par- economic effects: whether reform will make ticipation in energy by country). The World Bank Group * Private Sector and Infrastructure Network 2 Does Reform of Energy Sector Networks Improve Access for the Poor? MAP 1 SCORECARD FOR POWER SECTOR REFORM IN DEVELOPING COUNTRIES, 1998 IBRD 3H 0MAY 2000 9 1 This map was pndcad by she Map Osign uail of The Wasd lank. Th. ba... sdlem, dnimatin.asn ay'' , a iforimi shw ml this map d.amso imply, as th. pars t ofThe Wodd hats Greep, nay Judgmermalno lIse legal ratms af asp tersitrsy, or any enlDarsoast or acceptlasof a sabh boundaries The World Bank surveyed power reform ie 115 countries to see which of the followiag six reform slops had been toke n:NNumh er of reform steps •Corporatization. 0 0 a Laws permitting divestiture and onhuodling. C'1 Regalations. @ 2 * Vertical and horizontal anbundling. 3 a Laws permittieg concessions or greenfield investment. * Privatization of existing assets. CD4 Note that: The score does oot indicate the quality of reform or the sequence of steps taken. Countries not icl SUFvey a Fur simplicity, seccess in one region cuaints as saccess for the country. a Sectors under private ownership for lIt years are not considered part of the current reform movement and are omitted from the survey. Source: Bacon 1999. Generating and selling electricity: Box 1 summarizes the cost characteristics of the what it costs fouir fuinctions. it has been estimnated that in England and WX,ales generation accouints for The provision of electricity through a grid about 65 percent of the total cost of electricity, involves four functions: tratisimission 10 percent, distribution 20 percent, * Generation: converting primaryT energy into ancti suipply 5 percent (New~berv and Green electricitv. 1996). These proportions vary in different sys- * Transmission: the high-voltage, long-distance tentis. In particuilar, the start-up costs of a grid are transport of electricity, high and fixed, which means that grids have big * Distributtion:, the low-voltage transport of elec- economies of scale, in terms of both the num- tricity from the high-voltage system to the user. her of houiseholds connected and the amount of * Suppl,y: the selling of electricity to uisers- energy transMitted. Thuis for grid systems in metering, billing, and so on. dcx ,elopinig couintries, one mighit expect trans- mission andI distribution costs to he a greater This Note's main concern is wAith reformi of the propor-tion of the total. transmission and distribuition svstems-'the grid"-but it also dliscuisses the innovations in There are twokev points here. First, physical fac- electricity generation that made reform possible. tors m-ake the fixedi costs of transmission and dis- The World Bank Group 3 BOX 1 COST CHARACTERISTICS OF ELECTRICITY PROVISION Generation The costs comprise fixed capital costs and variable operational costs including fuel. Because each type of plant has a different balance between fixed and variable costs, for each type the optimal size-giving the maximum economies of scale-is different. Transmission Transmission costs cover building and maintaining the transmission system and operating it (dispatching plant and maintaining voltage and frequency within predetermined limits). The cost of building and maintaining the system depends on physical factors such as its size and the ter- rain. The cost of extending it depends on the expected peak demand, but once the grid is built, the cost is sunk and so does not vary with the number of users or the volume of electricity transmitted. The high fixed costs make it unprofitable for more than one transmission system to compete in an area. Furthermore, the technicalities of minute-to-minute balancing of supply and demand together with the high cost of system failure mean that the natural monopoly extends over the whole integrated system. Distribution As for transmission, the high fixed land low variable) costs depend primarily on the physical coverage of the system (both distance and terrain) and the level of local peak demand. However, because the operating function is much simpler (it does not involve generator dispatch), the economies of scale are not as great. A country that supports only one transmission system can support a number of (non-overlapping) distribution systems. Supply Many supply costs, such as bad debts and the costs of payment collection, vary with the number of customers. These costs are disproportion- ately high for low-income households, which are more likely to experience payment difficulties and suffer disconnection. But some supply costs are fixed: once supply has been extended to a village, the extra cost of reading another meter in that village is low. Supply costs vary with the distance of customers from the nearest demand center. The more remote and dispersed the customers, the more expensive it is to administer meter reading and bill collection centrally. tribution particularly high for grid extensions to As the demand for electricity increases, the fixed remote rural populations. Thc population den- costs can be spread. In dcveloping countries, sity in rural areas is typically low, which means howvever, it takes time for demand to grow once that the fixed costs are shared among relatively access is provided: people have to wire their few people. houses and buy electrical appliances before they start to buy electricity. Demand for elec- Second, the poor tend to have verv low demand tricitv entails both a switch (not necessarily for electricity, which means that the average cost complete) from other fuels for cooking, heating, per unit consumed will be high because the and lighting and new demand for electrical fixed costs are divided among few units. appliances such as televisions. Over time, as Furthermore, for the rural poor, this demand incomes rise, loads will increase. andl load fac- tends to be concentrated at peak tirmes (mainly tors will also rise as people buy appliances with in the evenings as people switch on lights). Since conistant loads such as refrigerators. However. the fixed costs of transmission and distribution this progression is difficult to predict and there- depend in part on peak demand, this demand fore the returns to investment in extension of pattern results in still higher costs for poor rural electricity grids to rural and poor people are populations. uncertain. These points are illustrated in table 1, which To summarize, providing access to electricity for gives indicative figures for the relative distribu- low-income households-in particular the exten- tion costs of connecting different numbers of sion of the grid to rural areas-depends critically rural households at different distances from the on the balance between the fixed and variable transmission system. The central column shows costs of transmission and distribution. The fun- the unit costs of distribution. The riglht-hand col- damental cost characteristics of grid provision do umn shows the unit costs including generation not faxvor the provision of access to rural and poor and high-voltage transmission. populations. Can reform make any difference? 4 Does Reform of Energy Sector Networks Improve Access for the Poor? TABLE 1 EFFECTS OF LINE LENGTH AND CONSUMPTION LEVELS ON THE RELATIVE COSTS OF ELECTRIFICATION IN INDONESIA (U.S. cents per kilowatt-hour) Generation and transmission 10 Medium-voltage extension and low-voltage distribution 3-kilometer spur line, 20 households 45 55 3-kilometer spur line, 50 households 20 30 1-kilometer spur line, 20 households 15 25 1-kilometer spur line, 50 households 7 17 Note:These costs are indicative averages for most developing countries with relativelyflat terrain. Afew countries are now adopt- ing new, lower-cost network designs. Source: World Beak 1996, p. 50. Buying electricity: why it is getting began to want to build power plant-large cheaper industrial users as well as independent power producers (IPPs). The recent wave of electricity reform was facili- tated by innovations in technology. Competition and private participation have had fur.her effects on costs. Rather than buying Generation equipment from a favored national supplier, as state-owned monopoly generators had done, 'Until the 1980s the electricitv industrv was new entrants import it if that means lower cost. viewed as a unified natural monopoly that pro- In turn, this has increased competition between duced and delivered electricity. For decades equipnment manufacturers, and thermal effi- economies of scale had increased in electricity ciency has increased, further pushing unit costs generation, reinforcing the view that it was a nat- down. The thermal efficiency of CCGT stations ural monopoly. is now nearly 60 percent (compared with 30 per- cent or more for other thermal stations ), and the In the 1980s improvements in turbine technology cost of installing the latest CCGT technology is were imported from the space progratn and mate- nowv about US$375450 per kilowatt.3 rials science and the price of gas fell (in part because of gas market liberalization in developed Thus generation market reform should cut costs economies). This had a radical effect on the and reduce prices for customers. Following the economics of generation: the fixed cost of introduction of competition in generation and installing a combined-cycle gas turbine (CCGT) the establishment of a bulk power market in plant in the early 1990s in the United Kingdom Argentina, bulk electricity prices have fallen was around USS600-650 per kilowatt. compared fairly consistently (figure 2). with US$750-800 for oil-fired plant, US$900-1,200 for coal plant, and USS2,250 for nuclear. Falling In other cases there have been difficulties, how- gas prices reduced the variable costs as well.2 ever. In England and WXales, for example, antici- pated reductions in bulk electricity prices failed Combined-cycle gas generating units of 50-100 to materialize after competition was introduced megawatts could by the 1990s be built and run and the bulk power market cstablished, even economically-at one-tenth the size of the ther- though primary fuel prices were falling (figure mal plants (1,000 mnegawatts or more) of the 3). Tiis has been blamed on the manipulation of 1980s (figure 1). This meant two things. First, bulk power prices by the larger generators. In generation could be a competitive activity even other words, competition was not fully func- in relatively small electricity systems. Second, tioning. Trading arrangements intended to elim- developers other than the state monopoly utility inate such behavior are planned. The World Bank Group 5 FIGURE 1 COST CURVES SHOWING OPTIMAL PLANT SIZE, 1930-90 U.S. dollars per megawatt Thermal plants 1930 Combined-cycle gas generation 1980 50 200 600 1,000 Megawatts Source: Hunt and Shuttleworth 1996. Some developing countries have also had ernments have begun to separate transmission, difficulties in harnessing the full benefits of IPPs. distribution. and supply. A key question in generation reform is how to set up a bulk power market that delivers the While transmission and distribution have in benefits of reduced costs while still attracting many cases been separated. and distribution private investors. Offering long-term power pur- split among a number of companies, both func- chase agreements to IPPs attracts investors, but tions retain their natural monopoly characteris- the greater the security (in terms of guarantecd tics in any one area because of their high fixed purchase volumes and prices) offered by the costs. However, the introduction of private par- contract, the less sharp the incentive for cost ticipation through competitive tendering for reduction and the less scope for the power pur- concessions (to identify the least-cost provider) chasing agency to adjust its purchasing to has captured many benefits in terms of lower achieve least-cost dispatch.4 costs. To ensure that the full benefits of competitive Increased competition in the equipment markets generation reach customers, it is necessary to has reduced the price of many of the fixed cost introduce competition in supply. If supply is pro- components. Installation has also proved vided through the local monopoly distribution cheaper w hen done by private contractors rather company, customers cannot shop around for than utility employees.9 cheaper electricity. The monopoly distribution company can shop around, but has no incentive More generally, the private sector is simply to do so as it can pass on generation costs to its more efficient as a consequence of its profit captive customers. However, competitive sup- seeking. For example, when private distribution pliers will need to purchase power as cheaply as began in Buenos Aires there was a dramatic possible, thus ensuring that lower generation reduction in theft. Since theft was particularly costs are passed to retail customers. prevalent in slum areas, this reduction in theft cut the difference between the cost of supply- Transmission and distribution ing these areas and the electricity tariff and enabled the distributor to supply slum areas Having recognized that the electricity industry with reduced subsidies (Albouy and Nadifi comprises a number of distinct functions, gov- 1999). 6 Does Reform of Energy Sector Networks Improve Access for the Poor? FIGURE 2 BULK ELECTRICITY PRICES IN ARGENTINA, 1992-99 U.S. dollars per megawatt-hour 60 50 40 30 20 August August August August August August August August 1992 1993 1994 1995 1996 1997 1998 1999 Source: CAMMESA 2000. Equipment costs can also be reducecl by relax- acted on in many countries (partial opening of the ing equipment specifications and adopting inter- market to supply competition is a requirement of national standards. In the United Kingdom, for the European Union directive on the single mar example, over the past five years the cost of elec- ket for electricity, for example). As a result of com- tric plant in real terms has fallen by 10-15 per- petition, in the United Kingdom1 the cost of meters cent (Fairbairn 2000). has fallen by 39 percent over the past five years. Ilowever, transmission and distribution remain However, competition in supply is so far con- local or national monopolies. This means, first, that fined largely to more developed markets, where incentives to reduce costs are not as sharp as they companies can offer a number of sutpply services wxould be under competition (although the profit (such as electricity and gas) together and can dif- motive supplies some incentive), and second, that ferentiate themselves by service quality and savings that are made will not be freely passed to brand. In developing countries the costs of sup- consumers. Therefore, where these monopolies ply can be reduced in other ways, notably are privately owned, regulation is necessarv. through increased local involvement. Employing someone to read meters in a village is cheaper if Incentive-based regulation, such as the CPI-X that person does not have to travel a long dis- price cap methodology, involves a balance tance from the nearest town. Local participation between giving utilities the incentive to reduce in bill collection and maintenance can also be c(osts and ensuring that cost reductions are effective. For example, in Bangladesh locally passed to the consumer. The utility keeps some managed cooperatives buy power from the grid of the savings, but must pass the rest to the con- and distribute it locally. They have a better sumer.i In the tJnited Kingdom incentive-based record on billing, maintenance, and reducing regulation has been broadly successful in reduc- losses than that of the main power utility in ing prices to domestic consumers, even though charge of urban distribution (World Bank 1996). bulk prices have not fallcn (figure 4). Electricity for the poor: does cheaper Supply mean better? The potential for competition in supply, which, Cheaper generation has reduced the total cost of with relativelv low fixed costs, is not a natural providing electricity. That should mean lower monopoly, has been recognized and is being prices for the poor who are already served by a The World Bank Group 7 FIGURE 3 AVERAGE ANNUAL DEMAND-WEIGHTED ELECTRICITY POOL OF ENGLAND AND WALES PURCHASE PRICE, 1991-2000 Pounds per megawatt-hour 30 25 20 15 10 5 0 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000a Note: The years refer to fiscal years, ending in March. a. Average for April to October 1999. Source: Electricity Pool of England and Wales 20. FIGURE 4 ELECTRICITY PRICE FOR STANDARD DOMESTIC TARIFF CUSTOMERS IN THE UNITED KINGDOM, 1990-99 Pence per kilowatt-hour (1989-90 prices) 8 6 4 2 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Note:The years refer to fiscal years, ending in March. Source: U.K. Department ol Trade, Statistical Office 1999. grid. Reductions in the fixed costs of transmis- affordable. One solution is to direct electricity sub- sion and distribution equipment, and innova- sidies much more precisely to the poor, for exam- tions to reduce the costs of supplying remote ple. through thc introduction of lifcline tariffs.8 areas, improve the prospects that grids will be More generally, the design of the tariff system is extended to rural areas. crucial in determining how the benefits of elec- tricity reform (in terms of lower costs) are distrib- However, there are two important caveats. First, uted among different customer classes. If for the poor to benefit, lower production costs electricity reform is to benefit the poor, tariff pol- must be passed on as lower prices. In many devel- icy must be designed with their needs in mind. oping countries tariffs have risen following reform as subsidies have been withdrawn (despite cost Second, the fixed costs of transmission and dis- reductions). In many respects this benefits the tribution equipment have not fallen enough to poor,' but it does make access to electricity less make it profitable to extend the grid to all areas. 8 Does Reform of Energy Sector Networks Improve Access for the Poor? Given the huge difference between cost of sup- particularly if the ftinds are redirected tossard social policies. For ply and (socially or politically) acceptable tariffs ftirer disctlsston of energy subsidies see 19orld Bank 1990 arc International Energy Agency 1999. for some rural populations, extensions of the Lifeline tariffs essentially invols e subsidizing electricitv only at rite grid to these people must be subsidized if they very loss levels of constirnption typical of poor hotiseholds The are to happen at all. There are two ways in suhsidies apply to secv small amotonts of electricitn and do not bycost too MLuch. This po)licy has been suiccessfuil in Titailandl: see which this can happen: within the utility bv luntivate and Barnes 199. cross-subsidy from profitable customers (under an obligation to extend service) or with subsi- dies from outside the utility, for example, from References a rural electrification fund. Albuy, Yves, and Nadia Nadifi. 1999. "Impact of Power Sector Reform on the Poor: A Resiews of IssLues ancl the Literatuire. ' borld Bank. Energy, Mining, and Telecommuinications Depanmert, Conclusion Washington. D.(- Bacon, Robert. 1995. 'Competitive Contracting for Privately Reform of grid-based electricitypr nCGenerated Posser.' Vies point 47. \Xorld Bank, Finance, Private Reformn of grid-based electricity provision will Sector, and InfrastrICtuIre NetworIk, Washington, D.C not revolutionize access by the poor. The cost --. 1999. "Global Energy Sector Refornm in Developing Countries: structure of grid provision, so unfavorable to A Scorecard.' World Bank, Energy Sector Management Assistance Frogramame, Washington, D.C. extending access to rural populations, is not fun- CANINILSA (Compania admic iictralora del mercado noasorista elec damentally altered by electricity reform. trico sociedad anonima) 2000 wwwv.canimesacom However, reform unambiguouslv moves the Eleciricitv Pool of England and Wales 2000. wsvw elecpuolicorn overall level of costs in the right direction. At the Fairbairn. Richard. 2000. "Analysis of Electrical Trade Price Indices Viewpoint is an open margin, cost reductions imply both increased PB Poser Ltd. forum intended to affordability of grid services and increased via- Hunt, Sally, and Graham ShUitlewornh. 1996. CompentitionanrdChoice encourage biiyii' E/ectnlcity Chichester, England: John Wiley and Sons bilitv of grid extensions. As long as the intro- International Energy Agency 1999 Lookhtg cit Eoergl Srmhaia/ies dissemination of and ItrainlEeg gny 99 okn tFeg,Sbil~ debate on ideas, duction of competition and profit-seeking Getting the Prices Right- Rot-ld Energy Oztlook 1999 Insights. Paris. innovations, and best private participation is combined with regulation Izagtiirre, Ada Karina. 2003. "Private Participation in Energys' practices for expanding and tariff design that is sensitive to the needs of Viewpoint 208 orld Bank, Posvare Sector and InfrastcrucCe the private sector. The Nemwork. WEashington. D.C. vievvs published in this the poor, electricity reform is a positive step. Nel bert-, David MI, and Richard Green. 1996. "The English Electricity series are those of the Industrv ." In Richard J. Gilben and Edssard P. Kahn, eds., authors and should not international Comparisors of Flectrcitc Regul/ation Cambridge: be attributed to the Reform of grid-based energy services has concentrated on the gen C:ambridge Universirs Press. World Bank or any of its eration and distribution of elecuricity. Electricity nets'orks are far Tuntivate. Voravate Tig. and Douglas F Barmes. 1997. I'¶hailand s affiliated organizations. more extensive rhan gas netisorks in most pails of the develop Approach to Rural Electrification' Hoss was It Successfuld" World Nor do any of the con- ing wsorld and reform of gas netwsorks has been less ssidespread Bank, Inergy Sector tlanagenment Assistance lirogramme, clusions represent This Note therefore discusses electricitv reforns, although many of Washington, D.C. official policy of the the important points apply to both indtistries, gisen the parallels U.K. Department of 'rade and Industry, Statistical Office. 1999 Dige:st World Bank or of its in terms of network econoniics. of U.K. Energy Statistics London. Executive Directors or 2 Although the wsidespread adoption of CCGT as the new technol- Wordcl Bank. 1996. RurralEnergl andDeve/opmentt Inproti ngEttergy the countries they ogy of choice was linked to the fall in the price of gas. the tech- Sitppliea for TAo Billion People. NVashington, D.C. represent. nology can run on diesel This discussion therefore also applies to countries ssith no access to gas To order additional 3 The cost estimates are from Richard Fairbairn of PB Power Ltd Stephen Powell (stephenpowellfnera.corn) and copies please call For a more detailed discussion of this issue see Bacon 1995. Mary Starks (mary.starks@nera.com), National 202 458 1 1 1 1 or contact This is one reason whv employment in the electricity industry has Economic Research Associates. London Suzanne Smith, editor, fallen dramatically following reformn This is a controversial social Room Fl 1 K-208, The effect of reform and one that has direct implications for the poor. World Bank, 1818 H ffowever, the subject is beyond the scope o this Note Street, NW, Washington, 6 CPI-X achieves this by fixing allowed prices for a given period, This .tote originally appeared as a chapter in D.C. 20433, or Internet dLoring sshich the utilitr can retain the profits arising from any Energy Sector ,14anagement Assistance address ssmith7@ cost reduction. At the end of this period the price cap is worldbank.org. The reviewed to ensure that over the long term the benefits are Programme (ESi1APJ, Energy and Development series is also available passed to consumers. Report 2000: Energy Services for the World's on-lhim lwfdldbe/)nk. Since energy subsidies are a larger proportion of GDP in many Poor (Washington, D.C.: World Bank, 2000). org/html/fpd/notes/). developing coLntries and benefit the swell-off more than the poor @ Printed on recycled (because the well off use more energy-, particularly electricity), For more inforrration on 1 A A D paper. reductions in subsidies will tend to benefit the poor in fiscal terms, ES AlAPgo to www.esmap.org. LtJaiI 11