GLOBAL127 ENVIRONMENT127 FACILITY R b ~ / WbekiInt I ap 'NLim '-ee`-\,6A- ' '- 7 RoberHE WQRLi'ANKs GEF Documentation The Global Environment Facility (GEF) assists developing countries to protect the,global environment in four areas: global warming, pollution of international waters, destruction of biodiversity, and depletion of the ozone layer. The GEF is jointly implemented bythe United Nations Development Programme, the United Nations Environment Programme, and the World Bank. GEF Working Papers - identified by the burgundy band on their covers - provide general information on the Facility's work and more specific information on methodological approaches; scientific and technical issues; and policy ahd strategic matters. G;EF Project Documents - identified by a green band - provide extended project- specific information. The implementing agency responsibie for each project is identified by its logo on the cover of the document. Reports by the Chairman - identified by a blue band - are prepared b-y the Office of the GE-F Administrator in collaboration with the three GEF implementing agencies for the biannual Participants' Meetings. GLOBAL ENVIRONMENT FACILITY The Cost-Effectiveness of GEF Projects Dennis Anderson Robert H. Williams Working Paper Number 6 UNEP THE WORLD BANK (D 1993 The Global Environment Facility 1818 H Street, NW Washington, DC 20433 USA All rights reserved Manufactured in the United States of America First printing December 1993 The views expressed in this paper are not necessarily those of the Global Environment Facility or its associated agencies. ISBN 1-884122-051 ISSN 1020-0894 The Cost-Effectiveness of GEF Projects This paper is the third among a series of GEF Working Papers to deal with the Program for Measuring Incremental Costs for the Environment (PRINCE). The GEF is a financial mechanism that provides grants to developing countries for projects aimed at protecting the global environment. PRINCE was initiated in February 1993 at a workshop held at the Tata Energy Research Institute in New Delhi. It covers methodological studies, field tests, and dissemination related to the technical issues of measuring incremental cost. This is a concept central to the GEF; the two conventions to which it is linked- the Framework Convention on Climate Change and the Convention on Biological Diversity; and the Montreal Protocol dealing with ozone depletion. This paper was prepared at the request of the Scientific and Technical Advisory Panel (STAP) of the GEF. It addresses such issues as the costs of carbon emissions (or their reduction) and their implications for project appraisal; the appropriate discount rate to be used in comparing costs, bearing in mind intergenerational concerns as well as cost-effectiveness; cost estimates of the benefits of innovation, particularly the contribution of investment towards reducing the cost of future investments when cost curves, as a function of investment, are steep; the role of Type I projects (where national economic benefits outweigh national costs) relative to Type II projects (which are not cost-effective from a national standpoint but provide global environmental benefits); and the possibilities of reducing transaction costs in certain GEF projects with economic potential. Dennis Anderson is Senior Adviser in the Industry and Energy Department of the World Bank. Dr. Robert H. Williams is Senior Research Scientist at the Center for Studies of Energy and the Environment at Princeton University. Thanks are due to Professor Amulya Reddy, Chairman of STAP, for initiating this paper and for his comments and encouragement; Professor David Pearce for reviewing the paper; and the participants of the PRINCE workshop chaired by Dr. Rajendra Pachauri. The other Working Papers currently in the PRINCE series are numbers 4, 5, 7 and 8. iii Contents Introduction 1 Investment Criteria 3 2 Costs and Portfolio Choice 11 3 Conclusions 20 References 22 Figures in text 1 Carbon accumulations and marginal costs 3 2 Transitions to the use of renewable energy as the GHG constraint is approached 4 3 Carbon accumulations' damage function 5 4 Scenarios for imputed carbon tax under different assumptions about initial estimates of the accumulations limit 10 5 Marginal costs of pollution abatement in electric power 11 6 Overlaps between Type I and Type II projects 15 Tables in text 1 Cost of electricity generation 7 2 Cost of gasoline and zero net CO2-emitting alternative automobile fuels 7 3 Cost comparison of alternative motor vehicies 8 4 Pollution abatement through price reforms in electricity demand and supply 12 5 Emissions control technologies: abatement efficiencies and costs 14 6 Historical photovoltaics cost data for Japan 17 7 Estimation of cost-saving benefits of GEF investments in photovoltaics 17 Appendix 25 Tables in appendix A l Cost of electricity generation 26 A2 Cost (ex-refinery or fuel processing plant) of gasoline and zero net CO2 -emitting alternative automobile fuels 28 A3 Cost of automotive fuels delivered to consumers 29 A4 Comparison of alternative automotive vehicles 30 A5 Characteristics of alternative vehicles 32 A6 Sensitivity analysis for carbon tax implied by backstop technology 33 A7 Cost summary for cruise-design fuel cell electric vehicle by Ira F. Kuhn 34 v Abbreviations BPEV Battery-powered electric vehicle FCEV Fuel cell-powered electric vehicle GJ Giga Joule GWh Gigawatt-hour (1 million kilowatt hours) ICEV (Gasoline-powered) internal combustion engine vehicle KW Kilowatt or KWe where "e" denotes an energy unit KWp Kilowatt peak MW Megawatt or MWe where "e" denotes an energy unit NPV Net present value OCC Opportunity cost of capital ODA Official development assistance OECD Organization for Economic Cooperation and Development PM Particulate matter PV Photovoltaic STAP Scientific Technical and Advisory Panel (of the Global Environment Facility) TW Terawatt Data notes: Dollars ($) are current US dollars unless otherwise indicated I short ton = 0.9072 tonnes I long ton= 1.0161 tonnes Introduction The Global Environment Facility's (GEF) energy- duction and use by 20 percent, and developing related investments aim to develop approaches for countries (whose per capita energy consumption is dealing with global warming. While there is still only one-tenth of that of the industrial countries) much scientific disagreement on the extent and were asked to reduce their rate of growth of energy likely consequences of global warming, the GEF consumption from 6 percent per year to 4 percent a represents a commitment by its member countries year (a major reduction), then this could be achieved to putting precautionary policies in place in the through reforms to energy pricing policies and event of such climate change. Indeed, as further other measures to improve energy efficiency. evidence on globai warming and its consequences is The global warming problem would be effectively gathered, the investments of the GEF will leave the "solved" at a negative cost by eliminating unneces- international community better placed to reduce sary economic waste in energy consumption.2 How- carbon accumulations to safe levels over the long ever, even with such improvements in efficiency, term. The GEF supports activities, technologies, global CO2 emissions each year would still be twice and policies that would be adopted on a large scale their present levels in forty years, and carbon accu- if it became necessary to significantly restrict car- mulations would likewise be twice their present bon emissions. In the Pilot Phase of the Facility, the levels by the middle of the century; the global main emphasis was on testing various approaches warming problem would have been delayed by a that were considered technologically promising; decade or so, but would remain substantially these approaches were reviewed by the Scientific unresolved. To fulfill its purpose, GEF funding will and Technical Advisory Panel (STAP) of the GEF need to be based on a scenario in which carbon in May 1992.1 As the Facility moves closer to GEF accumulations in the atmosphere are stabilized at II-its full-scale operational phase-the emphasis some safe level over the long term to avoid further has shifted toward finding cost-effective solutions global warming. to global warming and other issues. Introducing a carbon accumulations constraint into In seeking cost-effective approaches, the terms "ef- the analysis of energy investments amounts to using fectiveness" and "costs" both require scrutiny. If, the resource depletion rule often used in the past for for example, the member countries of the Organiza- fossil fuels. The rule has become discredited in tion for Economic Cooperation and Development recent years, because each time a limit to reserves (OECD) were required to gradually reduce their net was thought to be approaching, the limit was ex- carbon dioxide (CO2) emissions from energy pro- tended by new discoveries. But if a limit is set to the I STAP report (GEF 1992). 2 World Development Report 1992. use of fossil fuels, not by their availability, but by supported by the GEF as it is in other areas. Some the amount of carbon the atmosphere can safely groundrules will have to be setas to whatcomprises absorb, the imputed cost (or shadow price) of using a satisfactory portfolio of projects that meet the such fuels rises at a rate equal to the discount rate criteria of cost-effectiveness. This issue is discussed until the limit is reached. At this point they can be in chapter 2. replaced by altemative energy approaches, known as backstop technologies, which do not result in net A satisfactory portfolio also cannot be detennined carbon emissions. This means that the shadow price without analyzing how relative costs are changing. to be attached to carbon savings is undiscounited in The most promising investments in renewable ener- net present value comparisons of costs, at least up to gy are still small scale, and costs are declining with the point where the backstop technologies are wide- successive pilot investments and with technical ly deployed. Chapter 1 develops the argument fur- progress. The transaction costs of demonstrating ther, discusses possible complexities, and makes a and developing new approaches are also initially preliminary assessment of shadow prices suitable high. Thus, it is important to know the current cost for the appraisal of GEF projects. of such approaches, the prospects for reductions in costs of the technologies in question, and how the Anotherproblem for setting GEF criteria and ground GEF can help reduce such costs. A related issue is rules relates to the diversity and costs of potential that of the continuity of policies, which many peo- projects. All GEF finance could be consumed many ple have drawn to the attention of the GEF and the times over on a single option, such as promoting World Bank. Cost reductions will not be achieved energy efficiency, or developing aparticularrenew- by one-off investments, but only by a long-term able energy technology. This would preclude many commitment to a program of investments to devel- other investments that meet the criteria of cost- op the more promising technologies. Such a pro- effectiveness discussed below, and thus would un- gram will also help to reduce transaction costs that dermine the goal of instituting precautionary policies. currently hamper the success of many GEF invest- Risky situations require diverse portfolios, a rule ments. These issues are also discussed in chapter 2. thatis as relevantforthe energy developmentprojects Chapter 3 presents the conclusions. 2 Investment Criteria Global warming as a carbon stant (the assumption is relaxed later). If a limit accumulations problem were to be set on the safe level of accumulations, From a policy-making standpoint, the problem then there would eventually have to be a switch to posed by global warming (or the threat of it) is best the non-fossil alternatives, and the marginal costs treated as an issue of stabilizing carbon accumula- of meeting demand would rise from f to n. tions in the atmosphere. Several studies have con- centrated on how best to reduce annual emission rates by a particular period; for instance, it has been | found that most OECD economies could stabilize Figure 1. Carbon accumulations and (and probably reduce) their emissions by the year marginal costs 2000 simply by improving energy efficiency. Magn l n However, even if ambitious emissions reduction Marnrl _ targets were set, CO2 would continue to accumulate f in the atmosphere because net emissions rates in the T tim (t) industrial countries would still be large while those in developing countries would continue to grow. '18tions Even if all energy efficiency options were thor- oughly explored, the question would remain of how to control carbon accumulations in the long term. I This is the question the GEF needs to consider. The development and use of non-net carbon-emitting technologies--mainly renewable energy-will be The transition to the non-fossil fuel alternatives crucial, with energy efficiency playing a supporting could not be abrupt. The lead times and lags in- role. volved in developing and introducing the technolo- gies on a large scale would be considerable, perhaps The way the carbon accumulations constraint af- one-half to three-quarters of a century (all the more fects prices and costs is summarized in figure 1. reasonwhytheyneedtobedevelopedquickly),and The upper half shows the marginal costs of fossil the actual situation would probably look more like energy (f) and of the non-fossil alternatives (n) or that in figure 2 than figure 1. But based on figure 1, backstop technologies, the lower half shows car- each extra unit of fossil fuel that is consumed before bon accumulations rising over time as fossil fuel is the accumulations constraint is reached brings for- burned. For simplicity, f and n are assumed con- ward thetime whenthebackstoptechnologies would 3 be needed (the shaded area). The present value of the extra cost is then Figure 2. Transitions to the use of renewable energy as the GHG ct = (n-f)(1+r)-(T-) constraint is approached and the actual marginal cost of fossil fuel consump- ERCENTAGE_OF tOTAL PRIMARY ENERGY DEMAND tion is f, + ct-3 cis the carbon tax that is theoretically 0 needed to bring about investment in the non-carbonro0 alternatives; it is also the shadow price or marginal ?O. benefit to be attributed to any activity (such as energy efficiency, or the use of a low carbon- 0. emitting energy resource) that would delay the time ,0 E at which the carbon accumulations constraint is 2 - AORO A NUCLEAR reached-the value of buying time, so to speak. 1ti0 201 o 2030 2050 YEAkR The shadow price rises exponentially at a rate equal to the discount rate until it equals the difference TOTAL PRIMARY ENERGYDEMAND between the marginal cost of the backstop technol- afO ogies and fossil fuels. Thus, if it is estimated to be ,00 / $20 per metric tonne of carbon, it would be $52 per 2f0 .SI tonne in ten years' time (assuming a 10 percent i 200 discount rate) and $130 per tonne in twenty years' 'so: .WAL. time, and so forth, until the upper limit (given by n) 100/ is reached. An alternative would be to leave the net NYORO AND NUCE carbon benefits undiscounted in the comparison of . - . 0 2 the present worth of costs between one option and .2AR another. another. CAR80N EMISSIONS Accumulations constraints and environmental damage functions What is the relation between the above approach 20 _ and an approach based on an analysis of the social FOSSIL FUEL C costs of global warning that some economists have advocated?4 The two approaches are consistent in .0 principle, and are related as follows. The social cost .ENIWAIE ENERGY SCENARIO curve (or damage function) associated with carbon e - . . , . 0 accumulations is not known reliably, but may have YEAR2 a trough-like formn, as illustrated in figure 3; too low a level of carbon accumulations would lead to Source: Anderson and bird global cooling, and too high a level to global warm- N equivalent ing. Presumably the curve also shifts to the right or eq__va __n_. left according to changes in the earth's orbit, wob- ble, and tilt, which may help to explain past changes If the location, width, and curvatures of the trough in climate.5 were known reliably, it would be possible in theory I This result is similar to Hfotelling's tomrula for the optimal price of a depletable resource, whereby the optimal price of the resource equals the extraction cost plus a "user cost" that rises exponentially over time until the backstop or replacement technology becomes economical. In the present case, the resource in question is not the availability of fossil fuels, but the limits on their use set by concerns about climate change. For example, Nordhaus (1991), Barbier and Pearce (1990), and Cline (1991). Goudie, Environm ental Change, 2nd ed. (I1983). 4 The cost of capital and GEF criteria Figure 3. Carbon accumulations' A general principle of environmental policy-making damage function is that policies that tackle an environmental problem directly are both less costly and more effective than those that deal with it indirectly. Thus taxes or Margirnal regulations on pollution, such as on the sulfur or lead Costs content of vehicle fuels, or on the treatment and Safety Margin disposal of spent nuclear fuels, are far more effective \i _l ,,lin reducing pollution than, say, a general tax or a \ ~ - - -- (n-f) restriction on energy use. Indirect measures may \ | / 1 penalize clean and dirty fuels alike, and have only L 1f , small effects on pollution, while raising costs appre- 0 Accumulations ciably. By contrast, direct measures may reduce pollution to low levels at a comparatively low cost. to choose an accumulations limit appropriate to our A similar principle applies to proposals to lower the environmental epoch, at which the marginal costs, test discount rate when deciding on investments n - f, of maintaining the accumulations at a given intended to deal with global warming. It works both level equal the present value of the marginal social as a blunt instrument of policy and may also lead to costs to the world's communities of climate chang- decisions that contradict the aims of the policy;7 it es associated with carbon accumulations, as indi- favors capital intensive solutions over labor inten- cated in figure 3-not an unambitious objective for sive ones, and may sometimes work against invest- decision-makers. The intersection of the damage ments with more immediate promise (such as function and marginal costs gives an optimal level renewables) by giving added weight to those with of accumulations under certainty. But since the more distant prospects (such as nuclear fusion). In damage function is not known, some safe limit may general, lowering the discount rate is no substitute eventually have to be decided upon based on the for direct measures to address a pollution problem, findings of climate research and environmental and will not guarantee the required results. While science. If there is also evidence of instabilities blunt instruments may sometimes serve a useful arising, for example, from the positive feedback purpose, the preferred approach is to reflect environ- between global warming and carbon releases from mental concerns directly in policy and, with due sea and land masses overriding the various negative attention to the welfare of all parties (including feedbacks-such that the marginal damage func- future generations), to use a discount rate equal to the tion is thought to rise steeply-then presumably a opportunity cost of capital. safety margin would need to be incorporated in any agreed limit. This is the approach suggested for the GEF. Hence the question is, how are the international and inter- The marginal damage function may be a good deal generational concerns about global warming best flatter over the short term than over the long term, reflected in its policies? Two rules might usefully be due to the thermal inertia of the sea and land masses. followed. If so, it is possible that we may overshoot the accumulations constraint, and it would be neces- The first rule is to base the GEF portfolio on the types sary to initiate a process of decumulation.6 In these of investments that would most likely be needed in a circumstances, there would be a premium or rental scenario of global warming. The GEF then supports element to be added to the carbon tax or shadow a precautionary policy in which the aim is to leave the price formula noted above. international community well placed to respond to 6 According to some estimates, net carbon fluxes to the atmosphere would be negative if those from fossil fuels were reduced to less than about 2 to 3 billion tonnes per year (Bolin in Bolin et al. 1986; Hall and Rosillo-Caille 1990.) Thus decumulations are possible in principle if net emissions from energy related activities were reduced to 0 to 2 billion tonnes per year. 7 Markandya and Pearce (1988, 1991). 5 the global warming problem should the need arise. costs of the fossil fuel alternative, and the equaliz- Middle-of-the-road scenarios, such as those on which ing discount (r*) is calculated, a value of r* greater several cost-benefit studies have been based, are not than the opportunity cost of capital (OCC) will relevant to the GEF. Such scenarios overlook the indicate that the project meets the criterion of cost- asymmetry of risks (noted in figure 2), and also the effectiveness. (A value of r* < OCC would indicate possibility of a much worse outcome than they that the project is an outlier.) Suppose we have two project. The idea of a precautionary policy, like projects with different capital, operating, and main- insurance policies, is to prepare for downside risks. tenance costs, but both reduce total emissions by the Further, given the ambiguities in the evidence about same amount. Then the shadow value of reducing the possible extent and consequences of climate the emissions, even if the reductions are achieved at change, and that some relationships (such as ocean- different points in time, will be the same in both climate interactions and ocean currents) have still cases (because the annual values are undiscounted). not been modeled and estimated in a scientifically However, the present values of their capital, operat- satisfactory way, the economist has to pay as much ing, and maintenance costs will differ and they will attention to the variance as to the expected value of be competing on the basis of relative costs-on the the independent variable. This too argues for con- most cost-effective use of capital and recurrent sidering a downside scenario. Finally, if GEF in- resources-which is what the GEF's sponsors are vestments are based on a seemingly risk-neutral seeking. position or a most-likely outcome, this would work against the purpose of the Facility, which is to Lead times, lags and changes in costs prepare for contingencies. To estimate the shadow price (c,), assumptions need to be made about the time when the switch to the The second rule is to base GEF decisions on a backstop technologies would need to be begun in resource depletion formula of the form just derived. earnest in a scenario of global warning, and on the Some of the simplifying assumptions that were prospective costs of the backstop technologies rel- made above will be considered in the next two ative to those of fossil fuels. sections, and preliminary estimates of costs (the shadow price to be placed on carbon emissions) will Assumptions about the switch (T) then be derived. The formula gives an appropriate As noted earlier, it would take a long time to switch weight to those options that are consistent with to the backstop technologies in order to comply stabilizing (or even reducing) carbon accumula- with the carbon emissions constraints, and the situ- tions over the long tern. In net present value com- ation may be more like that shown in figure 2 than parisons of costs, the shadow price on carbon in figure 1. World demands for fossil fuels are very emissions is then undiscounted up to the point large-currently about 7.5 billion tonnes of oil where T is reached, since its value rises exponen- equivalent energy a year-while the output of the tially with the discount rate. (This is in accordance renewable energy industry amounts to much less with the above equation calculating the shadow than 1 percent of this (excluding hydropower, which price.) provides about 2.7 percent of primary energy sup- plies). The industry would not be geared up for The shadow price on emissions reductions is undis- substitution on a large scale for several decades, counted because of the accumulations constraint even assuming that the backstop technologies were and the decision to choose cost-effectiveness as the fully developed. criterion as opposed to the net present value of benefits, due to the difficulties in estimating the Thus carbon accumulations could not be stopped latter. The net present value of costs, under the abruptly as indicated earlier (by the solid lines in the criterion just noted, will be the present value of lower quadrant of figure 1), but would likely over- capital, operating, and maintenance costs minus the shoot agreed levels before they could be stabi- shadow value of reducing emissions. When this lized-unless it turns out that T is several decades expression is compared with the present value of the away, which would give ample time to plan ahead. 6 In this case it is better to think of T as a point in time applications, renewable energy is already the least- when the transition to the alternatives from fossil cost option-for example, the use of biomass for fuels would need to begin on a large scale. The year cogeneration, wind energy in favorable locations, 2010 for the present calculations is chosen for two photovoltaics for rural electrification, and the pro- reasons. First, major cost reductions and technical vision of supplementary power in electricity distri- developments in the manufacture and use of renew- bution networks. But substituting renewable energy able energy technologies (the main option that the for fossil fuels on a large scale would likely raise GEF can support) should be realized by then. Sec- costs. Table 1 presents one assessment of the long- ond, the evidence on the greenhouse effect will also term costs of using renewables for electric power be much clearer and (in a scenario of global warm- generation on a large scale. The details are provided ing) will enable governments and industry to make in the Appendix. a firm commitment to the widespread use of non- fossil fuels. Many agree that, for electricity generation, the backstop technologies may eventually become com- Marginal costs petitive with fossil fuels, at least in regions where The costs of the backstop technologies relative to solar radiation is consistently high.8 Technologies fossil fuels vary greatly for different markets and to provide substitutes for non-electric energy will, applications. They also vary over time. For some however, be very expensive. Non-electric energy currently comprises 60 percent of the primary ener- gy markets in the industrial countries and over 65 percent in developing countries. The main backstop Table 1. Cost of electricity generation technologies are biomass-derived fuels, primarily US centslKWh (1 990 prices) ethanol or methanol; hydrogen derived from nucle- [ Long-term | ar or solar power; and further electrification of the |Source ofpower Current expectations I energy markets, which will depend crucially on advances in storage technologies. Table 2 summa- Coal 5.0 May rise j rizes a recent assessment of likely costs of the Oil 6.0 gradually with various fuels delivered to the consumer. Gas (combined cycle) 4.5 fuel prices Nuclear 5.5 Rises with Ethanol and methanol from lignocellulosic (woody- environmental biomass) feedstocks could become competitive with factors gasoline in the long term if (ex-refinery) gasoline Photovoltaicsa 30-50 7.0 Thermal-solarb 15.0 7.0 Biomass 9.0 4.0-6.O Bio-s I I Table 2. Cost of gasoline and zero net Sources: See Appendix table A I for further details. Ref- CO -emitting alternative automobile fuels erence should also be made to Johansson et al. (1992) b o for renewables in general; Booth and Elliott (1990) for ! ____________________________________ biomass; and the OECD (1989) for nuclear and fossil Prospecrivel fuels. Anderson and Bird (1992) also provide a review Source of power Current long term i of estimates. Gasoline (ex-refinerv) 33 45 Note: The estimates shown in this table are averages of ranges which vary with the type of plant or technology, MEthanol from biomass 80-89 55-64 the discount rate assumed, and often, the country. The I Methanol from biomass 80-89 55-64 tigures are rounded. Hydrogen from High insolation areas (2,000 to 4,000 KWh/m/year) photovoltaic power 960 144-166 b Costs would rise with pressures on land resources Sources and basis of calculations: See Appendix table at high levels of production. A2. See Williams ,1990) and Johansson et al. (1992). 7 prices rise to the level indicated.9 But if biomass is lished in the market, would be comparable to to meet more than 50 percent of vehicle fuel require- those of gasoline vehicles ments, it will require intensive use of land as an * Their unit fuel costs would be greater, as summa- energy source. A typical net yield would be 300 rized in table 2 tonnes of ethanol per square kilometer, but might * Their fuel efficiency would be three times higher rise to around 700 tonnes with technical progress.", than the 20 percent attainable for the internal The total demand for vehicle fuels (diesel and combustion engine. gasoline) is around 1.7 billion tonnes of oil equiva- lent in developing countries and in OECD nations. Thus net costs of turning to hydrogen may be lower Demand from developing countries, which accounts than suggested in table 2. for 500 million tonnes, l is likely to triple in the next twenty years. Extremely large land areas would be Table 3 presents a reassessment of costs allowing needed for the cultivation of corn, sugarcane, and for these factors. Several long-term options are other crops if biomass fuels were to be substituted considered and are detailed in the Appendix. In each for gasoline and diesel on a scale large enough to case two quantities are calculated. One is the break- meet such demand. At the same time, the area even gasoline price, which is the gasoline price requirements of agriculture will rise appreciably (exclusive of taxes) at which the lifecycle cost of a with population growth and increasing per capita gasoline-powered internal combustion engine ve- incomes, depending on technical progress and yields hicle would equal the lifecycle cost of the alterna- in farming. tive vehicle. The other is the net lifecycle cost of reducing carbon emissions (in dollars per tonne of While biomass fuels are promising in terms of cost, carbon reduction). as the calculations in table 1 show, they would probably need to be complemented by solar-de- The use of two decimal places in the table is not rived hydrogen as a vehicle fuel at high levels of intended to indicate precision, but only to avoid the substitution, or by electrification of the vehicles, again with solar electricity being the primary ener- gy source. The advantage of these options are their Table 3. Cost comparison of alternative relatively low land intensity-the annual yields of motor vehicles solar schemes are over 30,000 tonnes of oil equiv- Undiscounted alent energy per square kilometer a year (in final net lifecycle energy units), assuming net conversion efficiencies Break-even cost of reducing of 5 percent, or fifty to one hundred times greater gasoline carbon than those of biomass. Type of price emissions' motor vehicle ($Igallon) ($/ton) It is, however, not sufficient to look at the cost of the Battery-powered 1.81 354 fuel alone. If hydrogen (produced, say, from photo- Fuel cell electric: voltaic-generated electricity) or biomass fuels were Methanol-based to become the premium vehicle fuel in a low- (biomass) 0.72 -175 carbon-emissions scenario, it would likely be in H2-based, using association with the use of a fuel cell in an electric [ biomass 0.83 -143 vehicle. The engineering economic studies summa- H2-based, using rized in the Appendix suggest the following: - PV electricity 1.62 117 Sources and basis of calculations: See Appendix. * The capital and maintenance costs of electric Assuming a gasoline price of $1.25 per gallon. vehicles using fuel cells, once they are estab- - See also the forthcoming review by Ahmed (1993) on the costs and status of solar and biomass energy, which reviews a large number of industry and other studies. Johansson et al (1992). British Petroleum Statistical Review of World Ener gy ( 1992). 8 unnecessary accumulation of rounding errors. In If the above figure of $120 per ton is taken as an fact, there is quite a margin of uncertainty in the initial working estimate of the marginal costs of the estimates. There are engineering-economic rea- (marginal) backstop technologies, and T to be year sons for thinking that the lifecycle costs of the fuel 2010 for the reasons discussed earlier, the present cell-powered electric vehicle (FCEV) could even- value of this would be $25 per ton, using a discount tually be lower than that of the gasoline-fueled rate of 10 percent. The appropriate shadow prices vehicle (as the two calculations in table 3 suggest would then be as follows (figures rounded): for FCEVs using biomass as the fuel source). However, the sensitivity analysis of the cost of fuel Year 1993 1995 2000 2005 2010 cells and photovoltaics (Appendix table A6) shows and after that the net costs could also be higher, depending on technological developments. Similarly, if gaso- line prices were lower than the $1.25 per gallon price ($Itonne) used in the analysis, the net costs of reducing emissions would also be higher. Only research, As noted earlier, the shadow price rises exponen- technical developments, and time will tell, and tially at the discount rate until T is reached. Such estimates are likely to be revised from time to time estimates are not rigid, and there would be good in the light of new problems and developments. reasons for revising the calculations periodically as Further suggestions on how best to deal with these new evidence on costs and on the greenhouse effect uncertainties are provided shortly. As a starting comes to light. point, however, a marginal cost of $120 a ton may provide a useful basis for analysis; this is roughly Uncertainties and risks the same as a carbon tax in the equivalent amount The main uncertainty is that the extent, likelihood of 30 cents per gallon being imputed to the use of and consequences of climate change are not known, vehicle fuels. even within broad limits. This does not, however, justify arbitrary ground rules for decision-making. Shadow prices Consider the risks at two levels: for climate change For the analysis of GEF projects, the shadow price and for the energy industry. to be attached to carbon emissions (ct) needs to be based on costs of the marginal backstop technolo- Suppose the absorptive capacity of the atmosphere gies, not on the costs of the most promising (non- were greater than is implied by the scenarios used marginal) options. This means that they are best for the above calculations, which are based on a based on costs in the non-electric markets (shown in fairly rapid development of the renewable energy tables 2 and 3), where the substitutes for fossil fuels alternative. If the time at which the greenhouse gas are likely to be more expensive, rather than in the accumulations limit is reached were to be delayed electricity markets, where the renewable energy by a decade, the shadow price (or imputed carbon options have good prospects of becoming compet- tax) would decline by 60 percent relative to the itive with fossil (and nuclear) fuels in the long termn. initial value (of $25 per ton) calculated above; The logic of this is that the more promising of the conversely, it would rise by a fraction of 2.5 if the backstop technologies, whose costs would be less limit were advanced by a decade. The situation than f,+c,, would then be given added weight when would thus be as summarized in figure 4. the net present value of costs are compared. At the same time, some marginal technologies with costs Decisions as to whether to raise or lower the shadow close to f,+ct would not be excluded-only the price ultimately depend on the findings of current outliers would be omitted, pending further develop- and future research on climate change. Such find- ments. Furthermore, those applications of the back- ings will also drive developments in the backstop stop technologies that have costs lower than f, (in technologies, which may justify the initial esti- the so-called niche markets) would have the highest mates of costs being raised or lowered. Thus any returns. policy response should keep options open, so that 9 Figure 4. Scenarios for imputed carbon tax important element of policy. But it is onl) an Figure 4. Scenarios for Imputed carbon tax under differen an aeconomically desirable element, and by itself will Eunder different assumptions about initial estimates ofthe accumulations limit notpreventcarbonemissions fromaccumulating in the atmosphere. Global warming could only be prevented by widespread recourse to the non-net carbon-emitting orbackstop technologies, the most Carb~on (a)promising of which are renewables. Tax or X * Hence the development and use of renewable ener- gy technologies merits the highest priority in the GEF portfolio, and their marginal costs set upper bounds to the shadow price (or imputed tax) to be a 2 '1990S placed on carbon emissions. Further, it is the costs (The growth rate in (a) - the discount rate) of the marginal technologies likely to be brought Initial estimates of the accumulations limit (a) correct into use-not the most promising ones with the (b) too high (c) too low. lowest costs-that set the upper bounds of the shadow price. Technologies at the margin, and of the shadow price can be raised or lowered relative course those below it, automatically qual.y for to the path initially agreed upon, based on the GEF finance-only outliers do not. Further, L xtra findings of climate research, energy studies and weight is given under this criterion to those appli- development. cations (such as in the use of photovoltaics for rura: electrification in favorable locations) that reduce The risks to the industries supported by the GEF any added costs of turning to renewables. should not be large. It is likely to be more than a * The shadow price needs to be based on accumu- decade before the findings of climate research es- lations constraint rather than on a yearly emis- tablish what the greenhouse gas accumulations lim- sions constraint. This means that it rises with the it is likely to be, and thus before the ground rules discount rate until the long-term cost of using the could justifiably be changed because of shifts in the marginal technologies is reached. The prelimi- perceived limit. There would thus be some continu- nary estimates made above suggest a figure of ity in policy. If the global warning problem proves about $25 per ton of CO2 rising at the discount to be somewhat of a false alarm, or less serious than, rate of 10 percent per year, up to $120 per ton. say, the projections of the Intergovernmental Panel * Shadow price estimates would be raised or low- on Climate Change (IPCC), technological develop- ered periodically as further evidence on costs is ments that already have useful market outlets would gathered. They would also eventually be raised have been stimulated, albeit on a larger scale than or lowered according to the emerging evidence would ideally have been required. on global warming. By basing its portfolio on the types of invest- A stock taking ments that would ultimately be needed in a sce- The GEF is part of a precautionary policy to address nario of global warming, and by working with an the problems posed by global warming. The Facil- accumulations constraint, the GEF would be ex- ity supports those activities and investments that plicitly reflecting intergenerational concerns and would leave its developing country members and risks in its portfolio. The Facility can then con- the international community better placed to reduce centrate on cost-effectiveness in its choice of carbon emissions and accumulations on a large investments-an aim that can best be served by scale, should the need arise. Several conclusions using the opportunity cost of capital as the test follow: discount rate. Technologies and practices with significant long-term potential for reducing car- In a global warming scenario, the achievement of bon emissions and accumulations would not be energy efficiency-perhaps the most discussed excluded by using the opportunity cost of capital option for reducing carbon emissions-will be an as the criterion. 10 Costs and 2 Portfolio Choice Energy efficiency and the backstop have a marginal economic benefit of approximately technologies 13 - 4 = 9 cents per KWh. CO2 and other pollutants In many developing countries, excess demands for would be reduced correspondingly. If prices were energy are being encouraged by large subsidies. For raised further, the marginal benefits (savings in electricity consumption, a recent survey of sixty- costs) would be less, but would still be positive so three utilities revealed that retail prices average long as the marginal costs of supply exceeded the little more than 4 cents per KWh while marginal marginal value of consumption (the price consum- costs, if the power systems were operating efficient- ers pay); pollution would decline correspondingly. ly, would average about 10 cents per KWh; the total The net benefits of raising prices (excluding the subsidies are thought to amount to over $100 billion economic benefits of reducing CO2 emissions) would a year, and are a source of both budgetary distress be zero at the point where prices reflected costs. and economic inefficiency.12 Moreover, this under- Beyond this point, further increases in prices would states the costs of poor pricing policies, since low tax consumers and, depending on the budgetary cash flows also lead the utilities to compromise on situation, sharply increase costs. Table 4 shows the required maintenance and the reinforcement of dis- calculation for various levels of abatement, and tribution networks; resulting electrical losses may figure 5 summarizes the results graphically. range from 20 percent to 40 percent of generation (compared to 10 percent attainable with good prac- tices). The thermal efficiencies of power stations are several percentage points below nameplate rat- Figure 5. Marginal costs of pollution ings, and, owing to poor maintenance, plant avail- abatement in electric power ability averages only about 60 percent (as compared (wh mf to *vel ) with 80 to 90 percent with good practices). Allow- 35 All ing for the managerial shortcomings induced by Energy EfficienCy N*O price inefficiencies, marginal costs are closer to 13 g 25 * PM cents than to 10 cents per KWh; thus the costs of Q 20 A 002 implied subsidies are much larger than the $100 M 15 Low Polluting billion just quoted. 10 ecnoge According to the above estimates, and taking present 10 2 40 50 60 70 80 90 100 day prices as a starting point, a 1 KWh reduction in .10 Pollution Abatement (A%) demand brought about by an increase in price would 12 World Bank Electric Power Policy Paper (1992). 11 Table 4. Pollution abatement through price reforms in electricity demand and supply (with special reference to developing countries) Price required Marginal cost Marginal cost P = P (I -A)" P, - pi including gains Abatement A % US centslKWh US centslKWh managerial efficiency 0 4.0 -6.0 -9.2 10 4.9 -5.1 -7.3 20 6.3 -3.7 -4.7 30 8.2 -1.8 -2.0 40 11.1 - 1.0 50 16.0 6.0 6.0 60 25.0 15.0 15.0 70 44.4 34.4 34.4 Basis: When prices are increased by a certain amount, the decrease in demand can be estimated from a standard price elasticity fornula. With unchanged technologies, the decrease in pollution is proportional to the decrease in demand. Alternatively, one can postulate a percentage level of abatement, relative to the case of unchanged prices, and estimate the price that would be needed to achieve it. This is the approach followed here, the results being shown in the first two columns. The demand function assumed is (Q/Qo) = (P/P)-', where Q, is the demand at price P,, Q. is the initial demand at prevailing prices P., and e is (the numerical value of) the price elasticity, which is taken to be 0.5, based on the survey by Bates and Moore (1992). Per unit abatement, A, is by definition (I-Q/Q.), so A = 1-(P,/P0) from which we get the pnce required to achieve A, shown in the second column. The benefits from reforming prices are calculated in two steps. The first, shown in column three, corresponds to the price efficiency benefits. The actual price consumers pay, P, , is the marginal benefit to them of an extra KWh of consumption; if P, represents the actual (unsubsidized) marginal cost (MC) of supply (sometimes called the efficiency price since P, = MC for efficiency), then the cost of reducing demand by one KWh is simply Pt - P,, and this is negative when, as occurs in the above case, P, < P,. (This calculation neglects the extemal benefits of reducing pollution, since we are comparing the cost-effectiveness of pollution abatement via energy efficiency and using low polluting technologies.) PO is taken to be 4 cents/KWh (slightly below the present average in developing countries) and P, 10 cents/KWh. Allow for what are sometimes called "X" or "managerial inefficiencies," after Liebenstein (1966). These tend to be correlated with shortages of finance, and can add 50 percent or more to costs. For Pt < P,, we have assumed they are proportional to the per unit distortion in prices, or (P, - P,)/(P,-P ), such that the closer P, is to P1, the lower the level of managerial inefficiency losses induced by price inefficiencies. Based on the evidence presented earlier, we have used a managerial inefficiency factor of M,=1-M (P,-P,)/(P,-P ), where M. = 0.35, so that when P,=P., managerial efficiency is only 65 percent of that when Pt=P,. Managerial inefficiency losses relative to best practice cases are assumed to be zero for P,> P,. (See Anderson and Cavendish (1992) for further discussion.) Also shown in figure 5 are the marginal costs and ing countries. In addition, there would be substan- long-term abatement efficiencies of low polluting tial economic gains, averaging about 5 cents per technologies. For turning to renewables on a large KWh (see figure 4) and totaling around $125 scale, the addition to marginal supply costs would billion a year (the current level of electricity probably not exceed 2 cents to 4 cents per KWh, but production in developing countries, which is 2,500 the long-term abatement efficiency would be 100 TWh, times $0.05). The benefits would rise over percent (table 5 gives further details and sources, time with demand growth. The additional effects and provides information on otherpollutants). Three of non-price-induced gains in efficiency, such as conclusions follow from such calculations: those arising from energy efficiency services, may conceivably push the "energy efficiency Energy efficiency gains through improvements frontier" out further, perhaps to the point where in price and managerial efficiency would likely pollution could be abated by 40 to 50 percent by reduce pollution by up to 40 percent in develop- a combination of price and institutional reforms. 13 1 3 At present, the additional contribution of non-price measures to improvements in energy efficiency is not known reliably, and merits research. 12 * Beyond a certain point, however, the cost curve * Afforestation programs on farmlands and water- rises rapidly-even tariffs of the order of 40 or 50 sheds (in which carbon-fixing is a by-product) cents per KWh, or four or five times the real cost * The substitution of commercial energy for tradi- of supplies, would not abate pollution by more tional cooking fuels than about 70 percent. Once the win-win options * The use of photovoltaics and micro-hydro for of energy efficiency are achieved, the most cost- electricity supply in remote areas where the cost effective measure by far is investment in renew- of diesel generators is high. able energy technologies which, as noted, are capable of achieving CO2 abatement levels of Type II projects are those for which the national 100 percent at a long-term cost of around 2 cents economic benefits are less than national economic to 4 cents per KWh.14 costs (NBNC). Projects falling under this category owing to the growth of per capita income, popu- include photovoltaics, solar-thermal, wind power, lation, urbanization, and the substitution of com- biomass gasifiers and gas turbines for power pro- mercial fuels for fuelwood. Without efficiency duction, sustainable biomass production to substi- reforms, demand and emissions would be four tute for fossil fuels, fuel cells, and various projects times their present level by 2005-20 10, and more that advance the "energy efficiency frontier." This than twice their present level even if ambitious final category includes alternative technologies in efficiency reforms were in place. Furthermore, lighting and water heating; advanced, high-efficien- given the low levels of per capita KWh consump- cy gas-turbine cycles; irrigation pump sets powered tion in developing countries, demands would by renewable energy; and methods for reducing the continue to grow well beyond this period. overall energy intensity of industrial processes. Thus in a scenario of global warming in which some The distinction between Type I and II projects is agreements on the safe limits to long-term carbon important because, as the World Development Re- accumulations are reached, recourse to renewable port 1992 concluded, it is essential that new interna- energy technologies would be unavoidable. There- tional financing for global environmental problems fore the development and use of these technologies not detract from the more urgent needs of economic must be a high priority in the GEF portfolio. development. The elimination of poverty and the achievement of economic stability and growth are Type I and Type 11 projects the main priorities for developing countries. Fur- In the GEF, Type I projects are defined as those for ther, the most serious environmental problems fac- which the national economic benefits (NB) are ing these countries are local, not global; they include greater than the national costs, and where there are the provision of water and sanitation (2 billion global benefits (GB) in the form of reductions in people are without access to satisfactory services), carbon emissions (NB>NC and GB>0). In other the abatement of local air and water pollution, the words, they are projects that developing countries protection of watersheds, and the prevention of soil could be expected to undertake in their own best erosion. economic interests, using official aid as and when conditions require and merit it, even if global warm- The GEF is supporting the applications of some ing were not an issue and the GEF did not exist. renewable technologies that have already estab- Examples include: lished a market niche in energy supply. They meet Type I criteria (NB > NC), and are likely to become * The economically efficient use of flared gas or an increasingly important source of energy and coalbed methane for power generation economic growth in the decades ahead; the best * Reducing waste in the production and use of examples are wind power and the use of photovol- energy through price and institutional reformns taic systems for small-scale applications such as 14 Some studies regard this to be a high estimate of costs. See Williams (1990). 13 Table 5. Emissions control technologies: abatement efficiencies and costs Emissions with and without policies Sector and Without Withl pollutant (index) without, % Costs Use of low polluting technologies Electric power: coal PM 100 <0.1 0.4 cents/KWh SO, 100 5 1.5 cents/KWh NO 100 5-10 0.5 cents/KWh Electric power: gas PM -0 -0 0.0 cents/KWh SO2 -0 -0 0.0 cents/KWh NO -0 5-10 0.5 cents/KWh Motor vehicles I Lead 100 0 4 cents/gallon NO 100 20 CO 100 5 12 cents/gallon VOCa 100 5 PM (diesels) 100 <10 4 cents/gallon Sulfur (diesels) 100 5 8 cents/gallon All fossil fuels CO2 (electricity) 100 0 2 to 4 cents/KWh CO2 (non-electricity markets) 100 0 $0.5 to $1.0/gallon Sources and notes: The abatement technologies and costs for electric power are reviewed in OECD (1989), Ken King (1991), Bates and Moore (1992) and Anderson (1991). For coal, the reference plant is a conventional boiler using 3 percent sulfur coals. With combined cycle plants for gas and coal (using coal gasification) the net costs can be negative, once efficiency gains are factored in. For motor vehicles, see OECD (1986, 1988) and Walsh (1990). For reducing CO2' the estimates are based on the costs of renewable energy; these are reviewed in Anderson (1991) and Anderson and Bird (1992), drawing on a variety of sources, including the U.S. Department of Energy's "The Potential of Renewable Energy: An Interlaboratory White Paper" (1990). The cost estimates shown probably err on the high side. a Volatile organic compound. water pumping, lighting, and rural electrification. tion is that, by giving more weight to Type II, the There are several government, industrial, and aca- GEF is helping to keep investment options open by demic studies of such technologies (several are supporting a more diverse portfolio of energy cited in this paper's references), including recent technologies than would otherwise exist. Several of reports of the World Energy Council (1992). Those the technologies used in Type I projects are well- engaged in research and development in industry suited to the circumstances of developing countries are looking to the GEF and to government programs (if only because solar insolations are much higher to support further developments and applications. than in the industrial countries), which gives Type Relative costs and prices have changed appreciably II GEF projects an important innovative function. with technical developments over the past twenty years and, as discussed below, there is a distinct Without the Type I-Type II distinction, the GEF's possibility that many applications, presently meet- resources could be rapidly absorbed by develop- ing only Type II criteria, will meet Type I criteria in ment projects currently financed by ODA. The the future, and thus become part of official develop- resources that can presently be allocated by the GEF ment assistance (ODA) and regular investment. to global warming projects amount to less than one Thus another reason for the Type I-Type II distinc- fiftieth of those provided by ODA for energy devel- 14 opment projects, and are also minute compared only meet Type II criteria to the point where they with investments in the energy sector, which cur- meet Type I criteria. Such developments should be rently exceed $100 billion a year for expanding encouraged. Figure 6 summarizes the situation. electricity supply alone. The GEF's resources are also small compared with the energy supply subsi- Aside from afforestation and some niche markets for dies in developing countries, which are estimated to the backstop technologies, Type I projects will gen- be $230 billion a year.'5 Put another way, Type I erally be those related to energy efficiency-using projects are already provided for by ODA whereas flared gas and coalbed methane for power genera- Type II projects are not, and the main contribution tion, for example, and an array of innovations to of the GEF will be to finance the latter. The danger improve end-use efficiency. Such projects have al- is that, if GEF resources are used regularly for Type most always been argued on the grounds that they I projects, the Facility's impact on Type II projects have good economic retums to investment, as in the would be greatly diluted, while its contribution to various demand-side management and integrated Type I projects would be small. resource planning approaches now being promoted in the United States.'6 They are being supported by Hence the distinction is sound, and the original idea ODA through structural adjustment and sectoral of the GEF concentrating on Type II projects re- reforms to improve institutional arrangements and mains justified. There will nevertheless be excep- price efficiency, and through the provision of fi- tional cases in which the GEF may advantageously nance for the development of energy efficiency consider some Type I projects. Such cases can be services in developing countries. Two recent World readily identified. As illustrated by the examples Bank policy papers have discussed the required noted above, the number of backstop technology policies and the role of official finance in developing applications that meet conventional economic cri- them in some detail.7 Type II projects, by contrast, teria is growing rapidly. Thus, if the GEF is blended will generally be the backstop technologies. with commercial finance and ODA in programs that combine Type I and Type II applications, the devel- To sum up, the following procedures and ground opmental and transaction costs faced by GEF users rules are suggested for GEF projects: in developing their projects elsewhere will be great- ly reduced, and the scope and effectiveness of the * First, compare the net present value of the costs of GEF greatly enhanced. the proposed project with the best fossil-fuel alter- native; the shadow prices to be placed on the carbon 1Technical developments in the backstop technolo- emissions of the latter can be estimated as discussed gies, including those listed earlier, are reducing earlier. This will determine whether the project is costs and thus graduating projects that presently cost-effective according to GEF criteria. Figure 6. Overlaps between Type I and Type II projects Tvpe I Type II Energy efficiency Backstop technologies Afforestation (non-commercial applications) Type Type Backstop technologies I (commercial applications) 1 5 World Development Repon' 1992 and Shah and Larsen (1992). [A See Hirst and Goldman (1991). : 7 "The Bank's Role in the Electric Power Sector: Policies for Effective Institutional, Regulatory and Financial Reforn" (1992a) and "Energy Efficiency and Conservation in the Developing World: The World Bank's Role" (1992b). 15 * Second, repeat the calculation with the shadow through innovations induced by the investment, prices equal to zero to estimate a conventional through learning-by-doing, or through a combina- rate of return. If the rate of return is satisfactory, tion of the two. For new industries, such benefits can it is clearly a Type I project, and GEF finance be substantial-as they were in the electricity indus- should be provided only if the leverage effect is try over much of the present century, when costs fell indeed exceptionally high, and if supporting the twenty-fold between 1900 and 1960 and the thermal effort can help the GEF and its users in develop- efficiencies of power plants rose ten-fold.23 The ben- ing Type II projects in general. efit often does not appear in cost-benefit analyses because the effect is fairly small and the investments Innovations and cost reductions being appraised generally use well-developed tech- A GEF requirement will be to support the develop- nologies for which the scales of production are al- ment of promising emissions reduction technolo- ready large. But the situation is different for Type II gies and practices that have not yet achieved their GEF projects, where production levels and markets full potential, and whose costs may be expected to are still small and cost curves are declining steeply. decline as applications and markets expand."8 This is already happening in several areas. In the case of Thus the actual costs of a GEF investment have the photovoltaics, for instance, unit costs of modules three following elements: have declined fifty-fold since 1970, to around $6,000 per kilowatt peak (KWp), and are projected to fall * The incremental capital costs again to the $1,000 to $1,500/KWp range as mar- * Plus the present value of incremental operating kets expand, and as improvements in materials, and maintenance costs conversion efficiencies, and manufacturing tech- * Less the present value of reductions in incremental nologies occur.19 Including balance-of-system capital costs in later years, per unit of investment in costs-such as structures, dc/ac invertors, control- year zero, times the levels of investment in later lers and installation-total costs are about $10,000/ years.24 KWp at present, but are projected to decline to $1,000/KWp over the long term.20 The scope for While the contribution of a current investment to cost reductions in other renewable energy forms, reductions in the unit costs of later investments may and in the efficiency of end-use and energy conver- be small, the overall benefits of investing in and sion devices is also considerable.2" More generally, developing a technology can be substantial if the the possibilities for technical improvements and prospective use of the technology is large. It is the cost reductions in all Type II projects that qualify product of two effects that is important-the contri- for GEF finance are far from being exhausted. bution to cost declines and prospective use. The contribution of an investment to reductions in The cost of photovoltaics the costs of future investments can be regarded as an A study of the Japanese photovoltaic industry by Dr. economic benefit.22 The cost savings may come Hamki Tsuchiya enables us to estimate the benefits Based on the notes to STAP by Professor Robert Williams. See, for example, the U.S. Department of Energy Interlaboratory White Paper (1990). 20 Ibid, and Johansson et al. (1992). 21 See Johansson et al. (1989). 22 See Arrow (1962). 23 See U.S. Department of Energy (1983). 24 More formally, let C = the present value of the costs of investments over a long time interval, K, (t = 0, I ...) be the unit capital costs in period t, I, be the investments in period t (in KW units in the case of electricity), and a, be the discount factor, a, = (1 + r) where r = the discount rate. Then CO = present value (PV) of 0 & M costs + _a,K,II But K, is a function (to) of the cumulative amount of investment, or K, = o(I, I Hence aC,/aI. = K,, + Ya(&K/aI )I (the summation is now over t > 0), neglecting changes in operating and maintenance costs, and the total present value of costs of an investment in period o are approximately (aCjal,,l,, - K,,I,, + PV of 0 & M costs + Xa,I,(dK,AI)I, where the third term will be negative under the conditions described, i.e., of cost declining with I. 16 bled, unit costs have declined by roughly 20 per- Table 6. Historical photovoltaics cost data cent. In an independent study using world sales for Japan data, Cody and Tiedje (1992) have obtained an Module cost Production Accumulated almost identical result. Year (Yen!Wp) (KW/year) production (KW) 1979 7,000 85.8 85.8 Table 7 below shows how the cost-saving benefits 1980 4,000 291 376.8 can be estimated given the relevant cost curves for 1981 3,500 1,024 2,424.8 the various renewable energy technologies. It is 1982 2,200 2,123 4,547.8 based on Dr. Tsuchiya's data forphotovoltaics with 1983 1,800 4,826 9,373.8 the following adjustments and assumptions: 1984 1,500 6,918 16,291.8 1985 1,200 10,800 27,091.8 * A present price of photovoltaics of $10,000/ 1986 1,100 13,400 40,491.8 KWp, with $1,000/KWp being the minimum 1988 900 13,000 65,941.8 expected volume of costs 1988 00 13000 6,98 * A value of b = -0.3 Source: Hamki Tsuchiya (1992). * Annual rates of investment and levels of accu- mulated production shown in the first two rows of the table. that occur when a current investment reduces future costs. His cost and production data are noted here A lower limit of $1 ,000/KWp is also assumed in the for convenience, and are not dissimilar to those outer years. As with the calculation of the shadow reported in U.S. Department of Energy and Europe- price of carbon emissions, the estimated benefits an studies. shown in the bottom rows of the table are only applicable in a scenario of global warming, assum- Dr. Tsuchiya estimates the following function: ing the investment levels shown in the first row. K = aX' On these assumptions, therefore, it would be more appropriate to appraise current investment in photo- where K = unit production cost (Yen/Wp), and X is voltaics on a net cost of $10,000-$3,800 - $6,000/ the accumulated level of production. His estimate KWp, rather than on a unit cost figure that ignores of the parameter b is -0.3, which means that each the contribution of the investments to future devel- time the accumulated level of production has dou- opment and cost reduction. Table 7. Estimation of cost-saving benefits of GEF investments in photovoltaics Year 0 10 20 30 40 50 Investment in Year, I, GW 0.075 1.0 5 25 75 250 Accumulated production: X1, GW 0.3 4.9 32 168 640 2,160 Production costs, K1, $/KWp 10,000 4,300 2,450 1,490 1,000 1,000 Contribution to cost reductions:a B, = bK I/X,, $/KWp 750 260 115 65 35 0 Present value of B, at r = 10% 3,800/KWp Note: Figures are rounded. The fornula follows from Table 6. 17 Extensions choices. This is especially true if GEF funding Further analysis of the costs and scale of production reduces such costs by facilitating market aggre- will enable estimates to be made forother renewable gation and organizational learning. Transaction energy technologies. Another factor to consider costs, including the difficulties involved in de- is the scale of manufacturing plant rather than veloping new approaches, can be high for a small accumulated annual production. If, for example, operation, but relatively small when the scale of annual production figures are used, the value of b an operation increases. is (numerically) higher, say about -0.4, and the * Pricing policies have worked against new tech- present value of cost savings would be higher than nologies by subsidizing fossil and hydro power. the figure just estimated. Such studies have been Although good pricing policies will be crucial proposed for the GEF and are now being initiated. for the development of new technologies, they need to be complemented by investments that Demonstration projects deal with institutional issues andpromote knowl- Significant opportunities for cost reductions also edge about new options. exist through operations that help to demonstrate * Once the alternatives are demonstrated in devel- the economic, technological, and administrative oping countries, it may be possible to reduce cost feasibility of renewable energy and energy efficien- by substituting labor for capital in the manufac- cy services.25 Examples include: ture and use of the alternative-developing coun- tries may have a comparative advantage in the - The establishment of industries that market ener- manufacture and use of renewable energy tech- gy services in developing countries nologies. * The development of niche markets for renewable energy-such as photovoltaics in rural areas For these reasons, the leverage effects of demon- - Some medium- and large-scale demonstration stration projects may be large, which means that projects for power generation to win confidence Type I projects would again qualify for finance (on in the new technologies and to provide an oppor- an exceptional basis) under this heading. tunity for the electric utilities (in particular) to familiarize themselves with the operational char- Continuity, replicability and aggregation acteristics of such options as wind power, and the in GEF projects use of photovoltaics for supplementary power on Achieving cost reductions in the manufacture of grid-fed distribution networks. renewable energy projects will require a sustained commitment from GEF and other sources over ten Each of these examples could involve projects with or twenty years.27 Significant cost reductions in good economic returns, but the traditional alterna- manufacturing will not be achieved by a few invest- tives (hydro and fossil-fired power stations) have ments over a short period, but only through a pro- been preferred for the following reasons: gram of investments in the more promising Type II projects over several years-particularly in photo- * There is a natural predisposition among the local voltaics, solar-thermal, and biomass projects. The utilities and the financing agencies to stay with situation is somewhat different for projects that better known approaches-even though these overlap the Type I and Type II categories. These approaches (for example, fossil fuel stations) projects often have high transaction costs and are have often proved unreliable.26 likely to become commercial sooner than Type II * Because the transaction costs of developing in- projects. But with respect to (Type II) renewable novative energy technologies can be high, GEF energy projects, several companies emphasize the support can be crucial in making cost-effective importance of continuity in operations, involving 25 This section also draws in part on notes prepared bv Professor Robert Williams for the STAP Committee. 26 The recent World Bank Policy Paper on Electric Power found availabilities to be as low as 40 percent in some countries and to average about 60 percent, as compared with 80 to 90 percent in good practice situations. 27 See Professor Robert Williams' note to STAP (September 24, 1992). 18 repeat projects. Any move to larger scale produc- * Bythefollow-upofsuccessfulpilotprojectswith tion manufacturing plants to achieve economies of larger scale operations scale and to introduce new, higher-volume-lower- * By aggregating or bundling a large number of cost manufacturing technologies, can only be justi- small-scale projects into larger operations fied by private companies if there is market growth. * By complementing Type I with Type II projects where the same technologies can be used in both There are several ways in which the GEF can commercial and developmental situations facilitate market growth: * By giving priority to countries whose pricing and institutional policies better promote renew- Through a commitment to continuity in its in- able energy and energy efficiency, for example, vestment operations, focusing on replicable by not subsidizing the use of fossil and hydro projects that are candidates for additional fi- energy. nance in other countries or in the same countries in the following years 28 See the World Bank's two policy papers on this subject (1992a, b). 19 3 Conclusions For GEF projects, the cost (or shadow price) of A review of innovations and costs in the backstop carbon emissions is the marginal cost of turning to technologies reveals that the costs of solar and the backstop technologies. The most promising biomass energy have declined remarkably over the technologies at present are those using renewable past two decades, to the point where they are com- energy-primarily solar and biomass-the costs of petitive in niche markets. (These markets are also which have declined appreciably over the past two growing.) The GEF can help reduce costs further by decades. In a scenario of global warming, they expanding applications. The contribution of any would be turned to increasingly to stabilize carbon single investment to innovation and the reduction of accumulations at a safe level (not yet determined). unit costs is small. The potential for reducing total The use of backstop technologies is consistent with costs over the long term, however, is significant stabilizing accumulations at any one of several when the declines of unit cost are multiplied by the levels over the long term, unless a runaway feed- prospective use of the technologies. An example is back occurs. investments in photovoltaics-their current costs (including balance of systems costs) are approxi- Costs and innovations mately $10,000/KWp. When projected for large- Estimates of marginal cost can be based on the scale use, such investments could help reduce formula for the optimal price of a depletable re- future costs by approximateiy $4,000/KWp. An source, in which the price of a resource (in this case allowance for such effects will make a difference in fossil fuels) equals the extraction cost plus a user project appraisal. cost that rises over time until the backstop or re- placement technology (in this case renewable ener- Type I and Type 11 projects gy) becomes economical; in the present case the The GEF will need to give priority to Type II resource in question is not the availability of fossil projects; such projects can only be readily justified fuels, but the limits to their use set by concerns once global concerns are taken into account, for about climate change. Preliminary estimates of example, by placing a shadow price on carbon marginal costs made in this paper amounted to $25 emissions of fossil-fuel alternatives when costs are per ton of carbon, rising at a rate equal to the being compared, and by considering the project's discountrateupto$120perton. Theseestimatesare potential for cutting future costs. Type I projects, not exact, but they provide a useful starting point such as those related to energy efficiency, are prof- for analysis, and would need to be revised periodi- itable and generate positive economic rates of re- cally in light of technical developments and new turn, even if global environmental concerns are evidence on the greenhouse effect. ignored, and they are supported by ODA. While 20 some of these projects can help reduce carbon * Giving priority to countries whose institutional emissions, they cannot solve the global wanning arrangements and energy pricing policies are problem. well suited to the success of GEF projects * Emphasizing replicability. Extensive use of GEF finance for Type I projects will dilute the Facility's effect and risk leaving its Continuity of GEF operations is needed to achieve mission unfulfilled. The development of Type II long-term cost reductions in Type II projects, and to projects will be crucial if the international commu- reduce transaction costs. Numerous companies, to- nity, especially developing countries, are to re- gether with government departments and research spond effectively to the global warming problem. institutions, have made the point that costs can only GEF funding might be used in exceptional cases be expected to decline if there is a long-term com- when the development of Type I projects (for exam- mitment to develop the markets and the applications ple, applications of renewables) will catalyze the of promising technologies. Future GEF operations development of Type II projects; the GEF may then will therefore need to support the principle of rep- help to reduce transaction costs and expand com- licability by financing follow-up projects. mercial uses. When used for such cases, the lever- age effect of resources would be expected to be GEF's mandate exceptionally high. The appraisal methods and criteria proposed above for GEF projects are based on conditions that would Transaction costs occur in a scenario of global warming, even though Transaction costs are not an insurmountable obsta- the seriousness of the global warming problem is cle to high-potential demonstration projects. The not known. It has been remarked that "the unequiv- leverage effect of GEF projects is thought to be ocal detection of the enhanced greenhouse effect high for many Type II projects. Examples include from observations is not likely for a decade or wind energy, biomass for power generation in more."29 But the approach is consistent with the some regions, and thermal-solar and photovoltaics mandate of the GEF, which is to help in the imple- for supplementary power or grid-fed distribution mentation of precautionary policies and leave the systems. Such projects have potential economic international community better placed to deal with viability according to standard profit-and-loss cri- a particular contingency; this means that its invest- teria. The key obstacles to the use of innovative ments, and the ground rules on which they are technologies relate in such instances to transaction based, will need to focus on that contingency. As it costs and to uncertainties encountered in introduc- happens, the costs of the technologies and practices ing them, not to their economic promise. Demon- the GEF is supporting, notably in renewable energy, stration projects are considered to be important in are declining significantly. Global concerns aside, this respect, and would be satisfactorily screened several such technologies have good economic po- by the cost and innovation criteria suggested above. tential. The GEF is extending early support to Transaction costs can be reduced by: developments that may have long-term economic merit. If so, this will not be the first example of * Aggregating or bundling many small-scale policies initially introduced to address an environ- projects into a larger program suited for project mental problem holding the seeds of an economic finance surprise. * Combining Type I and Type II projects where they are complementary 2'1 John Houghton, Chairman of the Scientific Assessment Working Group of the Intergovernmental Panel on Climate Change (1992). 21 References Ahmed, Kulsum. "Renewable Energy Technologies: A Review of the Status and Costs of Selected Technologies." World Bank Technical Paper No. 240. Washington, D.C.: World Bank 1993. Anderson, Dennis. "Energy and the Environment: An Economic Perspective on Recent Technical Develop- ments and Costs." Special Briefing Paper No. 1. Edinburgh: Wealth of Nations Foundation, 1991. Anderson, Dennis and William Cavendish. "Efficiency and Substitution in Pollution Abatement: Three Case Studies." World Bank Discussion Paper No. 186. Washington, D.C.: World Bank, 1992. Anderson, Dennis and Catherine D. Bird. "Carbon Accumulations and Technical Progress-A Simulation Study of Costs." Oxford Bulletin of Economics and Statistics, 54 (1): 1-29, 1992. Arrow, Kenneth. "The Economic Implications of Learning by Doing." Review of Economic Studies, V'ol. 29, 1962. Bates, Robin W. and Edwin A. Moore. "Commercial Energy Efficiency and the Environment." Background Paper for World Development Report 1992. Washington, D.C.: World Bank, 1992. Barbier, Edward and David W. Pearce. "On Thinking Economically About Climate Change." Energy Policy, January 1990. Bolin, Bert, B.R. Doos, J. Jager and R.A. Warrick, eds. The Greenhouse Effect, Climatic Change and Ecosystems. SCOPE 29 Report. New York: John Wiley and Sons, 1986. Booth, Roger and Philip Elliott. "Sustainable Biomass Energy." Shell Staff Technical Paper. London: Shell Centre, 1990. Cline, William. "Estimating the Benefits of Greenhouse Warmning Abatement." Washington, D.C.: Institute for International Economics, 1991. Cody, G. 0. andT. Tiedje. "The Potential for Utility Scale Photovoltaic Technology in the Developed World: 1990-2010." In Energy and the Environment. Edited by B. Abeles, A. Jacobson, and Ping Sheng, 1992. Goudie, Andrew. Environmental Change. 2nd ed. Oxford: Clarendon Press, 1983. Global Environment Facility. "Criteria for Eligibility and Priorities for Selection of GEF Projects." Report of the Scientific and Technical Advisory Panel of the GEF, Washington, D.C., May 1992. Hall, D.O. and F. Rossillo-Calle. "CO2 Cycling by Biomass: Global Bioproductivity and Problems of Degeneration and Afforestation." In Balances in the Atmosphere and the Energy Problem. Edited by E.W.A Lingeman. Proceedings of the 59th Hereseaus Seminar. Geneva, 1990. Hirst, Eric and Charles Goldman. "Creating the Future: Integrated Resource Planning for Electric U rllities." Annual Review of Energy and the Environment, 16: 91-121, 1991. Houghton, John. "Climate Change: The Current State of Scientific Knowledge." In Volume!: Special Session Papers on Photovoltaic Technology. Edited by A.A.M. Sayigh. Second World Renewable Energy Congress, Reading, United Kingdom, 1992. 22 Johansson, Thomas B., B. Bodlund and R. H. Williams, eds. Electricity: Efficient End-Use and New Generation Technologies, and their Planning Implications. Lund, Sweden: Lund University Press, 1989. Johansson, Thomas B., Henry Kelly, Amulya K.N. Reddy and Robert H. Williams, eds. Renewables for Fuels and Electricity. Washington, D.C.: Island Press, 1992. King, Ken. Environmental Considerations in Energy Development. Manila: Asian Development Bank, 1991. Liebenstein, Harvey. "Allocative Efficiency versus 'X-Efficiency'." American Review of Economics, 56 (June) : 392-415, 1966. Markandya, Anil and David W. Pearce. "Environmental Considerations and the Choice of the Discount Rate in Developing Countries." World Bank Environment Department Working Paper No. 3. Washington, D.C., 1988. "Development, the Environment and the Social Discount Rate." Research Observer, 6 (2): 137- 52. Washington, D.C.: World Bank, 1991. Nordhaus, William. "To Slow or Not to Slow: The Economics of the Greenhouse Effect." Economic Journal, 101 (July) : 920-37, 1991. Organization for Economic Cooperation and Development. Environmental Effects of Automotive Trans- port: The OECD Compass Project. Paris: OECD, 1986. Energy and Cleaner Air: Costs of Reducing Emissions. Summary and Analysis of Symposium Enclair 86. Paris: OECD, 1987. Projected Costs of Generating Electricity from Power Stations for Commissioning in the Period 1995-2000. Paris: OECD, 1989. Shah, Anwar and Bjorn Larsen. "World Energy Subsidies and Global Carbon Emissions." Background Paper for World Development Report 1992. Washington, D.C.: World Bank, 1992. Tsuchiya, Hamki. "Photovoltaics Cost Analysis Based on the Learning Curve." Tokyo: Research Institute for Systems Technology, 1992. United States Department of Energy. "The Future Role of Electric Power in America." Washington, D.C., 1983. "The Potential of Renewable Energy: An Interlaboratory White Paper." Washington, D.C., 1990. Walsh, Michael P. Motor Vehicle Pollution-A Global Perspective. Report SP-718. Warrendale, Pennsyl- vania: Society of Automotive Engineers, Inc., 1987. "Motor Vehicles and the Environment: A Research Agenda." International Conference on the Motor Industry and the Environment. Geneva, November 1990. 23 Williams, Robert H. "Analytical Framework for Scientific and Technical Advisory Panel Criteria and Priorities on Global Warming." Notes prepared for the Scientific and Technical Advisory Board Ad-Hoc Working Group. Washington, D.C.: Global Environment Facility, September 1992. "Roles for Institutional and Technological Demonstrations in the GEF Portfolio." Notes prepared for the Scientific and Technical Advisory Board Ad-Hoc Working Group. Washington, D.C.: Global Environment Facility, September 1992. "Low-Cost Strategies forCoping with CO2Emission Limits-A Critique of 'CO2Emission Limits: An Economic Cost Analysis for the USA' by Alan Manne and Richard Richels." The Energy Journal, 11 (4): 35-39, 1990. World Bank. World Development Report 1992: Development and the Environment. New York: Oxford University Press, 1992. "The Bank's Role in the Electric Power Sector: Policies for Effective Institutional, Regulatory and Financial Reform." World Bank Industry and Energy Department Report No. R92-133/2. Washington, D.C., 1992a. "Energy Efficiency and Conservation in the Developing World: the World Bank's Role." World Bank Industry and Energy Department. Washington, D.C., 1992b. World Energy Council. "Renewable Energy Resources: Opportunities and Constraints 1990-2020." Fifteenth Congress, Working Group Session No. 3.1. Madrid, September 20-25, 1992. 24 Appendix The Cost of Backstop Technologies The seven tables in this Appendix present the de- Absent a greenhouse gas constraint, the cost of coal- tailed calculations of Professor Robert Williams, based electricity will not rise much in the future. whose explanatory notes are also attached. They Ongoing technological advances (especially inte- were used as the basis for the calculations of the grated gasification/combined cycle technology) will shadow price of carbon emissions provided in the make it possible to provide coal electricity with text, and give a fuller description of the various extraordinarily low levels of local pollution at costs technological options and their costs. that are not likely to be higher than at present. Notes to tables Al and A2 Table A2 In these tables "long term" refers to the period 2000- Only a 12 percent discount rate is used in calculat- 2010. The crude oil price in the long term is assumed ing the production costs for synthetic fuels, since it to be $25 per barrel. is assumed that regulated utilities would not be involved-even for the electrolytic production of For the renewable options, a range of costs is pre- hydrogen. sented to indicate the uncertainties for the variables that seem the most uncertain. Gasoline is the focus because it is the highest quality petroleum-based liquid fuel, and the synthetic fuel Table Al alternatives are of comparable quality. Electricity generation costs are calculated for both 6 percent and 12 percent discount rates, to highlight There are several reasons for adding methanol and the sensitivity of the costs to the discount rate. Most hydrogen from biomass in addition to ethanol. The renewable technologies are more capital-intensive production of ethanol via enzymatic hydrolysis from than fossil fuel technologies, so that their costs are lignocellulosic feedstock is the focus of biomass- more sensitive to the discount rate used. based liquid fuel research and development at present, and its cost outlook is promising. However, it will be Also, while a low discount rate is appropriate for difficult to use ethanol in a fuel cell vehicle-which, utility investments in industrialized countries, a high because of urban air quality concerns, may well prove discount rate is more appropriate for developing to be the technology of choice for road transportation countries and for investments by independent power in the decades ahead. Both methanol and hydrogen producers in industrialized countries. can be used in fuel cell vehicles. 25 Table Al. Cost of electricity generationa (US cents per KWh, 1990 prices (2000-2010)) Current Prospectivellong-term Discount rate 6% 12% 6% 12% Coalb 4.1 5.6 3.9 4.8 Residual fuel oilc 6.0 7.4 6.7 8.2 Natural gas (combined cycle)d 3.2 3.9 4.8 5.6 Nucleare 3.8 5.9 4.3 6.2 Wind' 3.7-5.8 5.4-8.8 2.3-3.5 3.3-5.2 Solar thermalg 11.7-16.6 18.3-25.9 4.5-8.1 7.0-12.7 Photovoltaich 33 55 Moderate insolationi Thin-filmi - - 4.6-5.5 7.6-9.0 High insolation' Thin-filmi 3.5-4.1 5.7-6.7 Concentrating' 3.6-4.5 5.9-7.3 Biomassnm 7.4-8.9 9.1-10.7 3.7-4.6 4.5-5.3 a In all cases, plant life is taken to be 30 years and all corporate income and property taxes are neglected. The annual insurance charge is assumed to be 0.5% of the installed annual capital cost. Thus the annual capital charge rate is 0.0777 at a 6% discount rate and 0.129 at a 12% discount rate. b The present price of coal for utilities is taken to be the average 1990 price in the U.S. ($1.4/Giga Joule (GJ)) and the long- term price is assumed to be $2.0/GJ. Coal plants are assumed to be base-load plants operated at 75% capacity factor. Present technology is assumed to be pulverized coal steam-electric plants with flue gas desulfurization at an average efficiency of 33.9%. Future technology is assumed to be second-generation integrated gasification/gas turbine-based technology at an average effi- ciency of 42.1 %. For details see R.H. Williams and E. D. Larson, "Advanced Gasification-based Biomass Power Generation," in Renewable Energy: Sources for Fuels and Electricity, T. Johansson, F. Kelly, A.K. Reddy, and R.H. Williams, eds., Washington, D.C.: Island Press, 1992. c The present price of residual fuel oil for utilities is taken to be the average 1990 price in the U.S. ($3.1/GJ) and the long-term price is assumed to be $4.2/GJ, corresponding to a crude oil price of $25/barrel. Residual fuel oil plants are assumed to be load- following plants operated at 50% capacity factor. Both present and future technology are assumed to be steam-electric plants at an average efficiency of 35.4%. For details see Electric Power Research Institute, "Technical Assessment Guide, Volume 1: Electricity Supply-1986," EPRI P-4463-SR, December 1986. d The present price of natural gas for utilities is taken to be the average 1990 price in the U.S. ($2.2/GJ) and the long-term price is assumed to be double this ($4.4/GJ), consistent with the long-term crude oil price of $25/barrel. It is assumed that natural gas is burned in combined cycle plants that are load-following plants operated at 50% capacity factor. Efficiencies are assumed to be 42.7% at present and 50% for long-term future plants. For details see H. Kelly and C. Weinberg, "Utility Strategies for Using Renewables," in Renewable Energy: Sources for Fuels and Electricity. e The present cost of nuclear power is that estimated by the Electric Power Research Institute for a 1,100 MWe light water reactor. Under the assumptions that the nuclear industry could be revived, the unit capital cost in the U.S. could be cut nearly in half, and plants could be built in six years. For details see Electric Power Research Institute, "Technical Assessment Guide, Volume 1: Electricity Supply-1986," EPRI P-4463-SR, December 1986. The future cost is that estimated by the Electric Power Research Institute for a 600 MWe passively safe light water reactor, assuming a five-year construction time. For details, see Electric Power Research Institute, "Technical Assessment Guide, Volume 1: Electricity Supply-1989," EPRI P-6587-L, Sep- tember 1989. f Present wind power technology is taken to be the recently introduced variable speed turbine having an installed capital cost of $1,000 per KWe, with operating and maintenance and land rent costs of 1I.1 and 0.3 cents/KWh respectively. The range of costs reflects alternative wind regimes: the low end of the cost range corresponds to a hub height wind power density of 630 Watts/m2, where the capacity factor (CF) = 0.36, and the high end is for a power density of 350 Watts/m2, with CF = 0.20. For the long term, it is estimated that the installed capital cost could be reduced to $750/KWe and the operating and maintenance cost reduced to 0.6 cents/KWh. For the long term, the low end of the cost range corresponds to a hub height wind power density of 630 Watts/M2, where CF = 0.49, and the high end is for a power density of 350 Watts/M2, with CF = 0.27. For details, see A. Cavallo, S. Hock, and D. Smith, "Wind-Energy: Technology and Economics," in Renewable Energy: Sources for Fuels and Electricity. 26 Table Al. (continued) g Present solar thermal electric technology is an 80 MWe Luz parabolic trough system with an installed capital cost in the range of $2,800IKWe (with CF = 0.25 and an operating and maintenance cost of 1.8 cents/KWh) to $3,5001KWe (with CF = 0.22 and an operating and maintenance cost of 2.5 cents/KWh); no credit is taken here for use of natural gas as a backup. Future technology is a 200 MWe advanced central receiver system with high temperature thermal storage, with an installed capital cost in the range $1,800/KW (with CF = 0.43 and an operating and maintenance cost of 0.8 cents/KWh) to $2,500/KW, (with CF 0.32 and an operating and maintenance cost of 1.2 cents/KWh). For details see P. de Laquil, D. Kearney, M. Geyer and R. Diver, "Solar Thermal Electric Technology," in Renewable Energy: Sources for Fuels and Electricity. h In 1990 the average price of photovoltaic modules was $6.2/Wp, the total installed cost of the best photovoltaic systems was about $10/Wp, and the operating and maintenance cost was about $0.005/KWh. i Moderate insolation is defined as 1,800 KWh/m/year, the average for the U.S. j It is assumed that in the long term, thin-film modules achieve efficiencies of 15%, module costs are $45 to $50/m2, area- related balance-of-system costs are $37 to $50/m2, power-conditioning costs are $100/KW, and indirect costs are 25% of direct costs. The operating and maintenance cost is projected to be $0.3/M2 per year. For details, see "Introduction to Photovoltaic Technology," in Renewable Energy: Sources for Fuels and Electricity. k High insolation is defined as 2,400 KWh/m2/year, typical of sunny areas in the U.S. southwest. I It is assumed that concentrator photovoltaic systems are deployed in areas of good direct normal insolation (2,400 KWh/m2/ year). The low end of the projected costs for the long term is for ID tracking systems, for which it is estimated that modules can achieve efficiencies of 20%, module costs are $60/m2, area-related balance-of-system costs are $50/m2, power-conditioning costs are $ 100/KW, and indirect costs are 25% of direct costs. The high end of the projected costs for the long term is for 2D tracking systems, for which it is estimated that modules can achieve efficiencies of 35%, module costs are $150/ni2, area-related balance- of-system costs are $ 100/M2, power-conditioning costs are $100/KW,, and indirect costs are 25% of direct costs. In both cases the operating and maintenance cost is estimated to be 0.25 cents/KWh. For details, see H. Kelly "Introduction to Photovoltaic Tech- nology," in Renewable Energy: Sources for Fuels and Electricity. m The price of biomass for utilities is assumed to be in the range of $2.5 to $3.5 per GJ, which is 1.4 to 2.5 times the present utility coal price in the U.S. Biomass plants are assumed to be base-load plants operated at 75% capacity factor. Present technology is assumed to be a 23.4% efficient 28 MWe steam-electric plant. Future technology is assumed to be a biomass integrated gasifier/ intercooled steam-injected gas turbine with an average efficiency of 42.9%. For details see R. H. Williams and E. D. Larson. "Advanced Gasification-based Biomass Power Generation," in Renewable Energy: Sources for Fuels and Electricity. 27 Table A2. Cost (ex-refinery or fuel processing plant) of gasoline and zero net C02-emitting alternative automobile fuelsa ($!barrel of gasoline-equivalent) Current Prospectivellong-term Gasoline (ex-refinery) 33b 45c Ethanol from lignocellulosic feedstocksde 90-103 44-52 Methanol from lignocellulosic feedstocksdf' 80-89 55-64 Hydrogen from lignocellulosic feedstocksc,f 63-70 44-51 Hydrogen from wind powerg 140-210 90-140 Hydrogen from photovoltaic powerh 960 144-166 a The unit energy measure is the energy content (gross or higher heating value basis) of a barrel of gasoline, which is 5.54 GJ. The costs of all alternative fuels were calculated for a 12% discount rate. For all biomass fuels the plant life is assumed to be 25 years; for wind and photovoltaic systems it is assumed to be 30 years. In these cost calculations all corporate income and prcperty taxes are neglected. The annual insurance charge is assumed to be 0.5% of the installed capital cost per year. The "long term" is defined as the period 2000-2010. b Ordinary gasoline from crude oil @ $20 per barrel. c Reformulated gasoline from crude oil @ $25 per barrel. d Biomass feedstock delivered to the conversion facility is assumed to cost in the range of $2.5 to $3.5 per GJ. e It is assumed that ethanol is produced from lignocellulosic feedstocks using enzymatic hydrolysis. Present costs and future cost estimates were made by researchers at the U.S. National Renewable Energy Laboratory. For details see C. Wyman, R. Bain, N. Hinman and D. Stevens, "Ethanol and Methanol from Cellulosic Biomass," in Renewable Energy: Sources for Fuels and Electricity. f It is assumed that methanol and hydrogen are produced from lignocellulosic feedstocks at present using a Shell oxygen- blown gasifier adapted to biomass; the Shell gasifier is commercially available technology for coal gasification. The future technology is assumed to be an indirectly heated biomass gasifier (specifically, the Battelle Columbus Laboratory gasifier) de- signed to exploit the high reactivity of biomass (compared to coal); with this gasifier, a costly oxygen plant is not needed (but would be needed to gasify coal for the production of these fuels). For details, see E. Larson and R. Katofsky, "Production of Methanol and Hydrogen from Biomass," PU/CEES Report No. 271, July 1992. g It is assumed that hydrogen is produced via electrolysis from wind power (see table Al) costing at present 5.4 to 8.8 cents/ KWh (so that the produced hydrogen costs $25.8 to $38.0 per GJ) and in the future 3.3 to 5.2 cents/KWh (so that the produced hydrogen costs $16.2 to $25.0 per GJ). For details see J. Ogden and J. Nitsch, "Solar Hydrogen," in Renewable Energy: Sources for Fuels and Electricity. h It is assumed that hydrogen is produced via electrolysis from dc electricity (dc/ac power conditioning is not needed). With present technology, dc electricity from photovoltaic systems costs about 51 cents/KWh (see table Al), corresponding to a photovol- taic hydrogen cost of $173 per GJ. For the long term, it is assumed that photovoltaic hydrogen is produced from electricity generated in thin-film devices in areas of moderate insolation-1,800 KWh/m2 per year. As in the case of ac electricity production (see table Al), it is assumed that, in the long term, thin-film modules achieve efficiencies of 15%, module costs are $45 to $50/m2, area-related balance-of-system costs are $37 to $50/r2, indirect costs are 25% of direct costs, and operating and maintenance costs are 0.15 cents per KWh. Under these conditions the cost of dc electricity (assuming a 12% discount rate) ranges from 6.2 to 7.5 cents/KWh, corresponding to a range of hydrogen costs of $25.7 to $30 per GJ. For details see J. Ogden and J. Nitsch, "Solar Hydrogen," in Renewable Energy: Sourcesfor Fuels and Electricity. 28 Table A3. Cost of automotive fuels delivered to consumersa ($IGJ) NG->MeOH NG->H2 Biomass->MeOH Biomass->H2 Wind->H, PV->H2 Productiona,b Feedstockc 3.40 2.72 4.65 3.83 - - Capital 2.40 1.62 4.21 2.57 - - O&M 0.93 0.49 1.91 2.12 - - Subtotal 6.7 4.8 10.8 8.5 21.1d 27.8d Compression - - - - 1.4e 1.4' Overseas transport l.5 - - -- - Storage - - - 1.1h 1.g Local distribution - 0.5 - 0.5 0.5 0.5 Refuelingh 2.8 5.2 2.8 5.2 5.2 5.2 Retail Price 11.0 10.5 13.6 14.2 29.3 36.0 a Details of the production cost estimates for methanol (MeOH) and hydrogen (H ) derived from natural gas (NG) and bio- mass feedstocks are from E. Larson and R. Katofsky, "Production of Methanol and Hydrogen from Biomass," PU/CEES Report No. 271, July 1992. Details of the production cost estimates for electrolytic H2 are from J. Ogden and J. Nitsch, "Solar Hydro- gen," in Renewable Energy: Sources for Fuels and Electricity. b Lifecycle costs for MeOH or H2 production are calculated assuming a 12% discount rate. All corporate income and property taxes are neglected. The annual insurance charge is assumed to be 0.5% of the installed capital cost. c It is assumed that MeOH and H, are produced from NG feedstocks costing $2 per GJ, and from biomass feedstocks costing $3 per GJ (the mid-point of the range of biomass costs considered in table A2). d It is assumed that electrolytic H2 is produced from ac wind power or dc photovoltaic power at the average of the hydrogen production costs for these sources indicated in table A2. e The cost of compressing electrolytic H2 to 7.5 Mega Pascals (1,098 pounds per square inch area). Compression costs are included in the production costs for H2 from NG and biomass. f It is assumed that MeOH produced from NG (but not from biomass) comes from remote overseas sources. This cost is for transport from southeast Asia to the U.S. Gulf coast. See C. Wyman, R. Bain, N. Hinman and D. Stevens, "Ethanol and Methanol from Cellulosic Biomass," in Renewable Energy: Sources for Fuels and Electricity. g When H, is produced from intermittent sources, storage is needed near the production site so as to keep the pipelines full. h The cost for MeOH includes local distribution. The refueling station costs for H2 include the costs of compressing H2 to the high pressures needed for H, storage canisters on board the cars. 29 Table A4. Comparison of alternative automotive vehiclesa Vehicle type BPEVb FCEVC FCEVC FCEVO FCEVC FCEVC FCEVC ICEVd Primary US Natural Natural Biomass Biomass Wind Photovoltaic Petroleum energy source average gas gas Energy carrier Elect. MeOH 2 MeOH H2 H2 H2 Gasoline 1. Lifecycle cost (cents per km): Base vehicle 7.59 7.22 7.18 7.22 7.18 7.18 7.18 11.17 Fuel cell (inc. reformer) - 2.62 2.23 2.62 2.23 2.23 2.23 - Battery 7.08 2.03 2.15 2.03 2.15 2.15 2.15 -- Fuel storage - 0.02 0.81 0.02 0.81 0.81 0.81 Home rech. system 0.04 - - - - - - - Miscellaneous O&M 6.03 6.09 6.06 6.09 6.06 6.06 6.06 6.71 Fuel taxes 0.74 0.74 0.74 0.74 0.74 0.74 0.74 0.74 Fuel - 1.14 0.91 1.40 1.23 2.55 3.13 3.00 Electricity 1.48 0.22 0.20 0.22 0.20 0.20 - TOTAL 22.96 20.08 20.28 20.34 20.60 21.92 22.50 21.62 2. Gasoline-equivalentfuel economy: On fuel: miles per gallon - 62.4 74.0 62.4 74.0 74.0 74.0 25.9 liters/100 km - 3.77 3.18 3.77 3.18 3.18 3.18 9.08 On electricity: miles per gallon 120.0 131.5 131.5 131.5 131.5 131.5 131.5 - liters/100 km 1.96 1.79 1.79 1.79 1.79 1.79 1.79 - Average:: miles per gallon 120.0 71.5 84.2 71.5 84.2 84.2 84.2 25.9 liters/100 km 1.96 3.29 2.79 3.29 2.79 2.79 2.79 9.08 3. Delivered price of energy carrier: Fuel:' $/GJ - 11.0 10.5 13.6 14.2 29.3 36.0 9.5 $/gal gasoline- equivalent - 1.45 1.38 1.79 1.87 3.86 4.75 1.25 Electricity: $/GJ 19.4 19.4 19.4 19.4 19.4 19.4 19.4 $/gal gasoline- equivalent 2.56 2.56 2.56 2.56 2.56 2.56 2.56 - Average:. $/GJ 19.4 11.8 11.4 14.1 14.7 28.2 34.3 9.5 $/gal gasoline- equivalent 2.56 1.56 1.50 1.87 1.94 3.72 4.52 1.25 4. Breakeven gasoline price:9 $/gallon 1.81 0.61 0.69 0.72 0.83 1.38 1.62 1.25 $/liter 0.48 0.16 0.18 0.19 0.22 0.36 0.43 0.33 5. Lifecycle GHG emissions:` (gr Clkm) On fuel: - 34.2 28.6 6.2 9.2 4.5 4.5 86.0 On electricity: 48.2 43.9 43.9 43.9 43.9 43.9 43.9 - Average: 48.2 34.9 29.7 12.8 14.5 10.8 10.8 86.0 Index 56.0 40.6 34.5 14.9 16.9 12.5 12.5 100.0 3(0 Table A4. (continued) Vehicle type BPEVh FCEV FCEV' FCEV' FCEVC FCEV FCEvk ICEVd Primary US Natural Natural Biomass Biomass Wind Photovoltaic Petroleum energy source average gas gas Energy carrier Elect. MeOH H2 MeOH H2 H, H, Gasoline 6. Lifecycle cost of GHG emissions reduction: ($ItC) +354 -301 -220 -175 -143 +40 +117 - 7. Lifecycle GHG emissions per unit of energy consumed.i (kg C/GJ) On fuel: - 26.0 25.8 4.7 8.3 4. 1 4.1 27.2 On electricity: 70.5 70.5 70.5 70.5 70.5 70.5 70.5 - Average:e 70.5 30.4 30.5 11.2 14.9 11.1 11.1 27.2 8. Cost of GHG emissions reduction when onlyfuel cost is taken into account:k ($/tC) n.a. n.a. n.a. 288 423 1161 1540 - a Based in large part on a cost and performance model for automobiles developed in Mark A. DeLuchi, "Hydrogen Fuel Cell Vehicles," Institute of Transportation Studies, University of California, September 1992 (draft). b The battery-powered electric vehicle (BPEV) uses a bipolar Li/S battery, and has the performance and cost characteristics indicated in table A5. Electricity is assumed to cost $0.07/KWh ($19.4/GJ), the average U.S. residential price. c The fuel cell-powered electric vehicle (FCEV) uses a proton exchange membrane fuel cell for baseload power and a small bipolar Li/S battery for peak.ng power. It has the performance and cost characteristics indicated in table A5. It has a fuel economy of 74 miles per gallon (62.4 mpg gasoline equivalent) when operated on compressed hydrogen (methanol) and 131.5 mpg (3.59 miles per/KWh) when operated on external electricity. The compressed hydrogen is stored in aluminum canisters wrapped with carbon fiber at 8,000 psia. d The gasoline-powered internal combustion engine vehicle (ICEV) is a year-2000 version of the 1990 Ford Taurus. It has the performance and cost characteristics indicated in table A5. The gasoline price is assumed to be $1.25 per gallon ($9.5/GJ), withouit retail taxes-the expected retail price of reformulated gasoiine in the U.S. for crude oil @ $25 per barrel. e With methanol (compressed hydrogen) fueling, the fuel cell provides power for 0.787 (0.786) of each kilometer anid exter- nal electricity provides power for 0.181 (0.165) of each kilometer. Thus, regenerative braking provides power for 0.032 (0.049) of each kilometer. f See table A3. g The breakeven gasoline price is that gasoline price (excluding retaii taxes) at which the lifecycle cost of gasoline-powered ICEV equals the lifecycle cost of the alternative vehicle. h The lifecycle greenhouse gas (GHG) emissions, in gr C-equivalent per km, include both CO2 and other GHGs emitted throughout the entire fuel cycle, as well as the direct emissions from the vehicle. The biomass methanol and hydrogen options include the GHG emissions from the fossil fuels used to grow, harvest, and transport the biomass to the conversion facility. The hydrogen cases include the GHG emissions from the power plants that provide the electricity to compress these gases to high pressure at the refueling station, assuming the electricity needed to run the compressors is provided by the average mix of electric power sources in the year 2000. It is assumed that the electricity for recharging the batteries of the BPEVs and FCEVs is provided by the average mix of baseload power in the U.S. in the year 2000. i The lifecycle cost of GHG emissions reduction for option i is given by: [(lifecycle cost), - (lifecycle cost),CEv, ]/Rlifecycle GHG emissions)lCEV - (lifecycle GHG emissions),], where the lifecycle costs are given by line 1 in table A4, and the lifecycle GHG emissions are given by line 5. j Calculated as: (lifecycle GHG emissions in kg C/km)/(total energy consumed in GJ/km). k The cost of GHG emissions reduction for option i when only energy cost is taken into account is given by: [(average price of energy i - gasoline price, in $/GJ)] /[(lifecycle GHG/GJ gasoline) - (lifecycle GHG emissions/GJ energy i)], where the average prices are given by line 3, and the lifecycle GHG emissions are given by line 7. 31 Table A5. Characteristics of alternative vehicles BPEV MeOHIFCEV H/FCEV ICEV Weight (kg) 1,462 1,275 1,167 1,371 km driven annually 22,960 22,960 22,960 17,837 Range between refuelings (km) 400 400 400 640 Vehicle lifetime (years) 11.2 11.2 11.2 10.8 Selling pricea($) 28,247 24,810 25,446 17,302 Fuel economy: 1/100 km gasoline-equivalent 1.96 3.77 3.18 9.08 mpg gasoline equivalent 120 62.4 74.0 25.9 Maintenance cost ($/year) 388 450 434 516 Source: Mark A. DeLuchi, "Hydrogen Fuel Cell Vehicles," Institute of Transportation Studies, University of Califomia, Septem- ber 1992 (draft). The retail price breakdown for the BPEV and the hydrogen-powered FCEV is as follows: BPEV FCEV $ $ Traction battery and auxiliaries 13,625 4,205 Hydrogen fuel storage 0 2,692 Fuel cell stack and auxiliaries 0 4,496 Extra support structure for EV because of added weight 34 (14) Extra weight and drag-reduction measures for EV 107 107 Difference between EV and ICEV powertrain (2,839) (3,298) Net increment above ICEV cost 10,924 8,188 32 Table A6. Sensitivity analysis for carbon tax implied by backstop technologya Price of dc electricity delivered Lifecycle Carbon price hydrogen cost tax (centslKWh) ($IGJ) (centslkm) ($/tC) Low-cost FCEV,b high insolation,c + low pv cost parametersd 4.7 28.8 18.42 - 372 Low cost FCEVb + high insolationc 5.1 30.4 18.56 - 356 Low cost FCEVb 6.9 36.0 19.04 - 300 High insolationc + low pv cost parametersd 4.7 28.8 21.88 + 30 High insolationc 5.1 30.4 22.02 + 47 Base casee 6.9 36.0 22.50 + 102 High pv cost parameters' 7.5 38.2 22.70 + 126 High fuel cell costg 6.9 36.0 24.73 + 362 High fuel cell costg + high pv cost parameters' 7.5 38.2 24.92 + 384 a The backstop technology is assumed to be a photovoltaic hydrogen-powered FCEV displacing a gasoline-powered ICEV. In all cases it is assumed that electricity used for recharging the battery and for the compressor work required at the HI refueling station is provided at the same cost as for the cases in table A4, but with power sources having zero net CO2 emissions (appropriate for the long term). For this reason the base case carbon tax ($124/tC) is slightly lower than that shown in table A4 ($141/tC). For the calculations in table A4, it was assumed that electricity is provided with the average mix of power sources in the U.S. in the year 2000. b In this case the fuel cell, battery, hydrogen storage, and related equipment cost is $3,800 per car (as estimated by Ira Kuhn- see table A7) instead of the base case value of $11,400 (as estimated by Mark DeLuchi-see table A5). c High insolation = 2,400 KWh/m2/year (typical of the southwest U.S.). d Low photovoltaic cost parameters are module costs of $45/m2 and area-related balance-of-system costs of $37/l2. e In the base case, hydrogen is generated from dc photovoltaic electricity generated in an area of modest insolation (1,800 KWh/m2/year-the average for the U.S.) assuming mid-range values of PV module and balance-of-system costs. The FCEV has the performance and cost characteristics developed in Mark DeLuchi's model (see table A5). f High present value cost parameters are module costs of $50/m2 and area-related balance-of-system costs of $50/m2. g In this case the fuel cell costs $9,000 per car or $360 per KW-twice the cost estimated by DeLuchi (see table AS). 33 Table A7. Cost summary for cruise-design fuel cell electric vehicle by Ira F. Kuhna (year- 2002, 100,000 unitslyear) Basis Total cost ($i Hydrogen storage tankb 10.5 cubic feet (cf), 5,000 pounds 1,OOOC per square inch (psi) aluminum wrapped with carbon fiber (2 tanks @ 5.25 cf) Fuel cell system 37.5 KW gross; 30 KW net Fuel cell stack plates $8/lb 600 Solid polymer electrolyte membranes $5/ft2 260 Catalysts $2/KW 75 Gas management system $4/lb 180 Air compressor system $4/lb 100 Subtotal 1,215 Auxiliary power and controls Ultracapacitor 1.5 MJ 300 Step-up transformer $51/b 150 Auxiliary battery $2/lb 20 Power coniditioner for hotel load $10/lb 50 Subtotal 520 Motor-related Motor 75 shaft horse power (shp); 63 shp max. continuous 500 $5/lb Controller $5/input KW 320 Gearbox 10:1 step-down planetary 240 $4/lb Subtotal 1,060 TOTAL 3,795 The cost for the Ford Taurus ICEV parts displaced is $3,000-$4,000. Thus the net extra cost for the FCEV ranges from zero to $1,000. The removed parts weigh 935-980 lbs and the fuel cell system would weigh about the same. Thus the average cost of the fuel cell system would be about $4/lb. Note: Fuel cell capacity is sufficient for sustained 75 mph cruise on 0% grade or 55 mph on 3% grade, with 3 KW hotel load; 34 KW battery or 1.5 Mega Joule (MJ) ultracapacitor sufficient for acceleration and brake power regeneration; 75 hp peak power output. a Ira F. Kuhn, President, Directed Technologies. Inc.. 4001 N. Fairfax Drive, Suite 775, Arlington, VA 22203. Tel.:(703) 243-3383 Fax: (703) 243-2724. b Kuhn estimated that the alternative of a methanol storage tank plus reformer would cost $1,300, weigh 300 Ibs, and occupy 10 cubic feet, compared to S 1,000 for a compressed hydrogen storage system weighing 200 lbs and occupying 13 cubic feet. c Cost estimates provided by Structural Composite in Pomona. CA. Tel.: (714) 594-3939, a specialty tank manufacturing company. 34 Distributors of World Bank Publications ARGENTINA The Middle East Observer KENYA SOUTH AFR3CA, BOTSWANA Carlos Hirsch, SRL 41, Sherif Street Africa Book Service MEA) Lbd. For single t,Lle= Caleria Guemes Cairo Quaran House, Mfangano Steet Oxford University Pr1 Florida 165, 4th Floor-Oc. 453/465 P.O. Box 45245 Southern Africa 1333 BuenosAires FINLAND Nairobi P.O. Box 1141 Akateeminen Kirjakauppa Cape Town 8000 AUSTRALIA, PAPUA NEW GUINEA. P.O. Box 128 KOREA, REPUBLIC OF FIJI, SOLOMON ISLANDS, SF-00101 Helsinki 10 Pan Korea Book Corporatie, For suhscr#fian orden: VANUATU, AND WESTERN SAMOA P.O. Box 101, Kwangwhamun International Subscription Service DA. Information Services FRANCE Seoul P.O. Box 41095 648 Whitehorse Road World Bank Publications Craighall Mitcham 3132 66, avenue d'l6na MALAYSIA Johannesburg 224 Victoria 75116 Paris University of Malaya Cooperative Bookshop, Limited SPAIN AUSTRIA GERMANY P.O. Box 1127, Ialan Pantai Baru Mundi-Prenas Libras. S.A. Gerold and Co. UNO-Verlag 59700 Kuala Lumpur Casteilo 37 Graben 31 Poppelsdorfer Allee 55 25001 Madrid A-lOl Wien D-5300 Bonn 1 MEXICO INFOTEC Librerla Internacional AIDOS BANGLADESH HONG KONG, MACAO Aparesdo Postal 224-60 Consell de Cent 391 Micro Industries Development Asia 2000 Ltd. 14060 llaipan, Mexico DF. 08009 Barcelona Assistance Society (MIDAS) 464 Wyndham Street House 5, Road 16 Winning Centre NETHERLANDS SIU LANKA AND THEi MALDIVES Dhanmondi R/Area 2nd Floor De Lindeboom/InOr-Publikaties Lake House Bookshop Dhaka 1209 Central Hong Kong P.O. Box 202 P.O. Box 244 7480 AE Haaksbergen 100, Sir Chittampalm A. BEnck offia5 INDIA Gardiner MAwatha Pine View, 1st Floor Allied Publishers Private Ltd. NEWZEALAND Colombo 2 100 Agrabad Commercal Area 751 Mount Road EBSCO NZ Ltd. Chittagong 4100 Madras - 600 002 Private Mail Bag 99914 SWEDEN New Market For single til: 76, KD.A Avenue Brnh offic: Auckland Fritzes Fackbokstoretaget Kulna 9100 15 J.N. Heredia Marg Regeringsgatan 12. Box 16356 Ballard Estate NIGERIA 5-10327Stockholnm BELGIUM Bombay - 400 038 University Press Limited jean De Lannoy Three Crowns Building Jericho For suscrition orders: Av. du Roi 202 13/14 Asaf Ali Road Private Mail Bag 5095 Wennergren-Williams AB 1060 Brussels New Delhi- 110 002 lbadan P. O. Box 1305 S-171 25 Solna CANADA 17 (hittaranjan Avenue NORWAY Le Diffuseur Calcutta - 700 072 Narvesen Infonnation Center SWITZERLAND CP. 85, ISOIB rue Amrnpre Book Department Forsingle titiw Boucherville, Qu6bec Jayadeva Hostel Building P.O. Box 6125 Etterstad Wrairie Payot J4B SE6 Sth Main Road, Gandhinagar N-0602 Oslo 6 Case postale 3212! Bangaiore - 560 009 CH 1002 Lausanne CHILE PAKISTAN Invertec lT SA 3-5-1129 Kachiguda Mirza Book Agency For seuscrWition ordcn: Av. Santa Maria 6400 Cross Road 65, Shahrah-e.Quaid-e-Azam Librairie Payot Edificio INTEC Of. 20. Hyderabad -50 027 P.O. Box No. 729 Service des Abonnem;enrs Santiago Lahore 54000 Case postale 3312 Prarthana Flats. 2nd Floor CH 1002 Lausarae CHINA Near Thakore Baug, Navrangpura PERU China Financial & Economic Ahmedabad - 380 009 Editorial DesarroUlo SA THAILAND Publishing House Apartado 3824 Central DepartusentStcre 8, Da Fo Si Dong Jie Patiaia House Lima 1 306 Silom Road Beijing 16-A Ashok Marg Bangkok Lucknow - 226 001 PHILIPPINES COLOMBIA International Book Center TRINIDAD & TOBAGO, ANTIGUA Infoenlace Ltda. Central Bazaar Road Suite 1703. Cityland 10 BARBUDA, BARBADOS, Awartado Aereo 34270 60 Bajaj Nagar Condoainium Tower 1 DOMINICA. GRENADA, GUYANA, Bogta D.E. Nagpur 440010 Ayala Avenue, H.V. dela JAMAICA, MONTSERRAT, ST. Costa Extension KITTS & NEVIS, ST. LUCIA. COTE DIVOIRE INDONESIA Makati, Metro Manila ST. VINCENT & GRENADWNES Centie d'Edition et de Diffusion Pt Indira Limited Systematics Studies Unit Africaines (CEDA) Jalan Borobudur 20 POLAND #9 Watts Street 04 B.P. 541 P.O. Box 181 Internadonal Publishing Service Curepe Abidjan 04 Plateau Jakarta I;320 UlI. Piekna 31/37 Trinidad, West Lndies 00-677 Warzawa CYPRUS IRELAND TURKEY Center of Appeied Research Covernment Supplies Agenicy Forsucrtion onrds: Infotel Cyprus College 4-5 Harcourt Road IPS lournals Narlabah.e Sok. No. 15 6. Diogenes Street, Engomi Dublin 2 Ul. Okrezna 3 Cagaloglu P.O. Box 2006 02-916 Warszawa Lstanbul Nicosia ISRAEL Yozmot Literature Ltd. PORTUGAL UNITED IINGDOM DENMARIK P.O. Box 56055 Uvraria Portugl Microinfo LtdL SamfundsUltteratur Tel Aviv 61560 Rua Do Carmo 70.74 P.O. Box 3 Rosenoerns AII 11 1200 Lisbon Alton Hamnpshire GU342PG DK-1970 Frederiksberg C ITALY England Licosa Commissionaria Sansoni SPA SAUDI ARABIA, QATAR DOMINICAN REPUBLIC Via Duca Di Calabria. 1/1 Jarir BookStore VENEZUELA Editora Taller. C por A. Casellh Postale 552 P.O. Box 3196 Libreria del Este Restauracifn e Isabel la Cat6ica 309 5012S Firenze Riyadh 11471 Aptdo. 60337 Apartado de Correos 2190 Z-l Caracas 106C-A Santo Domingo JAPAN SINGAPORE TAIWAN. Eastern Book Service MYANMAR,BRUNEI ECYPT. ARAB REPUBLIC OF Hongo 3-Chome, Bunkyo.ku 113 Inform.ation Pubclcations Al Ahrm Tokyo Private. Ltd. Al Cala Sreet Golden Wheei Building Cairo 41, Kaliang Puddig, #04.C3 Singapore 1334 UNDP Field Offices* P.C. BO 3 C.siiii 9r of Grand Centr:l P.C. 3ox = 3 KAbUl AtZsc dsA n La &Pi.3o1kviid lh:3rio Sation laskzts 10310, New Yog. N.Y. 1zniti Desluncrzt 34 P.C. ox 54 A;artado Posalu 10163-1601 Shahiat S1Ztt C a b a r o a, 4540 P.. O o° Vilt No.35 Boss 32.1 low, Cctu P.C. 3 44 15V4557 Timtuo, AJbania a 3-0 Ta1110 T.horas, sAie CAx.i Pot 02W E*onia RBApblln of Iii Boltu postars So 7Q3!;7 7G9 Brasla, 0t BOita post. Ai.r Can, FP 1,47, P. O. 6530 P.O. lkX 204 Algisrs 16000, hmil Abidjrn 01, Addis Ababa, (Alwiyui: Alias Ropublquu do Cow Eh>9ia llhdad, Irq P.C. 3cx S CLLs POua 91O 1^s4 Socla Piivxte V1a4 Bag P. O. Box 210 Luanda, Peay1e' BuiSria UNDP Cuba Suva, Pi3 KhSn, Iozalita LRpublic of Angola I Unrt.d Nodm Baits postal 1490 PIz=a DOits petAi 2183 P.O. Box 51339 CaailUa d Com wo J j t m b a a , N.w Yotr, ?,Y. Librsville, CObon Jarumlao 91513 2257 DrnS 100 17, UJ A. IMsOS 1000 Capital P,O. R= 552 Podml bits Pos4b 172 P. 0. Bos 5605 ,uLThsauts, P.O . Dc 35224 Buenos Air, BangUi. C*dtTA1 Ntiosia, Cypts Anunad, 7orfdA At7oina Af;xa PIpublia P. 0. 30x 1423 P.O. Bar 2530 Ac;n, (ShlhA Haul XAZSkda United NtiIoCs P.O. Box 877 U K - % I 0 0 Rocn: 1312 UoLflsd lotarim Paor Cz,pr:snA A,csrado 2l L3 enicn Pr,psot ficw C-CrobxdA flnourk 'A' Alzy, X Hr2zzAn H4owl 01909 Guwtsu .md 11ooo B4ita potal. 46 B3it. pO, 42001 C'.ty, C.A P. O. It 320113 PionerslkaLSt. 2 Yaounde, SLpub D j i b o u t i , Ouatatl Nal Kenya Yerevan,A=la of Caoo Rdpubliu do Djibouti jBoil. potal. P. 0, Box 2995 UutiLa4S NNatiao P.C. 30x207 C;taky, Rspuwu Sfatt Ildrim Otis Doula Apsrudo 1424 of OUI.at 13030 QCwUI 3 itualllyst Strt CAnr0oon sAmto Doalnio, sutMo@f Kwatit Baiz 1, Azarijs DaUinianRoUbilUs C. P. 179 C.;- dal 62 P.C. Bo.x 1011 9 Toautm Srm P. 0. 3cx 25314 P-ta, RWaubUt de P.O. 2cx :7 bts&au Codsx Saond 1oor )42ZmA, 3&inli Caoo V§rd4 P y o a I * g, Rpublic oft RIga D et m o O r a t;i e GuirwRlu Lcis .?. 0. 3cr 224 Bacit pouaas 906 PICV!4'1 Dhaka ICO1 N'Ciammno Reoaub!4 ctXof aa P. 0, Box 10960 P. 0. BOX 114216 7"pbta RrFsub4 Rop%:IISu. du 0oorsb*own, BeirutAZ4bu=l c3alIadaih T7had P . O L o CulanA 1-703-4731 P.O. Bcx230 P.O. 3= 625 C Camila 197-D Qito, Euador BoilA paile 557 MAaru 10 0. Brldzotown, ALX.LAo,CSl PCri.azPrioo*, o B.arbdos P9 0. los 932 Hait 2 DonagqlJ t Posul Cade N. P.. Box 10274 0.P.o Box 103 San.1t1ut 1151S Aptrts"Postual 975 1ow mctucya 10 =M fc micAk Bo9lnI. PNpLa's Csleo, Arab T'gusigalpA D.C., Llberi-s, West B44:uA RkpubUa of Chim RapubLic o? 3 cp otduts Aric 3t$ poslas 506 Aparuda AAso, P. C. Bos II14 P.C. Box 3Q9 C d iralPon Oft s Czoncu., Ripubhi No. 091369 SAn S&ivsdor, Et Now DLdli ttcat. Box 62 of 3ocio BOjA, Colomia Salvdor L"diA LT.2000 ViwUs, Ll%uankl 3 P . 5 7 S BOits polo" 143 P.O. Box 5364 Ois.adougou0t ?orcj,j, F74r.J AsrA RsLxvindra mention P. 0. Box 351 Ouszrdou5ou, 1sva R.spub Erits (22nd fIoct TripolI, Great Burkina Faso of Comoacs Dinsza Vao,ha Socialal Puopl's UNDP Mil6ia, RoAd Lbyan a. P. 0. 30% 162 Bait3 postiat 465 Equa:triai Culon Barmty 402 , Arnb JAnahirty, Tbwhiru, Bhutan B A Br r a v YI I I a o UNDP, P.O. lndia PUopla's RpubUc Box 1601 P. 0. BoX 13AS 'GFF Wor;ing Papens are avalabte at UVDP Fzd COfflc and chrough WorlXd Bak Distribuor5. AnntarnalYo 101, NiArmy, Nzget Bota Postal4e 445 SwletAaiSM mr. floor) Rapubijo of Kgall, kwa2ds Waainr4oa .C. UAIaaacat P. 0. Box 2851 P. 0. Box 23217 2 0arkl, Abuj P. 0. Box 1146 Damascus, syrian P.O. Bx.3O135 Jlddah 21431, At. blpubIc Plet publbu of Lllogw.3, a1wi P. 0. Box 2073 Saudi Arabit Yu poSlvia LAgo, Nigjeia P. O. JoxS12 ( S b r b i a P. 0. Box 12544 P.O. BoxS3S Port-ot-Spain, Momn.mgs) S0712f Xuall t;PO Rox403 Riyadh 11421, Trinidad and Xza t oz a r a lunpur, Malaysia XarmoJ 3, PaIi:n Studi Ambis Tobqp mukocla P.O. Bcx 2058 P.. Q ox 1051 P. O BoX lOll Io.P.O. Bex 51 I Male, Repubio of Ihsamabad, nSia Prealown. SiSrra langkok loso1 , B a n i a Mildlves Leoor Ttzilaw Ierz*,g ovIz, Apatudo 6314 Croatia B at1apd120 FnaluaS, Pina= PeivateM iaBg BoltsP ,911 5 I a v r a i a lamka, MaU Apia, Samoa Loin, ToSo Macedoria CailIIA da coar. *lo Visual Bolo poswe 620 1107 CaI PoStal 109 Bol po*i*8E3 Lzui11t1aM1C1caw" Nouekbhott, Asur4loa, haray SRo TOME, Sio O3ST 7uis,, TwI, %am m 17 Mauritania Tomi and Prinipl P.O. Box SOO Apana441f0 P. X. 407 A-1 400Vian P.O. Box 253 Naclorw Unidas olo Mnistry of 06043 Ulus Part Louit, LinA, PomU Environranmt, Anka,T rnkey Boits poalo 7248 M2urhalu Economic, 812aMc p Ripubll P.O. BoxR72 Piaiswgad 6 Slchnavogo of Zale Aperado Poatal Dcmeglo Alxpon zlneal Relations Povstatya 105-39 Pod o Lork Mot Fleurl Street P. Q. Bx 31956 1151 Mezxica, Box P.O.BOx656 K.iv 262010, Luska, Repubi of D.F., Mexo Patay cthy, Moto Victoria, M.hi UI"ind Zimbla Marila Seychelles Room 222/224, Pit1ippi. P, , 0Box 3490 r.O. Box 4775 Hotl Codiu Bolte pals 154 Abu Dhabi, United Ilanrs, Zlinbabwe 277012 Cllainatu P.O. Box 2865 Dakar, Senegal Aab Enirstas Ztpobli; of 106l Maia Moldova Philippiln P.O. Box 9181 P.O.Box24 Dar-6e-Saglam, P.O. Box 4927 P. 0. Box 1041 Modiaclo, Somali Uriud RapubiU of U I s s n b m a t a e, Part Moresby, Democrat! TanD2tn Momsolla Papua New Guiri Republik P. 0. Box 7114 Catier ON Y1, P.O. Box I UNDP MoSadlalo KauVala, Republic RAbst-CbsUta Warsaw 12, Poland a/o UNDP Nairobi of Uganda Rabat, Moroocc K e n y a t t a Box3233 Intrrnt Ional Camile dA Corrwo P. 0. Box 4595 Doha, State of Conference 1207 Maputo Qaur Catm' 11200 Mootavidoo, Repub iic of P.O. BOx30218 Ur y Mozambique P.O. Box 551 Nairobi, Yenya Sanai a 4Taras Shovychank P.O. Box 630 Republic of Yanie P. 0. Box 1305 strt Yaonr, Myanw Colombo, Sri Tashkent 7009 Cantrl PAt OffIso Lanka UzWldAn Privts BU 133129 Box 143 hndhookdbana3bib S.oul, A*u Code P.O. Box 913 ApItfldo 69005 100l401 1XGamum,Republic P. O. Box 107 Rupublhic OfrX' of Sudan xthlmandu, Nepal 2 PItng lhte Stride AureI VIPAI= Privta Bag hn Apartado Postal nr. 16 M a X b a a a Ho Ch}MIMnh 3260 s4gtoNl 2, P. 0. SwZllAnd Socialist RIIbliO M a a a g u a Box 1-701 of Viat Nam Nicartsus 79362 ahares, Pidais de$ jVit;0j Romania CH-i1211 lt a;v 1889 P Street, Boitx poitala 11207 10 N.W. - The GEF Administrator 1 818 H Street, NW Washington, DC 20433 USA Telephone: (202) 473-1053 Fax: (202) 477-0551 United Nations Development Programme GEF/Executive Coordinator One United Nations Plaza New York, NY 10017 USA Telephone: (212) 906-5044 Fax: (212) 906-6998 UNEP United Nations Environment Programme GEF Unit/UNEP PO Box 30552 - Nairobi, Kenya Telephone: (254-2) 621-234 Fax: (254-2) 520-825 The World Bank GEF/Operations Coordination Division Environment Department 1818 H Street, NW Washington, DC 20433 USA Telephone: (202) 473-6010 Fax: (202) 676-0483 3 Printed on recycled paper ISBN 1-884122-05-1