Polloy Resach l ..WRKNGPAPERS World Development Report Office of the Vice President Development Economics The World Bank August 1992 WPS 957 Background paper for World Development Report 1992 Carbon Taxes, the Greenhouse Effect, and Developing Countries Anwar Shah and Bjorn Larsen A universal case cannot be made for national carbon taxes. Nevertheless, such taxes make eminent sense for many develop- ing countries - on the grounds of equity, efficiency, ease of tax administration, and an improved local environment, even ignor- ing the potential benefits from controlling global carbon emis- sions. Policy RahWokingPaedismianatetfhpi ofwo inpogaus d wgethecxchangcof ideas maongBanl staff and aflomtbiedidcvelopmimucLbescpaps.distrbdbydcRcswchAdvusoxySff,canythenamesofteuthws,reflect alydxeviews,sndsbcdbemedudczedilacoincy.Tfindiniprwnmandconclusionswethcauoown.Theyshoid nat be atbtd totheWodd Bn, its Bond of Diomimits managn, or ny of is mnber eounti Policy Rosrarhl World Devrelopment Report WPS 957 This paper- a product of the Office of the Vice President, Development Economics -is one in a series of background papersprepared forthe WorldDevelopmentReport] 992. The Report, on development and the environment, discusses the possible effects of the expected dramatic growth in the world's population, industrial output, use of energy, and demand for food. Copies of this and other WorldDevelopmentReport background papers are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact the World Development Report office, room T7-101, extension 31393 (August 1992, 68 pages). Shah and Larsen evaluate the case for carbon * A carbon tax can significantly reduce local taxes in terms of national interests. They reach pollution and carbon dioxide emissions. Cost- the following conclusions: benefit analysis shows countncs with few or no energy taxes substantially gaining from carbon A global carbon tax involves issues of taxes in tenrs of an improved local environment. intemational resource transfers and would be difficult to administer and enforce. It is thus * A carbon tax of $10 a ton produces very small unlikely to be implemented in the near future. output losses for Pakistani industries analyzed in this paper, and the output losses are fully offset by * National carbon taxes can raise significant health benefits from reduced emissions of local revenues cost-effectively in developing countries pollutants - even ignoring the global implications and are not likely to be as regressive in their of a reduced greenhouse effect. impact as commonly perceived. Such taxes can also enhance economic efficiency if introduced * Tradable permits are preferable to carbon as a revenue-neutral partial replacement for taxes where the critical threshold of the stock of corporate income taxes or in cases where subsi- carbon emission beyond which temperatures dies are prevalent. The welfare costs of carbon would rise exponentially is known. Given our taxes generally vary directly with the existing current ignorance on the costs of reducing level of energy taxes, so a carbon tax should be carbon emissions and the threshold effect, a an instrument of choice for countries such as carbon tix aopears to be a better and more India and Indonesia, which have few or no flexible instrwment for avoiding large unex- energy taxes. pected oosts. ThePolicy ResearchWorking Paper Saiesdisseminates thefindigs of workumder way intheBank Anobjectiveofthe series is to get these findins out quicldy, even if presntations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent official Bank policy. Produced by the Policy Research Dissemation Center Carbon Taxes, The Greenhouse Effect and Developinn Countries Anwar Shah* and Bjorn Larsen* World Bank, Washington, D.C. Prepared as a Background Report for the World Development Report 1992 * This paper explores a range of ideas originating from Mr. Summers. The authors are grateful to Andrew Steer for his guidance and support and to Ken Piddington, Nancy Birdsall, Dennis de Tray, Shankar Acharya, Sweder van Wijnbergen, Shahid Chaudhry, Mohamed El-Ashry, David Pearce, Ravi Kanbur, Sudhir Shetty, Nemat Shafik, Patricia Annez, Gordon Hughes and Dennis Anderson for helpful discussions and/or comments. The World Development Report 1992, "Development and the Environment," discusses the possible effects of the expected dramatic growth in the world's p pulation, industrial output, use of energy, and demand for food. Under current practices, the result could be appalling environmental conditions in both urban and rural areas. The World Development Report presents an alternative, albeit more difficult, path - one that, if taken, would allow future generations to witness improved environmental conditions accompanied by rapid economic development and the virtual eradication of widespread poverty. Choosing this path will require that both industrial and developing countries seize the current moment of opportunity to reform policies, institutions, and aid programs. A two-fold strategy is required. * First, take advantage of the positive links between economic efficiency, income growth, and protection of the environment. This calls for accelerating programs for reducing poverty, removing distortions that encourage the economically inefficient and environmentally damaging use of natural resources, clarifying property rights, expanding programs for education (especially for girls), family planning services, sanitation and clean water, and agricultural extension, credit and research. * Second, break the negative links between economic activity and the environmen:. Certain targeted measures, described in the Report, can bring dramatic improvements in environmental quality at modest cost in investment and economic efficiency. To implement them will require overcoming the power of vested interests, building strong institutions, improving knowledge, encouraging participatory decisionmaking, and building a partnership of cooperation between industrial and developing countries. Other World Development Report background papers in the Policy Research Working Paper series include: Dennis Anderson, "Economic Growth and the Environment" Dennis Anderson and William Cavendish, "Efficiency and Substitution in Poilution Abatement: Simulation Studies in Three Sectors" William Ascher, "Coping with the Disappointing Rates of Return of Development Projects with Environmental Aspects" Edward B. Barbier and Joanne C. Burgess, "Agricultural Pricing and Environmental Degradation" Robin W. Bates and Edwin A. Moore, "Commercial Energy Efficiency and the Environment" Wilfred Beckerman, "Economic Development and the Environment: Conflict or Complementarity?" Richard E. Bilsborrow, "Rural Poverty, Migration, and the Environment in Developing Countries: Three Case Studies" Charles R. Blitzer, R.S. Eckaus, Supriya Lahiri, and Alexander Meeraus, (a) "Growth and Welfare Losses from Carbon Emission Restrictions: A General Equilibrium Analysis for Egypt"; (b) "The Effects of Restrictions of Carbon Dixide and Methane Emissions on the Indian Economy" Judith M. Dean, "Trade and the Environment: A Survey of the Literature" Behrouz Guerami, "Prospects for Coal and Clean Coal Technology" David 0. Hall, "Biomass" Ravi Kanbur, "Heterogeneity, Distribution and Cooperation in Common Property Resource Management" Arik Levinson and Sudhir Shetty, "Efficient Environment Regulation: Case Studies of Urban Air Pollution" Robert E.B. Lucas, David Wheeler, and Hemamala Hettige, "Economic Development, Environmental Regulation and the International Migration of Toxic Industrial Pollution: 1960-1988" Robert E.B. Lucas, "Toxic Releases by Manufacturing: World Patterns and Trade Policies" Ashoka Mody and Robert Evenson, "Innovation and Diffusion of Environmentally Responsive Technologies" David Pearce, "Economic Valuation and the Natural World" Nemat Shafik and Sushenjit Bandyopadhyay, "Economic Growth and Environmental Quality: Time Series and Cross-Country Evidence" Anwar Shah and Bjorn Larsen, (a) "Carbon Taxes, the Greenhouse Effect, and Developing Countries"; (b) "World Energy Subsidies and Global Carbon Emissions" Margaret E. Slade, (a) "Environmental Costs of Natural Resource Commodities: Magnitude and Incidence"; (b) "Do Markets Underprice Natural Resouce Commodities?" Piritta Sorsa, "The Environment - A New Challenge to GATT?" Sheila Webb and Associates, "Waterborne Diseases in Peru" Background papers in the World Bank's Discussion Paper series include: Shelton H. Davis, "Indigenous Views of Land and the Environment" John B. Homer, "Natural Gas in Developing Countries: Evaluating the Benefits to the Environment" 3tep'ien Mink, "Poverty, Population and the Environment" Theodore Panayotou, "Policy Options for Controlling Urban and Industrial Pollution" Other (unpublished) papers in the series are available direct from the World Development Report Office, room 17-101, extension 31393. For a complete list of titles, consult pages 182-3 of the World Development Report. The World lDevelopment Report was prepared by a team led by Andrew Steer; the background papers were edited by Will Wade-Gery. Table of Contents 1.0 Introduction The Problem and the Status of Current Policy Discussions 2.0 Global Carbon Taxes: Potentials and Perils 3.0 Economics of a National Carbon Tax 3.1 Revenue Potential of Carbon Taxes 3.2 Distributionel Implications of Carbon Taxes 3.3 Efficiency Costs of Carbon Taxes 4.0 The Impact of Carbon Taxes on Greenhouse Gases and Local Pollutants 5.0 Carboi Taxes, Industrial Performance and Economic Growth 6.0 Tradeable Permits 7.0 Some Conclusions Appendix A: The Differential Incidence of Carbon Taxes References 1.0 Introduction The last few years have witnessed a dramatic growth in worldwide concern over global climate change and a proliferation of proposals to limit or reverse global environmental damage. Carbon taxes and tradeable permits figure prominently in proposed economic policy responses. Despite this policy interest, empirical work relevant to developing countries is almost completely lacldng. Even for developed countries, research on carbon taxes is of recent origin (see e.g. Jorgenson and Wilcoxen 1990, Poterba 1991, Pearce 1991 and Goulder 1991) and still largely in progress. A careful analysis of carbon taxes in terms of their impacts on efficiency, equity, economic growth, government revenues and environmental protection, is needed for an informed debate on policy development (see Summers 1991). This paper takes a first step in this direction by quantifying the efficiency and equity implications of carbon taxes for a few selected developing countries. The paper is organized into seven sections. The remainder of Section 1 outlines the global warming issue and suggested policy responses. Section 2 briefly outlines the potentials and perils of global carbon tax regimes. Section 3 deals with the economics of a national carbon tax. Calculations on the revenue potential and differential incidence of carbon taxes are 1,-esented for India, Indonesia, Pakistan, USA and Japan. For Pakistan, detailed calculations on the distributional implications are also presented. Section 4 provides estimates of the impact of carbon taxes on greenhouse gases and local pollutants for the sample countries. Impacts of carbon taxes on industrial performance for selected industries in Pakistan are traced, in Section 5, using dynamic production structure empirical models. Section 6 evaluates the use of tradeable permits as an alternative to carbon taxes. A final section presents a summary of the conclusions. The paper concludes that whereas a global carbon tax may be a more distant policy option, national carbon taxes -- if introduced in revenue-neutral fashion by reducing corporate income taxes -- offer significant potential in combatting global change and local pollutio- as well as reforming the tax system. Further, a conservative evaluation of the benefits of reducing local externalities overwhelms the negative output effects of carbon taxes. Thus, even ignoring global externalities, a case for carbon taxes for some countries can be made purely on own national interest considerations. The Problem and the Status of Current Policy Discussions In recent years, worldwide concern about the atmospheric accumulation of so-called "greenhouse" trace gases - carbon dioxide (CO,), methane (CH4), nitrous oxides (N20), tropospheric ozone (03), and chlorofluorocarbons (CFCs) - has been mounting. By trapping some of the sun's heat in the atmosphere, these gases permit the existence of life on earth. Their rapid accumulation, however, can contribute to a rise in the earth's temperature (commonly termed the "greenhouse effect" or "global warming"). C02 is estimated to contribute 80.3% of total warming potential (Nordhaus 1991). Scientists fear that if the current pace of accumulation continues unchecked into the 21st century, a point might be reached when the absorptive capacity of the earth's atmosphere would become exhausted and a natural disaster of unprecedented proportions would consequently ensue. Even without this point being reached, significant warming of the earth's surface is expected to have major economic consequences (see 1 Churchill and Saunders, 199 1, for an overview of the scientific and economic issues of relevance to developing countries). Developing countries with agrarian economies and/or coastlines would be particularly vulnerable to natural calamities associated with global warming. It must be emphasized that there is considerable uncertainty at the present time regarding global climate change, its magnitude, its regional manifestations and its consequences. Much scientific work remains to be done. The uncertain state of our present knowledge of global warming coupled with the potent-ally large and irreversible damages that might result, call for public policy responses that are both flexible and reversible. The possible use of carbon taxes and tradeable permits to deal with global climate change has initiated a controversial debate. These debates reflect a wide spectrum of views on this issue. Some argue that in view of the uncertainties regarding climate change, inaction would be the best policy (Eckaus, 1991). At the other extreme, some environmentalists argue that we may have already missed the boat and immediate economic policy responses that may forsake growth are needed (see Postel and Flavin, 1991). A majority, however, take a middle view. Energy economists argue that energy policy options consistent with restraining the greenhouse effect also make good economic sense. Churchill and Saunders (1991), for example, exhort developing countries to seize the initiative and 'increase incentives for sustainable energy use, shift to cleaner alternative fuels and technologies, and improve efficiency in energy production, distribution and end use" (p.28). Some public finance economists espouse the same middle-of-the-road view by presenting a somewhat different perspective that emphasizes reliance on flexible and less distortionary tools to deal with an uncertain but potentially serious problem. Summers (1991) has argued that corrective taxes, e.g. taxes on carbon contents of fossil fuels can raise significant amounts of revenue at a relatively small deadweight loss while furthering global and local environmental protection and discouraging "bads", and therefore represent "what we pay to preserve civilization". It has also been argued that in developing countries, carbon taxes offer a potential for enhancing the environment as well as financing developmental expenditures, and could therefore serve as a means to enrich civilization. It is interesting to note the large energy subsidies that prevail in a handful of large carbon-emitting countries. Getting energy prices right would prima facie represent a first order priority in any economic policy response designed to curtail greenhouse gas emissions. Larsen and Shah (1992a) examine energy pricing practices around the world. In determining the level of subsidy, they use border prices of fossil fuels as reference prices (as proxies for marginal opportunity costs of production). Total world energy subsidies in 1990 are estimated to be in excess of US $230 billion and, in revenue terms, equivalent to a negative carbon tax of US $40 per ton of carbon. The removal of such subsidies could reduce global carbon emissions by 9.5%, and would translate into a 21% reduction in carbon emissions in the subsidizing countries. To achieve an equivalent reduction in tons of emissions in the OECD countries, a carbon tax of US $60 per ton would need to be imposed in the OECD countries. This would result in a total annual cost (in terms of foregone output, adjustment costs, etc.) of US $15.5 billion. This amount would then represent the upper bound for OECD compensatory transfers to the subsidizing countries. It is also worth noting that very large (37-68%) reductions in global carbon emissions could be achieved, were Japanese or German standards of energy efficiency to be universally adopted. 2 While this debate continues to rage, some countries have already moved to adopt tax policies that, intentionally or otherwise, bear on the issue of global warming. In late 1989 and in response to ozone depletion, the USA introduced a tax on the sale of CFCs at an initial rate of $3.02 per kg, representing a 200% tax on the sale price. This tax is scheduled to rise to $6.83/kg by 1995 and to $10.80/kg by 1999. Total revenue intake during the first five years is estimated to total $4.3 bil.on. The USA has not yet imposed a carbon tax but legislation is currently pending in the U.S. Congress for the phased introduction of a carbon tax to start at $5 per ton of carbon in 1991, rise to $25 per ton in 1995. Proposals to increase excises on gasoline and introducing a Federal BTU tax are also under disussion. Note that the U.S. Government remains uncommitted to a targeted policy response to climate change other than advocating economic poiicies which make good economic sense independently (the so-called "no regrets" policies). Among European countries, Finland took the lead in introducing the world's first carbon tax at a rate of $6.10 per ton of carbon on all fossil fuels in January 1990. Netherlands and Sweden ($45/ton tax) have followed suit in February 1990 and January 1991 respectively. The European Community is currently debating a proposal to introduce in a revenue-neutral manner a community-wide carbon-cum-energy tax at US $3 per barrel. The tax would increase by $1 a barrel each year in real terms until it reached $10 per barrel (roughly equivalent to a carbon tax of $70 per ton) in the year 2000. At the international level, momentum has steadily built behind the proposition that global warming and other aspects of climate change are of major consequence and require a concerted global policy response. In 1990, the UN General Assembly formally launched international negotiations on a "Framework Convention on Climate Change" and assigned this task to an "Intergovernmental Negotiating Committee"(INC). The INC has held conferences in Geneva (1990, 1991) and Washington (1991). These conferences have debated international protocols to limit emissions of "greenhouse gases. A global climate change "framework convention" is likely to be ratified at the June 1992 UN Conference on the Environment and Development to be held in Brazil. The discussion in these international fora has cl nitered on both domestic and global policy options to combat global climate change. These have included: immediate term options such as a global carbon tax or permits (tradeable or otherwise) and emission limits; intermediate term measures such as increased energy efficiency, afforestation, biomass, nuclear energy and population control; and long term measures such as backstop technologies that use solar, solar-hydrogen and other environmentally safe sources. Developing countries are fully involved in the debate on these issues. One argument often advanced is that the greenhouse effect results from the accumulation over a long period of trace gases contributed primarily by industrial activity in developed countries, and consequently that developing countries should not be asked to sacrifice their current developmental-goals in order to address a problem created by past policies of developed countries. In fact, if one were to construct an index of "global warming debt" by level of development, this particular argument would have some empirical validity (see Smith, 1991). It is also frequently asserted that any global attempt to limit environmentally harmful emissions would ultimately slow the economic development of LDCs. Attempts to develop energy intensive manufacturing capability in the early to mid stages of development would be more costly, and hence more difficult. Also as importers of energy intensive manufactures (primarily capital goods), developing countries would end up bearing the burden of the policy response applied to emission generating activities. In general, it is 3 commonly perceived that, unless accompanied by compensatory transfers, the relative costs of action are likely to be higher for developing countries, given that their relative contribution to the accumulation of these gases is expected to grow faster than that of the OECD countries over the next century. The available literature offers little guidance in determining the validity of these arguments. The following section provides pre!iminary and tentative guidance on these questions. 2.0 Global Carbon Taxes: Potentials and Perils Taxes on the carbon content of fossil fuels have been advocatAx in recent years as part of a proposed concerted international effort to combat global climate change. While both the need for and the mechanics of such taxes remain unsettled issues, a general consensus is emerging that, if adopted globally, such taxes would represent a flexible, reversible and lower cost alternative to regulatory responses, including the widely-discussed notion of equal percentage reductions in greenhouse emissions by all countries. The latter measure is unlikely to lead to the equalization of marginal emission reduction costs from all sources and would not, thereforv, result in a cost efficient outcome for the world as a whole (see Hoel, 1991). Tietenberg (1985) reports that cost savings associated with moving from equal percentage reductions to a market based instrument such as a carbon tax, could be substantial (exceeding 40% of total costs). Maler (1989) also reports that a uniform percentage reduction strategy for greenhouse gas emissions would capture only one third of the total potential gains from optimal allocation. A uniform level carbon tax (i.e., tax per unit of carbon emissions equal for all gases and all countries), if imposed by a global agreement, would equalize the marginal costs of emission reductions (by fossil fuel and by location), and would therefore be cost-efficient. Several alternative designs for such an agreement are possible, with each presenting its own particular shortcomings. Consider the case of a domestic carbon tax that is imposed by an international agreement. Since perspectives on global warming vary among countries, national commitment to impose such taxes will also vary. If a country has signed such agreement under international pressure, that country can make the carbon tax an ineffeclve instrument by reducing existing energy taxes, by taxing close substitutes of fossil fuels (e.g. hydroelectricity), and providing subsidies to complements or products that are fossil fuel energy intensive, and by lax enforcement of the agreed-upon carbon tax (see Hoel 1991). Thus by following a suitable strategy, a free ride becomes possible. A global carbon tax imposed by an international agency, on the other hand, would impinge on national sovereignty and therefore would not likely be accepted internationally. A third alternative would have globally imposed but nationally administered and collected carbon taxes; countries would make a positive or negative net transfer to an intemational agency based upon an agreed revenue disposition scheme. Basic criteria for such redistribution would be population and GDP, or a combination of these factors. Additionally, a small fraction of the revenue pool could be distributed on the basis of special considerations, e.g. to provide an inducement to countries which might view global warming as beneficial (such as Russia, Canada and Nepal) to join an international agreement. Tables 2.1 and 2.2 provide illustrations of net transfers involved based on the three revenue redistribution schemes outlined above, using either standard GDP or GDP adjusted by purchasing power parity -- so-called Penn GDP. From these tables, it is apparent that a revenue redistribution alternative 4 Table 2.1 Net transfers by couatry under akeiate carbon tu regimes (Using UN Nadonal Accounts GDP) 1erld too CDP iWrld t ttc Vrld tait Nwt rev *w ltc per Carbon Carbot fell ee rivaoFd "rOvInite MC revuuas fever"" Not dletribut#d rtvma Net Capita 2 migale w "tea toe atytete rovetlA, distrit4d Per CaPIta tta distributed per capita raveut ball aid half pat capit., reveenDt (UfS') of acrid to per (less to by populallot (UMS 1.of by CDP (111111 It of by pqiil*ticfn USS)l I of 1987 4MIaliad GDP el't 6SitOItcnh M5 3 mi 15? OUmSliiS liii) PGaDP Oen/S (II) mill MSS(il 3 BAUO4MISI *~~~~ 66, 0.063 0.119 30 32 0.182 1160 10.82 6.502 54 C.al 0.131 6,11 M.l S.111 *229 0.163 0.3I66 84 90 0.373 1165 10.11 4.482 13 .0.14 .0.062 630 5.0? 1.212 Cal" *286 10.382 1.8 n 31 569 1.017 lien 5.19 21.052 934 X.46 -1.561 6408 0.66 0.231 tualt ~* 322 1.652 0.56? l8i 1454 0.57I an6 9.30 2.892 785l .0.84 *0.263 4127 4.23 I.-2n PAEIIIAN a 323 0.242 0.394 128 132 0.392 1'140, 9.84 5.02n 102 .0.29 .0.092 621 4.17 1.4111 fohalA a 443 0.4821 0.346 153 263 0.352 1906 9.59 2.162 232 .0.18 *0.043 1069 4.11 1.061 11368*13 * 198 ~~~~~0.083 0.774 461 42 0.773 100 6.49 1.0921 16 .2.80 .0.472 s8 1.8s 0.3111 SOIpI. ARM1XVIAIVtc 70 0.353 O.536 380 I90 0.542 11510, 7.3I 1.011 ¶09 *I.63 -0.23213 28 .0 fUt1A. SIN flptWI. a 889 0.713 2.063 1834 392 1.062 2) .22 .0.811 so *I5.62 -.1.6% 148 *11.42 -1.283 fatal 1S.1821 829 AV"ra" 316 i.07n 31.1 1.0111 7.78 2.452 .2.30 -.o.2 2.74 0.863 AVn va I CAIe tter 0.491 ,0ag@ *171 1.413 0.550 943 77 0.551 910 1.69 0.102 40 -4.16 -.0.41 670 .1.25 -0.0711 SAZIL 2145 0.923 0.166 356 503 O.l3 1573 7.5t6 0.35X 928 3.01 0.142 1Il1 5.29 O.J53 SOUIl AF6CA *493 1.38r 0.919 *202 n5 0.92° 36 *IIu o , -0.41 15 -I5.29 0.611 sic -13.5n .0.543 vtazwta * 2629 0.421 0.485 1276 233 0.692 203 .*64 * 2 14 .4.71 *0.182 175 **3.I7 -0.i2 gORI*. mim1c 0 3 0.8221 0.342 i067 449 0.341 46 0.45 0.012 40 -1.12 ' 0.043 435 ° -6.1O 0 Total .2716 Average 22t65t Soo 0.365 851 O.381 2.55 0.11o s9 .16 00 .OX0 0.39 oo03X Ave ala, S. Africa 0.114 084613 * 1697 2.291 1.967 ~~~~~ ~~3338 ItS? 1.972 419 -22.26 *1.312 196 .26.18 -l.663 30? -23.22 *I.492 IIJO3AVI 2869 2.0 0.49 1 406 328 !~.491 260 -2.91 *0.t02 204 -.31S -0.192 232 .4.11 -0.143 USSR ~ ~ 825 18.453 0.430 i578 10129 0.433 3140 -24.66 .0.392 1213 .10.30 *0.if 18 174 -.1 CUCIMIIL0V5AIA * 9242 ~~1.1173 0.44. .110 640 0.442 173 .29.96 *0.321 440 112.82 *0.141 307 *2I.40 -0.233 CUOa.W. AK.1 a9261 .3 .7 36 9 .6 6 -42.5? -0.81a 574 -19.23 -0.172 319 .30.90 -.02731 total 24.13X ¶3248 Average 7489 0.470 3520 0.4112 *24.06 -0.3211 *12.28 -0.163 *l8.18 O.0.41 AV* */a Paadvi 0.43u IISSi* 11364 1.161 0.346 3926 638 0.32 11 -261.14 -O..5 5.5 -4 .43 .0.861 Sri - 1 50.142 CAWAD* 16056 I.9"S 0.263 4221 1091 0.261 287 -31.09 -012 27 6.3 .4319 *28 *5 cermw. We,t *18249 1.231 0.159 289 11M 0.16.2 080 *17.86 .0.101 341d6 26.871 0.151 2048 4.50 0.021 1*11110 61*116 * a434 22.691 0.277 5ill 12461 0.281 2711 .40.co -0.221 135 .30 0.031 8232 *I7.35 *01)92 JAP36 19437 4.321 0.100 1942 2371 0.1021 1358 *8.30 -0043 26 40.07 0.213 4310 15.88 0.0812 total IBM9 33.392 18334 Aver"Ag 0.214 1908 0.212 .27.96 -0.153 16.01 0.092 -.S~3 -0.031 $&ple total 77.S61 AFIiCA "4S 2.601 0.405 261 1540 0.402 6561 8.51 1.322 1164 -0.64 -0.102 38162 3.94 0.61% SOUlS AMaiNCA 1951 2.461 0.249 486 1350 0.253 3091 8.26 0.322 1660 1.12 0.061 2376 3.69 0.192 cum"0 11136 21.133 0.210 2343 11600 0.211 5504 -12.31 -0.112 16870 10.6S 0.10 lox187e -0.83 -0.012 asmhN S 0251681 A836612897 26.043 0.281 3620 14300 0.282 4392 -25.08 -0.191 i5591 3.27 0.032 9992 *i0.91 .0.082 002881A 9372 9.311 0.319 2992 718 0.322 267 -18.80 -0.20111 688 -1.23 *0.Oil 478 -10.02 -0.112 ASIA 1333 24.491 0.350 466 13400 0.353 31947 6.46 0.482 11123 -0.S$ -0.043 21835 2.94 0.223 USIA 6328B 18.59X 0.429 3569 10100 0.433 3147 .24.57 0.1301 7213 .10.29 *0.i21 SIBO -17.39 .0.212 L43" 5~~633 100.002 0.306 1112 54910 0.313 54910 0.00 0.0021 54910 0.00 0.0021 54910 0.00 0.002 iota: carOt eia001m era ftam femail fuel coomtion antly. Emlealwn. from deforetattion era not included. Table w.2 Net transfers by country under global carbon tax regimes (Using PENN GDP) 13 23 ~~~~~~~~~~~~~~~~~~~~~~~~~~~3Ii World tex MAP kmitd tex W*et World to: Not leywee.. Not per Csrbon Corbon tla talxi rivegew 04m Not rev h revenu Not diettlbuted evris Noet k43105 1 e.lglom. onleulow Sevemme rowevee. distributed per copias rewrues distributed per nopltO otOviwAl belt .ee hall per capita revause (UM of orid to Pem per (tlet to Pam by poitetia (USE) 1 of P.m by Pam CDP CUSS) I of by Pam GOP USS) 1t of 197 emission GP cepIto sI0ton) CDP tS) (mlil U21 Cap (ofIT US51 Pm CGDP ad peouletIon Pa OW (ko's) (kg) ml .11111 (Sl itB' USE) MICInIA * 248 0.161 0.339 64 0.342 tX 53 10.28 4.141 71 o0.1e 7 0.00 628 *.0S 2.0r6 MAMUOIW * *631 0.069 0.034 30 32 0.041 1160 10.8* 1.301 230 1.92 0.23X 706 6.3 0.r6115 I0011 * 947 2.611 0.193 lot 1454 0.191 66 9.30 0.8 2014 0.70 0.010 $441 1.00 0.531 lINPAIW 104? 0.081 0.442 463 42 0.441 100 6.49 0.06X 25 -1.84 5 0.161 63 2.33 0. in 1000W11 X 1269 0.463 0.121 153 263 0.121 1906 9.39 0.761 S80 1,65 0.15 1243 i.7,2 0.4*1 PAKISTAN I 1391 0,241 0.092 126 132 0.091 1140 9.84 *0.6 30 2.43 0.1726 60 6.1C 0.441 SOIPI. AlAS MIWLIC* 1454 0.3S1 0.261 360 S 90 0.261 25e 7.32 0.501 19 0.06 0.011 376 3.70 0.251 0110* a 2233 10. 36X 0.239 S33 5699 0.245 1166 5.79 0.261 6)70 0.63 0.031 9126 3.21 0. 14 1.t.1 14,391 7900 'Vote"e 1534 0.213 328 0.211 7.64 0,511 0.62 0.031 4.53 0.261 WIIUIMALA * 3422 0.421 0.373 1276 233 0.370 203 .1.64 -0.051 167 .3.63 -0.111 165i -2.63 *0.062 MEXICO * 3760 1.411 0.249 943 77 0.231 910 1.69 0.041 62) 0.6$ 0.022 666 1.117 0.031 010664. RIPUALIC Of * 42J6 0.621 0.252 1067 449 0.251 468 0.4$ 0.011 474 0.60 0.011 Art 0. 53 0.011 INA IL *4613 0.921 0.07? 356 503 0.061 1373 7.56 0.161 1740 6.715 0.191 16S7 6.16 0.162 $W0111 AFRICA * 592 1. 381 0.387 2292 7`59 0.391 366 *U1.M0 .0.201 523 -7.13 *0.121 446 9.947 -0. 161 total 4.9511 271 AvMete" 4414 0.194 65? v. 191 2.513 0.061 3.20 0.070 2.67? 0.0701 V%OoSLAVIA 5000 0.602 0.281 1403 Ull 0.281 240 '2.91 -0.061 312 *0.69 -0.011 *66 -1.60 *0.041 Pft0k* *159 2.292 0.968 3336 1257 0.601 419 .22.26 0.402 560 -1O.0 -0.331 490 -20.36 .0.371 usIs 6791 16.4511 0.S27 3516 10129 0.531 3148 -24.66 *0.361 5126 *17.66 *0.261 4138 -21.6 '0.311 Cit OSI.0ASIA * 6521 1.170 0.462 4110 640 0.481 173 29.9 -0.351 35 *16.38 -0.22l 264 e2l.16 o0.26. caronn. lost & 6767F 1.631 0.612 5369 694 0.611 183 -42.5? *0.491 769 -30.31 -0.351 26? '36.44 -0.421 fotal 24.131 13246 Aisl"e. 6Y17 0.326 3529 0.521 -24.06 .0.362 -17.28 -0.261 -20.46 -0.311 AUSTIAIA * 1176 1.161 0.333 3926 638 0.331 161 *28.14 *0.241 511 *7.63 *0.070 346 -17.99, '0.1511 JAPAN * 12506 4.321 0.155 1942 2371 0.161 1356 68.30 -0.070 4073 13.94 0.112 2713 2.6* '3.021 CANADA * 15730 1.992 0.266, 4221 1091 0.27,01 207 *31.09 *0.202 1063 -0.25 0.001 686 -15.67 .0.o0 0.irminn Waut * 16693 3.231 o.17n 2696 177 0.170 660 -17.66 -0.111 2716 ¶6.06 0.102 Iri6 -0.69 -0.011 wmill htAlO a 1130$ 22.691 0.292 511 12461 0.29 2711 -40.00 *0.231 11363 -4.42 -0.031 7047 -22.21 -0.131 total 33.390 l633 Avera"e 3826 0.247 3906 0.251 -27.96 *0.182 3.14 0.021 -12.41 -.0061 s2omie total n6.651 2 of insorid toota AFRICA 1061 2.601; 0.245 261 1540 0.251 6361 a.5 0.601 1677 0.23 0.021 419 4.37 0.411 142313 AMEWRICA 3976 2.4611 0.122 466 1550 0.121 3091 6.26 0.161 2949 5.75 0.14 3020 6.01 0.151 LimPS 11146 21.131 0.210 2534 11600 0.211 3504 -12.31 -0.111 14716 6.30 0.061 10111 -3.01 *0.0S1 aIni II CISIMAL MAlIN ¶2626 26.04X 0.262 3620 14300 0.262 4392 -25.06 *0.201 13S16 -1.96 -0.021 895 11.3.3 .0.111 OCEANIA 9624 1.311 0.305 299 716 0.302 26? *18.80 -0.191 629 -3.1I -0.041 446 -11.25 -0.111 ASIA 2126 24.402 0.219 466 13400 0.222 31947 6.46 0.302 16294 1.01 0.0$1 24120 3.73 0.161 USSR 679 16i39; 0.sz5 3569 10100 0.53X 314? -24.57 *0.36X 5126 -17.57 -0.262 4136 -21.07' -0.311 wiMIo 4169 100.0019 0.267 1112 $4910 0.270 54910 0.00 0.002 54910 0.00 0.001 54910 0.00 0.001 Notes Carbn eneseioue er. Irm ossiml usti comuction only. cl.elim' Orow dafer.tetien are rat Included. based on population alone would be unacceptable to most industrialized countries, whereas one based solely on GDP would not be agreeable to developing countries. Note that under the formula that uses population as the sole factor, net transfers to developing countries would dwarf current official development assistance. It is possible that a formula that uses a combination of both factors and therefore redistributes only a very small fraction of total carbon tax revenues, might find acceptance by a majority of countries. A recent proposal by Norway would have mandatory greenhouse reduction targets imposed on industrial countries; such targets could be exceeded only by financing the transfer and/or adoption of green technology in developing countries. If such a proposal is received well in OECD countries, some of these countries might well choose to adopt carbon taxes to achieve the agreed-upon targets, then partly use the proceeds from carbon taxes to finance technology transfer to developing countries. In general, the prognosis for the acceptance of a global carbon tax regime is quite pessimistic. The degree of scientific uncertainty that surrounds global warming makes it unlikely that a majority of countries would agree to an international convention that is seen to forsake their current growth. The critical question then is that if one ignores the important yet uncertain phenomenon of global warming, is there a case for the adoption of national carbon taxes on other grounds, such as tax reform or a reduction in environmental externalities? The following sub-sections present a benefit-cost calculus of carbon taxes based on these latter considerations. 3. Economics of a National Carbon Tax As discussed earlier, taxes on the carbon content of fossil fuels to combat global climate change have been widely advocated and also recently implemented in selected countries. In the following section, the case for carbon taxes is examined in terms of their revenue potential, efficiency and distributional implications, and impacts on global and local externalities. For the purpose of these calculations, a small fossil fuel carbon tax of the order of $10/ton of carbon contents is selected. Such a tax results in 2% and 8.6% increases in the aggregate price of fossil fuels, and 1.0% and 5.6% reductions in consumption of fossil fuels, in Japan and India, respectively (Table 3.1). Partial equilibrium calculations presented in this paper, offer reasonable and defensible approximation of the impact of small carbon taxes; the same confidence could not be asserted for those taxes of $100/ton or higher which are frequently discussed in global mo'dels. 3.1 Revenue Potential of Carbon Taxes The revenue potential of carbon taxes is extremely large. For example, a $10/ton carbon tax, individually imposed by all iaiions of the world could raise $55 billion in the very first year of its operation (see Table 3.1). For some countries, like China and Poland, such revenues would amount to about 2% of GDP and would be sufficient to wipe out central government's budgetary deficit. On the average, countries having a 1987 per capita GDP of less than US$900 could raise revenues exceeding one percent of GDP and 5.7% of government revenue. For the OECD countries, comparable figures would be 0.21% of GDP and 1.0% of government 7 revenue. Carbon taxes in general are easier to administer than personal and corporate taxes and thereby less prone to tax avoidance and evasion. Due to tax evasion, the latter taxes raise revenues that are considerably less than their potential yield. Carbon taxes therefore present an attractive alternative to income taxes in developing countries. But how do such taxes fare in terms of equity and efficiency? 3.2 Distributional Implications of Carbon Taxes The existing literature on industrialized countries typically portrays carbon taxes as regressive charges. This is because expenditures on fossil fuel consumption as a proportion of. current annual income, falls with income. Poterba (1991) relates carbon taxes to annual consumption expenditures -- a proxy for permanent income -- and still finds a regressive incidence, although one considerably less pronounced than with respect to annual income. These results nevertheless cannot be generalized to developing countries, where the incidence of carbon taxes would be affected by institutional factors. Some important factors that may have a bearing on the tax-shifting are: market power, price controls, import quotas, rationed foreign exchange, the presence of black markets, tax evasion and urban-rural migration. Case (a): Full Forward Shifting. The degree of tax-shifting depends upon the relative elasticities of supply and demand for the taxed commodity. For example, carbon taxes on production or use of fossil fuels can be fully forward-shifted in the short run if the firms in the industry have full market power, or the demand for the taxed commodity is perfectly inelastic, or the supply is perfectly elastic. In Table 3.2, columns (a) and (b) present carbon tax ($10/ton) incidence calculations for Pakistan using data from the 1984/85 Household Income and Expenditure Survey and employing two alternative concepts of household income. Column (a) relates carbon tax payments to household current income by income class and column (b) to household expenditure by income class. In either case, the carbon tax burden falls with income, thereby yielding a regressive pattern of incidence. Such regressivity is nevertheless less pronounced with respect to household expenditures, thereby confirming the same conclusions reached by Poterba (1991) for the US. 8 Table 3.1 Revenue Potential of a US $10/ton domestic carbon tax (Using UN National Accounts GDP) Carbon tax GOP Carbon Carbon revenues Carbon tax Popu- per emissions emissions Tax Tax to total revenues lation capita to per Revenues revenues Gov't rev gov't to gov't (mill) CUSS) GOP capita (Tax:S1O/ton) to to GDP revenues deficit 1987 1987 (kg/$) (kg) mill USS GDP (X) X K X BANGLADESH * 106.1 166 0.179 30 32 0.18X 9.12X 1.96X NIGERIA ^ 106.6 229 0.366 84 90 0.37K 15.71X 2.33X 4.18X CHINA *1068.5 286 1.868 533 5699 1.87r 21.19X 8.81X 262.31X INDIA '797.53 322 0.567 182 1454 0.57X 14.73K 3.85X 6.65X PAKISTAN *102.48 325 0.394 128 132 0.39X 17.29X 2.28K 4.63X INDONESIA *171.44 443 0.346 153 263 0.35X 21.33X 1.62X 21.97K ZIMBABWE * 8.99 s9a 0.774 463 42 0.77M 33.10K 2.34X 7.03X EGYPT. ARAB REPUBLIC* 50.14 109 0.536 380 190 0.54X 38.07X 1.41K 9.03K KOREA, OEM PEOPLE'S R 21.37 889 2.063 1834 392 2.06X Total 8292 Averages 1.07X 18.78X 5.71K 25.20K MEXICO * 81.86 1715 0.550 943 772 0.55K 17.41X 3.16K 4.06K BRAZIL ^141.43 2145 0.166 356 503 0.17X 33.29X 0.50K 1.42K SOUTH AFRICA * 33.11 2493 0.919 2292 759 0.92X 23.02X 3.99X 16.11K VENEZUELA * 18.27 2629 0.485 1276 233 0.49X 21.61K 2.25K 27.25K KOREA. REPUBLIC OF 42.08 3121 0.342 1067 449 0.34K 17.27X 1.98X -77.26K Total 2716 Averages 0.38X 25.16X 1.53K 4.56K POLAND 37.66 1697 1.967 3338 1257 1.97X 38.78K 5.07X 137.92X YUGOSLAVIA * 23.41 2848 0.492 1403 328 0.49X 6.86K 7.18K -1288.78K USSR * 283.1 8325 0.430 3578 10129 0.43X CZECHOSLOVAKIA * 15.57 9242 0.445 4110 640 0.44X 48.35K 0.92K 527.11K Gensrmwn East * 16.65 11261 0.477 5369 894 0.48K Total 13248 Averages 0.47X AUSTRALIA * 16.25 11364 0.346 3926 638 0.35X 26.50 1.30K 28.73K CANADA * 25.85 16056 0.263 4221 1091 0.26X 20.29X 1.30X -10.32K Germany, West * 61.17 18249 0.159 2898 lm 0.16X 29.34K 0.54K 15.01K UNITED STATES *243.77 18434 0.277 5112 12461 0.28K 20.23X 1.37K 8.45K JAPAN *122.09 19437 0.100 1942 2371 0.10X 13.77M 0.73 2.82X Total 18334 Averages 0.21K 19.77X 1.08X 7.81X Note: Carbon emissions are from fossil fuel coabustion only. Emissions from deforestation are not included. 9 Table 3.2 Carbon Tax ($10/Ton) Inddence - Pakistan 1984/8S (carbon taxes (TAX) as percent of monthly income (Y) or expenditure (EXP)) Monthly Income Full Forward Shiftin Canital Owners Caoital Owners (0.69) Consumotion (0.31) (Rupees) TAXIY TAX/EXP TAX/Y TAXIEXP TAX/Y TAXIEXP (a) (b) (c) (d) (e) (f) -600 1.49 1.19 0.66 0.53 0.92 0.74 601-700 0.89 0.83 0.62 0.58 0.71 0.66 701-800 0.91 0.86 0.64 0.60 0.72 0.68 801-1000 0.80 0.77 0.68 0.66 0.72 0.69 1001-1500 0.81 1.81 0.72 0.72 0.75 0.75 1501-2000 0.81 0.85 0.76 0.79 0.78 0.81 2001-2500 0.82 0.87 0.74 0.79 0.77 0.82 2501-3000 0.74 0.80 0.77 0.83 0.76 0.82 3001-3500 0.76 0.83 0.75 0.81 0.75 0.82 3501-4000 0.78 0.83 0.77 0.83 0.77 0.83 4001-4500 0.68 0.78 0.78 0.90 0.75 0.86 4500+ 0.51 0.67 0.80 1.06 0.71 0.94 Regressive Regressive Progressive Progressive Proportional Progressive 10 Case (b): Complete Absence of Forward Shifting. Under a variety of circumstances, the burden of carbon taxes can fall entirely on capital owners. This can happen if price controls apply and legal pass-forward of the tax is disallowed, or if supply is completely price inelastic. The carbon tax will then be fully borne by fixed factors of production. With binding import quotas or rationed foreign exchange, carbon taxes will reduce rents received by quota recipients, rather than affect prices paid by consumers. Under the assumption of zero forward shifting, the burden of a carbon tax is attributed to capital income alone. The allocation of tax by capital income is then related to household income and household expenditures. Both these calculations yield a progressive distribution of the carbon tax burden (see Table 3.2: columns (c) and (d)). Case (c): Partial Forward Shifting. Clearly, (a) and (b) above are polar cases and are unlikely to be fully satisfied for energy products in any country. There are only a handful of empirical studies which examine shifting assumptions for developing countries. One such study was carried out for excise taxes in Pakistan by Jeetun (1978). He finds 31% forward shifting of excises in Paldstan. Given than a tax on the carbon content of fossil fuels at their production stage is by its very nature an excise tax, it would be reasonable to use this assumption for assessing the distribution of the carbon tax burden. In Table 3.2, columns (e) and (f), 31% of the carbon tax is attributed to final consumption and 69% to generai capital income; these series are then related to household incomes and expenditures by income class. This results in a roughly proportional incidence of carbon taxes under the former series and a progressive incidence pattern under the latter series. Comparison with the Incidence of Personal and Corporate Income Taxes: The above analysis suggests that the regressivity of carbon taxes should be less of a concern in developing countries than in developed countries, This conclusion is further reinforced when one examines the incidence of personal income tax in a typical developing country. Personal income tax may not necessarily turn out to be a progressive element in the overall tax system, given both tax evasion and urban-rural migration effects, and their significance in lower to middle income countries. With respect to tax evasion, Shah and Whalley (1991) argue that, if the bribe rate is high and tax compliance low, the redistributive impact of the bribe system is likely to dominate the direct redistributive effects of income taxes. The relevant issue then is who receives the bribes. If public service is dominated by a seniority system, then high officials with higher income and wealth receive a large portion (or the majority) of the bribe, along with professionals (accountants) who often act as "middlemen" in this process. Increasing income tax can thus trigger a reverse distributional process from middle class businessmen and others to wealthy elites, an entirely opposite conclusion to that commonly reached. Thus tax evasion either reduces or offsets the progressivity of the tax system. The perceived progressivity of personal income tax is further clouded by the operation of the Harris-Todaro effect. In developing countries, personal income tax is imposed on urban sector incomes only. Under such circumstances, if expected wages are equalized across modern and traditional sectors through rural-urban migration effects, some of the burden of the (urban) tax is shifted to the rural sector through intersectoral wage effects. Thus, rural workers, although they face no legal liability to pay the tax, bear part of the burden of the tax through reduced wages. The potential importance of this effect is illustrated by Shah and Whalley (1991) using 1984-85 data for Pakistan. They 11 find that incorporation of the Harris-Todaro effect in incidence calculations clouds the progressivity of the personal income tax in Pakistan (see Table 3.3). Shah and Whalley (1991) also present calculations establishing the progressivity of corporate income taxes that take into account complications introduced by foreign and public ownership of the corporate sector in Pakistan. The above analysis suggests that concerns over the regressivity of carbon taxes may be over-stated. If the lowest income group is protected from the regressive impact of carbon taxes by direct subsidies or alternate measures, then the regressivity of carbon taxes may not pose a serious policy concern. Further, if carbon taxes are used to reduce personal income taxes, traditional concerns that such a tax change would represent a move to a less progressive tax structure are not fully justified. Thus, a commonly perceived and widely accepted case against carbon taxes, based on equity grounds, does not hold up under a closer scrutiny. 3.3 Efficiency Costs of Carbon Taxes By design, carbon taxes distort production, investment and consumption decisions and thereby internalize the social costs of global and local externalities. For every dollar of carbon tax revenues raised, consumers lose more than a dollar in direct and indirect costs. It is the indirect or hidden costs of carbon taxes relative to other forms of taxation that are of interest to policy makers. The literature commonly refers to these costs as marginal welfare costs of taxation. In evaluating the potential of carbon taxes, one needs to determine what will be the impact on economic efficiency if the same revenues were to be raised by carbon taxes rather than by existing (and distortionary) taxes on income. The empirical literature on this question is regrettably sparse. Poterba (1991), for example, provides estimates of average and marginal deadweight loss associated with carbon taxes relative to a no-tax scenario. Such calculations are interesting, yet, as the following analysis demonstrates, pre-existing taxes have a major bearing on welfare costs. Further, it is the differential (relative to other taxes), rather than the absolute incidence of carbon taxes, that offers useful policy insights. Goulder (in progress) is pursuing this line of inquiry for the U.S. using a computable general equilibrium model. Browning (1987) has argued that a properly specified partial equilibrium model of taxation's welfare costs offers superior insights on the measurement of welfare costs since, in such an analysis, the contribution made by key parameters to the final estimate remains transparent, whereas it is obscured in CGE models. He further demonstrates that almost aUl the differences in welfare costs of taxation for the US can be traced to different assumptions regarding key parameters, rather than differences in the nature of models (i.e. partial vs general equilibrium). In the following, two measures for the differential costs of carbon taxation and a measure for the absolute burden of carbon taxes are presented. All these measures explicitly recognize existing taxes. Derivations of these expressions are laid out more fully in Appendix A. Case (a): Welfare Costs Under a Revenue Neutral Change That displaces Equal yield Personal Income Taxes by a $10/ton Carbon Tax. An evaluation of the welfare costs of carbon taxation is carried out here by using a frequently employed concept of applied welfare economics known as the Hicksian compensating variations. According to this measure, welfare loss is defined as the additional income required to maintain the consumer's original utility level, given 12 (a) lncidence of Peronal Incoe Taxes in Pakistan under AlternatIve Approaches (tax a a percentae of total Inom) ) Form of income tax shift to rural sector No ptrsonal income tax Reduced wages for low-income nnalhboutbo d in rural sector rural households Reduced rural uwges overall i income (rupees) Urban Rural Total Urban Rural Total Urban Rural Total Under 7,200 0 0 0 0 0.74 O.S8 0 0.54 0.6t2 7,200-8,400 0 0 0 0 0.83 0.63 0 0.60 0.45 2,400-9600 0 0 0 0 0.88 0.63 0 0.64 0.46 9,600-12,000 0 0 0 0 0.73 0.46 0 0.52 0.34 t 12,000-18,000 0.02 0 0.01 0.01 O.S7 0.3S 0.01 0.41 0.2S 28,000-24,000 0.04 0 0.02 0.02 0.70 0.38 0.02 0.32 0.18 24,000-30,000 0.02 0 0.02 0.01 0.01 0.01 0.01 0.29 0.13 30,000-36,000 0.20 0 0.13 0.09 0.02 0.06 0.09 0.26 0.16 36,000-42,000 0.22 0 0.16 0.10 0.02 0.07 0.10 0.31 0.17 o. 42.00048,000 0.40 0 0.29 0.18 0.03 0.13 0.18 0.19 0.18 48,000-S4,C"0 0.77 0 O.J0 0.3S 0.01 0.23 0.3S 0.18 0.29 j AboveS4,000 1.33 0 1.04 0.61 0.11 0.47 0.61 0.13 0.48 Nota: Calculations ie no pmonal income taX m rural sector are based on actual tax collecidons by income class as reponed in Pakistan, Governmeat of (19835. All figures Irom this survey are adjusted to bring the total in line with data firm Pakistan, Government of (1988). Income tax collections on household incm derived from urban sources or from graduated surcharges on land revenue are effectively erto. -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ t [~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~f (b) InclId of Corporate Taxes In Pakistan under Alternative Approache (tax as a percentae of total Inca) Tax incidence Income atgorydsect to excluding taxes Annualhouseold anthex burden paidbystate capitland Capital andforeign income (rpees) Capital consumption and labor enterprises Under 7,200 1.18 1.71 I.S6 0.8S 1 7,200-8,400 1.06 1.SS 1.64 0.77 8,400-9,600 1.04 3.S3 1.70 0.76 * 9,600-12,000 1.26 1.62 1.69 0.91 12,000-18,000 1.46 1.70 1.69 1.06 _0% 18,000-24,000 1.70 1.79 1.68 1.24 24,000-30,000 1.68 1.76 1.69 1.22 _ 30,000-36,000 1.7S 1.78 1.68 1.28 36,000-42,000 1.77 1.78 1.66 1.29 42,000-48,000 1.81 1.79 1.6S 1.32 48,000-S4,000 1.89 1.76 1.63 1.34 Above S4,000 2.01 1.74 1.64 1.46 the vector of new consumer and producer prices resulting from the policy change. Thus it is the additional income that would make the consumer indifferent to the new vector of consumer prices. A Taylor-series approximation of the expenditure function, yields the following expression for the welfare cost of the tax system under the equal yield scenario mentioned above. LN=LL/=-1e E(T1)2 - (T1+T2)21p 2 - Tr + T2X) PiW, 2 Pi W1(1) 1 [ ( (TH) 2 - (T1 + T2H) 2 + zeRwi 1 2 W H. 2 W~~2 J1H where '6xP = own price elasticity of fossil fuel demand ;=- = cross price elasticity of fossil fuel demand with respect to after tax wages. Pc = elasticity of labor supply with respect to prices of fossil fuels. Cnw = elasticity of labor supply with respect to after tax wages. Pi>l= composite price of fossil fuels before carbon tax. Xi1 = quantity of annual consumption of fossil fuels before carbon tax. W1 = after tax hourly wages before revenue neutral labor income tax change. H1 = manhours of labor per year. 8 = share of fossil fuel expenditures to total expenditures. '1 = pre-existing unit taxes on fossil fuels. T2 = unit carbon tax (US$10/ton). TH I = pre-existing labor income taxes per manhour. T2 = reduction in per manhour labor income taxes. The first term in the expression above captures the direct effect of higher fossil fuel prices on fossil fuel consumption. The two middle terms are the indirect effects (cross effects) of higher 14 after tax wages on fossil fuel consumption and higher fossil fuel prices (lower real wages) on labor supply. The fourth term captures the direct effect of higher after tax wages on labor supply. The key parameters needed for the evaluation of this expression are: hours worked per year; current labor income tax rate; prices of energy products; quantity of energy consumption; current tax rate on energy; carbon tax rate (per unit of energy); elasticity of labor supply; and elasticity of energy demand. The data required to calculte these parameters for India, Indonesia, Pakistan, USA and Japan were collected from a variety of sources. Table 3.4 presents data on carbon emissions, carbon prices and energy taxes for the sample countries and Table 3.5 reports a summary of results on welfare effects based on the above model. These calculations suggests that replacement of personal income tax by an equal yield $10/ton carbon tax represents a welfare deteriorating proposition in the sample countries. Estimates of the welfare loss (compensating variations) range from a low of 1.5 cents per dollar of carbon tax revenues in Indonesia to a high of 17.5 cents per dollar in Pakistan. On economic efficiency considerations alone, therefo.e, carbon taxes cannot be supported as a replacement for personal income taxes. The difference in the welfare costs of a US$10 carbon tax arise primarily from variations in elasticity values (quite similar for our sample countries), pre-existing fossil fuel taxes, labor income taxes, carbon prices (i.e., market value of total fossil fuel consumption divided by carbon emissions) and energy price changes from the carbon tax. The price of carbon, a key parameter in the welfare cost calculations, is a function not just of fossil fuel prices, but also of the types and mix of fossil fuels consumed. A country that is a large consumer of coal will have a low price of carbon relative to a country that is a large consumer of natural gas or oil, even if the latter has the same level of fossil fuel prices. The relatively low welfare loss indicated for Indonesia is primarily attributable to lower levels of energy taxation in Indonesia, and the relatively large loss for Pakistan is due to high pre-existing energy taxes. In the case of Japan, the welfare loss is substantially lower than for Pakistan despite high pre-existing energy taxes. This results from the high price of carbon in Japan, which implies the percentage increase in energy prices due to the US$10 carbon tax will be low. The welfare loss for India compares well with that for the U.S., even though pre-existing taxes in India are much lower (see Tables 3.4 and 3.5). This is because the price of carbon in India is only half of the price in the U.S -- the result of India's high consumption of coal. The welfare gain associated with the direct effect of lower labor income tax on labor supply is very small for India, Indonesia and Pakistan (at most 0.5% of the total welfare loss), because labor income taxes are low relative to wage income in these countries. Higher labor income taxes make the equivalent effect substantially larger in the U.S and Japan (20% and 15% respectively). For the first three countries, the indirect effects are small but positive, which indicates that the positive effect of higher real wages on energy consumption that results from the lowering of labor income taxes, dominates the negative effect of higher energy prices on labor supply. Again this is caused by low initial labor income taxes. In absolute terms, the indirect effects are negative and larger for the U.S. and Japan. This is because the negative effect of higher energy prices on labor supply dominates the positive effects on energy consumption of higher real wages associated with income tax reductions, and because the initial effective taxation of labor income is higher in the U.S. and Japan than in developing countries. These results imply that analyses which ignore pre-existing taxes will be in error, and could consequently result in possibly quite misleading policy advice. The difference in measured 15 Table 3.4 Carbon Emissions, Carbon Prices and Energy Taxes in Selected Countries, 1987 Country Carbon Emissions Carbon Price Energy Taxes (Million tons) (S/ton) (S/ton) India 148.2 117 10.69 Indonesia 26.6 200 0.00 Pakstan 13.2 253 65.13 USA 1246.1 198 26.64 Japan 237.1 538 104.80 Sources: Carbon emissions - World Resources Institute (1990) Carbon price - Authors' calouations based on data from Asia Development Bank and Energy Information Administration Energy taxes - Authors' caloulations based on data from International Energy Agency. 16 Table 3.5 Summary of Welfare Effects of a $10/ton Carbon Tax, 1987 Carbon Tax Welfare Loss (-) or Gain (+) Revenues Million Million % of Carbon % of Total % GDP US$ USS Tax Revenues Revenues A. Revenue Neutral Change by Equal Yield Reduction in Personal Income Tax India 1482 -129 -8.7 -0.39 -0.06 Indonesia 266 -4 -1.5 -0.03 -0.005 Pakistan 132 -23 -17.5 -0.39 -0.07 USA 12461 -1049 -8.4 -0.11 -0.02 Japan 2371 -269 -11.4 -0.07 -0.008 B. Revenue Neutral Change by Equal Yield Reductions in Corporate Income Tax India 1482 +250 +16.9 +0.8 +0.11 Indonesia 266 +23 +8.7 +0.2 +0.03 Pakistan 132 +12 +9.0 +0.2 +0.04 USA 12461 -773 -6.2 -0.08 -0.017 Japan 2371 +213 +9.0 +0.06 +0.007 C. Raising Additional Revenues with NO Change in Existing Taxes India 1482 -130 -8.8 -0.40 -0.06 Indonesia 266 -4 -1.5 -0.03 -0.005 Pakistan 132 -23 -17.7 -0.40 -0.07 USA 12461 -1269 -10.2 -0.14 -0.03 Japan 2371 -291 -12.3 -0.08 -0.009 D. Raising Additional Revenues with NO Change in Existing Taxes but Accounting for Subsidies India 1482 0 0 0 0 Indonesia 266 +1 +0.4 +0.01 +0.005 Palistan 132 -23 -17.7 -0.40 -0.07 USA 12461 -1269 -10.2 -0.14 -0.03 Japan 2371 -291 -12.3 -0.08 -0.009 Source: Calculations based on the models presented in Appendix A. 17 welfare costs can be substgntial if pre-existing taxes are high, as is the case for Pakistan. If these pre-existing taxes were to be ignored, one would obtain for Pakistan fairly low estimates for the welfare costs of carbon taxes, similar to those for Indonesia. For India, the welfare costs of carbon taxes in a no tax case scenario, would then be twice the level of Pakistan and Indonesia (since, due to the Indian use of inexpensive coal, carbon prices in India are nearly half of those in Pakistan and Indonesia.) Case B: Revenue Neutr4l Introduction of a $10/Ton Carbon tax by Equal Yield Reductions in the Corporate Income Tax. Feldstein's (1978) model is adapted to derive the following expression for the welfare costs of taxation (see Appendix A, Case B for details): =N - 1 [ (Tx) 2- (TX+Tx) 2 - - (pi) I P TXR_ (X+TX) ( TJR +r 2 (2 Jp-'T Tl2 TR 1 Illar/rT+1) [ ~1 T1 (1+) 2I) ]p fR1 (2) P1P~ 'TRa 2) 2( R+TR) 2 (p R) 2 where 71sr = elasticity of corporate savings with respect to after tax rate of return. ri = after corporate tax rate of return on corporate savings. T = number of years from time of savings to dis-saving. cl = marginal propensity to save. Ti1 = pre-existing unit tax on consumption in the period of dis-saving; i.e. unit tax on return on corporate savings. R T2 = reduction in unit tax on return on corporate savings. R Pa = after corporate tax discounted price of consumption in period of dissaving, i.e. after tax price of savings. Ri = savings in real terms such that p1RR, = nominal after tax value of savings. The first term portrays welfare loss associated with a $10/ton carbon tax, and is equivalent to the corresponding term in Case A. The second term captures the interaction of reductions in corporate income taxes and a simultaneous increase in carbon taxes. An increase in the after tax price of energy products is likely to affect the consumption of energy products, and a reduction in after tax return on savings is likely to affect savings decisions. This term could 18 be either positive or negative. The third term represents the welfare gain associated with a reduction in corporate income taxes. Corporate income may be considered as a return on savings, i.e. on a firm's total assets or shareholders' equity. Thus the third term captures the welfare effects of changes in after tax rate of return on savings in the corporate sector. Corporate income taxes induce intertemporal inefficiencies by reducing savings and increasing current consumption. Key parameters needed for the evaluation of this expression include: energy and retirement (future) consumption expenditures and prices, taxes on energy and retirement consumption, savings, marginal propensity to save out of exogenous income, uncompensated elasticity of savings with respect to after tax rate of return, and price elasticity of energy demand. These parameter values are obtained from a variety of sources. The model's results, presented in Table 3.5, suggest that, with the major exception of the U.S., an equal yield introduction of carbon taxes in part replacement of corporate income tax would uniformly represent a welfare-improving proposition for the sample countries. The estimated net welfare gain varies from a high of 0.11% of GDP for India, to a low of 0.007% of GDP for Japan. These positive net welfare effects lend support to the widely-supported view that corporate income taxes are far more distortionary than labor income taxes. For the U.S., the revenue-neutral introduction of a $10/ton carbon tax to replace corporate tax revenues is, in contrast to the above, a welfare-deteriorating proposition. The welfare loss is estimated to equal 6.2% of carbon tax revenues or 0.017% of GDP. The effect is due to lower marginal taxation of corporate income in the USA in compari3on with other sample countries. Case C. Raising Additional Revenues From Carbon Taxes With No Change in &isting Taxes. The following expression for the evaluation of net welfare captures the direct effect of carbon taxes on energy demand through price increases, and also their indirect effect throu-gh reduced real wages -- the latter being associated with an increase in consumption taxation.' r g 1 ( T1X) 2_TrX+rX 2 2 (P) pX1 (3) - ew I P W W1H1O The key elasticity parameters required for the evaluation of the above expression are the demand elasticities for fossil fuels and supply elasticity for labor. The results presented in Table 3.5 suggest that although the welfare costs of carbon taxes are significant, they represent only a small fraction of carbon tax revenues. Estimates for the sample countries range from a low of 1.5 cents per dollar for Indonesia (0.005% of GDP), to a high of 17.7 cents per dollar for Pakistan (0.07% of GDP). The welfare losses for India, Indonesia and Pakistan are only slightly higher than those obtained in case A. This is because, given very ineffective pre-existing labor 'For a formal derivation, see Appendix A, Case C. 19 income taxes and substantial tax evasion, the direct welfare effect of labor income tax reductions is very small for these countries. The difference in the two cases is larger for the U.S and Japan because of higher pre-existing labor income taxes and levels of tax compliance. Poterba (1991) finds a much lower welfare loss for the US (average welfare costs of 3 cents per dollar of carbon tax revenues, or about 0.01 % of GDP) in the revenue increase scenario by assuming no pre-existing taxes and no wage effects from carbon taxes. Thus levels of pre-existing taxes (on energy, income etc.) are critical in the estimation of the overall welfare effects associated with tax changes. Calculations that ignore these effects will understate the welfare cost of tax policy changes. Case D: Raising Additional Revenues From Carbon Taxes With No Change in Existing Taxes but Accounting for Subsidies. The efficiency costs of carbon taxes will be over-stated if, as in Cases A through C, subsidies are ignored. An efficient energy pricing policy calls for price to equal long run marginal cost (in the case of no externalities). Thus it is interesting to re-evaluate this welfare calculation by recognizing existing subsidies (Larsen and Shah 1992). For the sake of simplicity, only the welfare cost of the carbon tax's direct effect on fossil fuel consumption is calculated, and the indirect effect on labor supply of higher fossil fuel prizes is ignored. This is justified because the indirect effect on labor supply is less than 1% of total welfare costs. In order to calculate the welfare cost, petroleum products, natural gas and coal are considered separately and the same own price elasticity of demand is applied to all product groups. Furthermore, the welfare calculation ignores the substitution effect between coal and petroleum products in cases A-C, thus overstating true welfare costs. Significant fossil fuel subsidies exist in India and Indonesia. The price of coal in India was only 85% of long run marginal cost in 1990 (Bates and Moore, 1991), implying a 15% subsidy. By (conservatively) assuming a similar level of subsidy in 1987, the year used here for welfare calculations, a US$10 carbon tax leads to an approximately 26% increase in the price of coal at 1987 prices. Thus a large proportion of the tax acts to remove the subsidy and should be considered a welfare gain. The welfare cost of the carbon tax on petroleum products and natural gas is estimated to be equal to the welfare gain of the subsidy removal on coal. The overall welfare effect of a US$10 carbon tax is therefore approximately zero, rather than the - 8.8% of carbon tax revenues in Case C. Similarly, petroleum products in Indonesia are priced significantly below world prices -- approximately 35% lower in 1987. Following the same approach as for India, the carbon tax on petroleum products in Indonesia represents a welfare gain, although it is too small to eliminate the subsidies completely. The welfare gain is larger than the welfare costs of the carbon tax on coal and natural gas. Thus, in comparison with Case C's welfare loss of -1.5%, the net effect is a small welfare gain of 0.4% of carbon tax revenues. This section illustrates not only that are pre-existing taxes critical in estimating the welfare effects of carbon taxes, but that many subsidies are also. Calculations that ignore subsidies will over-state the welfare costs of tax policy changes. In conclusion, the case for carbon taxes on efficiency considerations alone depends on whether they are introduced in a revenue-neutral manner, whether they replace corporate income taxes, and whether fossil fuel subsidies exist. According to the calculations presented here, such taxes do not fare so well against personal income taxes, at least for countries with pre-existing energy taxes and no subsidies. Clearly, however, an overall assessment of carbon taxes must 20 therefore consider their impact on greenhouse gases and local pollutants, as well as on industrial performance and economic growth. These issues are taken up next. 4.0 The Impact of Carbon Taxes on Greenhouse Gases and Local Pollutants Through their impact on aggregate use and composition of fossil fuel consumption, carbon taxes may reduce the emissions of local and regional pollutants such as nitrous oxides (NOx), carbon monoxides (CO), particulates (PM) and sulphur dioxides (SO2) as well as carbon emissions. This section deals with the impact of carbon taxes on NOx, SQ and PM emissions. These extent of these latter three emission types depend on technology, combustion processes and sulphur content of fossil fuels; emission coefficients therefore vary greatly across sectors and countries. Tne data on emissions are derived here from available sectoral emission coefficients and sectoral fossil fuel consumption (OECD 1989, and Radian Corporation 1990). Table 4.1 illustrates the impact of a US$ 10 carbon tax on fossil fuel prices, and on CQ, SO2, NOx and PM emissions for selected countries. The impact of the carbon tax on C02, SO2, NOx and PM depends on the percentage increase in the end-user price of each fuel, in addition to the price elasticity of demand and emission coefficients. It is calculated as follows: : = ;ji elij 8j = Eij e'ij Qij eij apij/pu (4) where: Z is tons of reductions in CC)2, SO2 or NOx; i are sectors; j are fuels (coal, natural gas and petroleum products); e',j is the emission coefficient of Z for fuel j in sector i; Qj is consumption of fuel j in sector i; eu is the own price elasticity for fuel j in sector i; and bpij/pij is the percentage increase in price of fuel j in sector i from the carbon tax. Interfuel substitutions are ignored. The elasticity of energy demand, being fairly similar across all the sample countries, does not contribute to the cross country differences in emission reductions. The price of coal shows the largest increases primarily because of the low price of coal per ton. The increases for petroleum products and natural gas are only marginal in comparison because of their much higher current prices per ton. India shows the highest estimated emission reductions principally because coal is the predominant fossil fuel in consumption; it experiences relatively large reductions due to the high price increase induced by the carbon tax. Reductions are lowest in Japan because of high pre-existing energy prices that induce very low price increases from the carbon tax and thus low reductions in fossil fuel consumption. SQ emission reductions are highest in Pakistan because most such emissions are from high sulphur (5-6%) coal. SO2 emission reductions are also quite high in the United States because of the large share of coal in consumption. Because of low coal use, Indonesia experiences relatively modest emission reductions. In all sample countries, percentage PM reductions tend to follow percentage reductions in the other pollutants. A benefit-cost analysis of a US$10 carbon tax can now be made by comparing the welfare losses (rable 3.5) of a revenue-increasing carbon sax (with no reductions in either labor 21 p Pakistan Indonesia India United States Japan Fossil fuel consumption (milion local currency) 58209 8793837 222744 246502 15759000 Carbon (C) emissions (million tons) 13.2 26.6 148.2 1246.1 237.1I Price of carbon (per ton): Local Currency 4409 330595 1503 198 66465 USS 253 200 117 198 538 Energy Taxes (USS/on of carbon) 65.13 0.0 10.69 26.64 104.80 s Carbon tax (USS/ton) 10 10 10 10 10 Carbon tax (local currency/ton) 174 16500 129 10 1235 Elasticity of enerW demand -0.64 -0.6 -0.651 -0.6 -0.55 Price increase (from carbon tax) of coal 37.8% 17.5% 26.2% 18.3% 8 7% pArwleum products 3.2% 5.8% 2.3% 3.4% 0.15% natural gas 2.6% 4.4% 3.0% 4.3% 1.4% 0 Emissions of (000 tons) PM 44 87 1192 6478 463 S02 321 337 2207 17900 1600 NOX 203 434 2090 17400 1400 Emission reductions (%) C -4.5% -3.9% -13.3% -5.3% -1.6% ' PM -11.6% -5.0% -15.3% -7.8% -0.6% 0 S02 -19.1% -4.6% -15.9% -10.0% -2.3% NOX -3.8% -3.8% -11.9% -5.6% -1.2% (1) Welfare cost of a USS 10 per ton carbon tax ad (revenue increasing tax) million USS (Table 3.5) -23 -4 -130 -1270 -292 o (2) Cost of carbon (C) reductions (USS/lon) 38.7 3.9 6.6 13.8 78.9 0 (1) divided by tons of C reductions (3) Price level (GDP1Penn GDP 1987) 0.23 0.35 0.34 1 1.55 (4) Benefit-cost ratio* High (SO2+NOX+PM) 1.8 17.9 9.5 11.2 1.3 Medium (SO2+NOXt PM) 1.6 12.9 7.5 8.7 1.0 Low (SO2+NOX+PM) 0.5 2.2 1.9 2.1 0.2 * 'High' is based on Glomsrod et al (1990); 'Medium' is based on Bernow and Marron (1990); 'Low' is based on EPA/Energy and Resource Consultants, Inc. referenced in Repetto (1990). Source: Authors' calculations Table 4.2 Marginal Benefits of NOx, SO2 and PM Reductions (US $/ton) NOx S02 PM Glomsrod et al. * ("High") 10300 1400 3300 Bernow and Marron ("Medium") 6500 1500 4000 EPA/Energy and Resource Consultants ("Low") 230 637 2550 * The first study is for Nonvay and the two last for the United States. Source: Glomsrod et al (1990), Bernow and Mafron (1990),'Repeuto (1990). income taxes or corporate income tax) with the benefits of ernission reductions. Welfare cost calculations are for the case which does not account for subsidies. Thus welfare costs are substantially overstated for both India and Indonesia. Benefits are estimated given only S02, NOx and PM emission reductions; no attempt is made to estimate the benefits of reductions in emissions of C02, CO, lead and ground level ozone. The monetary value of emission reductions for any of these gases will be highly uncertain, in part because the damage emissions cause depends on: the aggregate level of emissions, climatic and topographic conditions, population density around emission sources and on concentration levels of the pollutant. The main monetary benefits per reduced ton of S02, NOx and PM emissions, come from improvements in health and reduced corrosion (see Table 4.2, for results from three independent studies). Glomsrod et al (1990) and Bernow and Marron (1990) report the highest estimates based on their studies for Norway and the United States, respectively. EPA/Energy and Resource Consultants Inc. report (for the United States) significantly lower benefits, in particular for NOx. This low benefit estimate for NOx may result from excluding chronic health effects. Benefit figures are adjusted by Penn GDP relative Purchasing Power Parity indices (Summers and Heston, 1991) for each sample country, thereby allowing more meaningful cross-country comparisons. Note that this procedure assumes a degree of transferability for different countries' externality measures that is unlikely to be satisfied in practice; estimates of such measures are therefore likely to be crude at best. Notwithstanding the above caveat, the comparison of costs and benefits (Table 4.1) suggests that, on local environmental grounds alone, Indonesia, India and the United States can benefit substantially from a carbon tax. Benefits exceed costs by a ratio of more than 7 in two cases, and approximately 2 in the case of the lowest benefit estimates. In the case of Pakistan and Japan, because of high pre-existing energy taxes and thus high a welfare cost for carbon taxes, the benefit-cost ratio is significantly lower, although still greater than one. 23 It is important to note that, although the monetary benefits of emission reductions are uncertain, there emission reductions have additional benefits that are not accounted for here as already mentioned. Furthermore, welfare losses are based on the worst-case scenario of a revenue-increasing carbon tax not compensated for by a reduction in other taxes. Last, but not least, significant energy subsidies in India and Indonesia are not incorporated in the welfare calculations, which consequently overstate welfare losses. Note also that these benefit-cost ratios do not depend on the price elasticity of demand for fossil fuels, which is assumed identical for each fuel. Both the welfare costs of carbon taxes and the quantity of emission reductions are proportional to that elasticity parameter, which is therefore canceled out in the ratio of benefits and costs. The latter depends primarily on pre- existing taxes on fossil fuels (which affects welfare costs) and on the valuation of emission reductions of S02 and NOx in both relative and absolute terms. Furthermore, the calculations presented here do not attempt to identify least-cost policies for local pollutant reduction. They merely quantify various additional benefits from carbon taxes that are frequently ignored in the literature. One means of accounting for the non-uniformity of emission externality costs across countries is to adjust the benefits of emission reductions for variations in population density and rural/urban population ratio. Here, an equal weight is applied to population density and urbanization. In consequence, benefits are larger by an average factor of two for Pakistan, Indonesia and India. Thus the benefit-cost ratio is larger than one for Pakistan even when lowest benefit estimates are used. For Japan, benefits are as much as twelve times higher. This is the result of a very high population density, which brings the ratio to 2.4 in the case of lowest benefit estimates and to as much as 14 in the case of highest benefit estimates. In this circumstance, Japan would benefit even more from a carbon tax than the United States. A cost analysis of carbon reductions is also illustrated in Table 4.1. The cost of carbon reductions is stated in terms of the welfare costs of a revenue-increasing US $10/ton carbon tax, divided by tons of carbon reductions. The large cost differences across countries are caused mostly by differences in pre-existing energy taxes (high pre-existing energy taxes implying high welfare costs) and percentage carbon emission reductions. To illustrate this point, the cost of carbon reductions may be stated as follows: C = (W/R )( R/E) = (W/R )(t*CJ/E) = W/E (5) where: C is the cost per ton of carbon reductions; W is the total welfare cost of the carbon tax; R is total carbon tax revenues; E is tons of carbon emission reductions; t is the carbon tax rate (US $10/ton); and Ce is the total tons of carbon emissions (thus CJIE is the reciprocal of percentage carbon emission reductions). Equation (5) reveals that C is high if welfare cost per tax revenue dollar (W/R) is high (Table 3.5, C), and/or if percentage carbon emission reduction is low (Table 2.2). The cost per ton of carbon emission reduction is lowest for Indonesia, even though percentage emission reduction is low. This is because of virtually non-existent energy taxes, which imply very low welfare costs per tax revenue dollar. Cost per ton is highest in Japan because of the combination of high welfare costs per tax revenue dollar and very low percentage emission reductions. The results in Table 4.1 also suggest that optimal carbon taxes 24 are not uniform across countries because of different levels of pre-existing energy taxes and impact on local pollutants. The preceding analyses of fossil fuel consumption and emission reductions considers only aggregate fuel reductions and not interfuel substitution. But since a carbon tax may induce significant interfuel substitutions, it is to be expected that the estimated emission reductions in Table 4.1 are overstated, given own-price elasticities. However, allowing for interfuel substitution would reduce the welfare costs of the carbon tax, such that the overall ratio of benefits to costs would most probably be only marginally affected. In conclusion, the above analysis suggests that a carbon tax has significant benefits in terms of both local pollutant and CO2 reductions. A monetary benefit-cost analysis indicates that, for countries with low or non-existent energy taxes, a carbon tax can be justified on local environmental grounds alone, even ignoring its benefits from a public finance viewpoint. 5.0 Carbon Taxes, Industrial Performance and Economic Growth Carbon taxes by changing the relative prices of inputs can impact on the production, financing and investment decisions of firms. In this section, the Bernstein-Shah dynamic model of production structure (forthcoming) is used to examine the impact of carbon taxes on the economic performance of Pakistan's apparel and leather products industries (1966-84). Several features of this dynamic model are noteworthy. The costs of adjustment are treated as internal to the firm and are explicitly modelled. These capital adjustment costs imply that capital input does not necessarily attain its long-run desired level within any one contemporaneous period. The model formulation allows for estimation of this speed of adjustment. Investment in capital results in some foregone output in the short run. The model distinguishes short run, intermediate run and long run effects of tax policy initiatives. These effects are influenced by the varying degree of capital adjustment. The model also treats the determination of output supplies, variable and quasi-fixed input demands simultaneously. Thus both the direct and indirect effects of tax policy changes are captured in the model. Moreover, the dynamic nature of the model allows for direct and indirect effects to be estimated in all three runs of production. In addition to the explicit modelling of adjustment costs, the Bernstein-Shah model incorporates several features of producer behavior which are absent from the Jorgenson-Wilcoxen framework. Output supply is endogenous and not solely a function of factor demand or of investment. Furthermore, product markets are not assumed purely competitive and the nature of firm interdependence, as measured by the conjectural elasticity parameter, governs the structure of product markets. Finally, the model recognizes financial capital market imperfections as firms are constrained by the rate of return that can be earned on their financial capital. Rates of return on equity and debt capital are treated as exogenous to firm's behavior, and cannot therefore be influenced by shareholders. Under such circumstances, the interest of owners is best served by maximizing the expected present value of the flow of funds to shareholders and bondholders. In other words, the firm's objective function is to maximize the expected present value of financial capital. The above mentioned product and financial market imperfections are germane to most developing countries. 25 Own Price Elasticities Carbon Tax Elasticities Imoact of a USStO/ton Carbon Emissions reductions Aggregate Aggregate Apparel & Aggregate TI Apparel Leather manufacturin Ava%e1_ Leather Appa Leather manufacturin Leather manufacturin ; ~~~~~~~Y -0.00081 -0.00098 -0.032X 40.039% 40.205% " r ~~~~~~L 40.00086 40.0W090 40.034% -0.036% 40.137% A ; ShortRun M -1.521 4.979 4.514 -0.00193 4.00133 -0.076% 4.0539 0.4829 -4.8% -7.6% Short Run K -0.00052 -0.00072 -0.021% 4.028% 0.499% Y 4.00111 -0.00145 4.044%9 4.057% L 4.00108 -0.00147 0.043% 4.058% '4 Kntcmediatc Run M -1.402 -1.135 -0.00178 4.00154 -0.070% -0.061% .4.9% X ~~~~~~~K 0.0008 40.00121 -0.034%C 40.048% 9~ ~ ~ ~~~~~ -0Y .00220 40.00272 -0.087E 40.10796 Intemiediate Run M, 1 4.00198 4.00317 4.079% 4.121% Long Run M -2.461 -2.879 - 4.00313 4.00392 -0.124% 4.155% -10.4% ii' K 0.00201 4.00255 -0.079% 4.101% I Notations: Y = Output L = Man Years worked M = Itermediate Inputs K = Capital stock Source: Model Resuts VA 'U; Table 5.2 Impact of a US $10 Carbon Tax on Manufacturing Value Added and Local Externalities PAKISTAN Paldstan AmmI La ther Industti PaldstAn A tem M=h%enu6jngi Short Intermediat Long Short rn nn run run Output effect (%) -0.035% -0.051% -0.098% -0.205% Output effect (in 000 USS) -102 -148 -284 -20900 Value added effect (in 000 USS) -19.12 273-3 Emission Reductions (%)* NOX 4.9% -5.0% -10.5% -7.4% S02 -4.7% -4.8% -10.2% -18.4% PM -4.8% -4.9% -10.3% -12.9% C02 -4.8% -4.9% -10.4% -7.6% Cost of C02 reductions - USS/ton (Ilos of value added divided 44.2 61.9 55.9 14.5 by tons of C02 reductions) Benefit-Cost Ratios associated with the impact of a US S10 carbon tax on value addod and local polluta na High 2.5 1.8 1.9 3.9 Medium 1.6 1.1 1.2 3.5 Low 0.09 0.06 0.07 1.1 *Emission reductions ae peretage of emissions from the Appami an Leather industrie or fiom total manufactraing industries. Includes sulfiur dioides (S02), nitrous oxides (NOx) and paticulate matten (PM). 'High" is based on Glomsrod et al (1990); 'Mediwnm is bued on Bernow and Marron (1990); 'Low" is based on EPA/IEnrgy and Resource Consultan, Inc. The ast study does not include chronic health effect of NOx emissions. Source: Model based calculations 27 Accounting for them, thus adds a sense of realism to the analysis of producer responses in these countries. The estimation model is characterized by an after tax rormalized shadow variable profit function, output supply and input demand and capital input demand equations. The model fits the data quite well. Furthermore, estimated parameters satisfy the conditions that the after tax shadow variable profit function be concave in capital and net investment and convex in prices. Table 5.1 provides estimates of carbon tax elasticities with respect to input demands and output supply in the short, intermediate and long runs. These tax elasticities are then used to calculate carbon tax effects at mean sample values. A $10/ton carbon tax on the apparel and leather industries leads to reductions in output and input demands in all periods, with the leather industry experiencing slightly higher reductions in output than the apparel industry. This difference is primarily attributable to the slightly higher energy intensity of the leather industry. Long run output impacts are (-)0.09 % for apparel and (-)O. 11 % for leather goods, both of which are higher than intermediate and short run impacts. Higher adverse effects in the long run arise because the model estimation suggests energy inputs serve as complements to both labor and capital in the two industries. To examine the same effects for manufacturing industries in Pakistan overall, a flexible accelerator type dynamic factor demand model developed by Shah and Baffes is implemented using time series data for the period 1956 to 1985. This model employs a flexible and non- restrictive technology and captures the short run divergence of fixed factors from their equilibrium values as well as the speed of such adjustment. Parameter estimates from the model suggest some pairwise substitutability among energy (materials) inputs and capital and labor. The model results suggest that the imposition of a $10 per ton carbon tax on Pakistani manufacturing industry will result in an output loss of 0.21 % in the short run (see Table 5. 1).2 The primary reason for larger output losses in aggregate manufacturing than in the apparel and leather industries is the substantial impact of the carbon tax on the price of coal. Coal is used primarily in industries other than apparel and leather. Thus energy prices for aggregate manufacturing increase substantially more than for the apparel and leather industries. A comparison of value added losses with the health benefits of reductions in local externalities throws some (albeit limited) light on the cost-benefit calculus of carbon taxes. Table 5.2 reports estimates of costs associated with carbon taxes, as well as benefits arising from a reduction in local externalities. Data limitations restrict the analysis to NOx, SO2 and particulate matter (PM) only. The dollar values on local externalities are based on the same three studies used in Section 4.0, adjusted for purchasing power parity. Benefit to cost ratios are higher for aggregate manufacturing than for the apparel and leather industries because relatively large emission reductions from reduced consumption of high sulfur coal more than offsets the higher loss of value added. Ratios are larger than one except in the case with the lowest benefit estimates for the apparel and leather industries. These tentative calculations suggest that losses of value added are offset by health benefits associated with NOx, SO2 and PM emission reductions, even if the reduction in global externalities associated with curtailing CO2 emissions is completely ignored. Table 5.2 also reports estimates for the average cost of 2 Jorgenson-_wlcoxen (1990) obtain -0.5% long run output effect for the U.S. for a $15.45/ton carbon tax 28 carbon reductions associated with a US $10/ton carbon tax in terms of US$/ton of carbon reductions, ignoring the benefits of reductions in local externalities. Calculations suggest that such costs are higher in the apparel and leather industries. The is primarily because total carbon emissions relative to value added in these two industries are much lower than in the overall manufacturing sector, while model results suggest that the elasticity of output or value added with respect to energy prices is only slightly lower. Thus losses of value added relative to carbon emission reductions are higher in the apparel and leather industries. 6.0 Tradeable Permits Tradeable emissions permits represent an alternative instrument that can ensure marginal costs of emission reductions are equalized across domestic sources and across countries. Given both perfectly competitive markets and certainty, permits are equivalent to emissions taxes (see Hoel 1990). Tradeable permits afford direct control over quantities of emissions as opposed to a carbon tax regime's indirect influence through prices. They are also easier to implement as an initial allocation of such permits reduces the resistance of existing emitters. Furthermore, tradeable permits in terms of their regulatory effects are more transparent to policy makers and administrators (see Oates and Portney, 1991). Tradeable permits have also been cited for their potential as a hedging instrument against risk and a vehicle for international technology transfer. Epstein and Gupta (1990) have argued that tradeable permits could serve as an instrument to reduce the risk of investing in backstop technology R&D. They argue that agents or nations that invest in R&D are exposed to a high probability of failure, although also to high profits in the event of success. If R&D investments turn out a successful technology that significantly reduces the costs of carbon emission reductions, the price of emissions permits will fall. If the investments yield no return the price of permits is expected to rise. This means that risk averse investors can purchase futures on emission permits3 as a hedging against risk. In this case, total investments in R&D can be expected to be higher than if a market for emission permits did not exist. One could further argue that carbon taxes would also induce higher levels of investment if tax revenues were pooled (fully, or in part) in an R&D fund or used to subsidize R&D. A closer analysis of the effectiveness of these alternatives seems appropriate given potential gains from the development of backstop technology. Emissions permits will induce international technology transfers if initial emissions allocations are such that industrialized countries will purchase emissions permits from developing countries. If this is the case, developing countries may purchase more energy efficient technology from industrialized countries until the marginal benefit is equal to the permit price. This transfer could potentially be quite substantial and significant for developing countries. Its magnitude depends on the costs of emission reductions and initial permit allocations (Larsen and Shah 1992b). If costs of emission reductions are high (after some smaller initial reductions) in industrialized countries, then developing countries will want to purchase more emissions permits from developing countries than if costs are low. This would imply larger revenue accumulations $ According to the New York Timnes, the Chicago Board of Trade will create a private market for trading in sulfur dioxide emission permits and forward contracts, and a futures market is also considered. 29 in developing countries which could be used to purchase more energy efficient technology. Technology transfers may turn out to be significant for developing countries because, in addition to reducing energy dependency, new capital embodies technological progress and thus contributes to increased total factor productivity. Total factor productivity gains are considered an important component for economic growth and improved international competitiveness. In practice, tradeable permits are subject to important limitations. These include: the "thinness" of permit markets, the presence of large buyers and sellers, and lack of any mechanism to deal with overshooting the mark. In the U.S., it is observed that the main reason the permit markets are not as well-functioning as envisioned is the "thinness" of the market, especially on the supply side, that is largely due to trading restrictions ano unclear definitions of property rights. When permits are infrequently traded, clear price signals are absent, thereby impairing the functioning of the permit system. On the other hand, a carbon tax is in itself a clear measure of the cost of emissions. To avoid ill-functioning permit markets, the number of potential traders should be sufficiently large. In the case of carbon emission permits, an insufficient number of traders may be avoided by integrating international (inter-country) and domestic markets. Market power is then eliminated and sufficient liquidity provided, especially if the market is open to outside parties as well as "emitters". In this case, any agent -- a producer or consumer -- obtains emission permits at a price quoted at trading boards, in much the same way as foreign exchange is traded and rates are quoted in international markets. There are alternative market arrangements, although an international (inter-country) market seems a minimum requirement because the costs of emission reduction can be expected to differ substantially across countries. Emission permits, traded internationally, allow marginal costs of reduction to be equalized across nations. Permits may be traded independently within nations so that marginal costs are equalized across domestic sources. It is also possible that permits will only be traded internationally and that carbon taxes will be used domestically. Alternatively, some countries may use emissions permits to reduce domestic emissions while other countries use taxes. In the latter case, there may be separate international and domestic permit markets. Any market arrangement that reduces the number of traders below that in a globally integrated market is exposed to the danger of market inefficiencies (market power, iltiquidity). However, the transactions costs of such markets may be too high to justify the establishment of a market that involves all "emitters" of carbon gases, from large industrial firms through to the individual household using fossil fuels. A carbon tax avoids these transactions costs. In global trading of permits, large countries can influence prices. For a large seller, it is optimal to have higher emissions than the level indicated by the marginal cost of reductions (the market price for quota); and the opposite holds true for a large buyer (see Hoel, 1990). A potential problem with permit markets is that the supply of permits is by no means guaranteed to be intertemporally fixed. New information about the costs of emission reductions and of global warming will induce policy makers to change the total supply. Furthermore, such changes cannot be preannounced at the initial time period since the changes are a function of the ncw information in future periods. New information is therefore similar to random shocks. This exposes permit holders to the risk of permit price changes that cannot be ignored. Two ways of getting around this problem are to establish a futures market, or to let permits expire at the end of each time period in order to issue the new supply at market determined prices. 30 Clearly, additional transactions costs will be unavoidable, thereby making tradeable permits less of an attractive instrument. It is not clear whether or not there will be a regional or global policy response to the greenhouse effect. In the event of such a response, the most talked about scenario is to set a target of a certain percentage global emissions reduction below their current (or some future) level, or to stabilize the current (or some future) global stock of emissions. The most frequently discussed target is a 20% reduction below current levels by a specific year, although a 50% reduction is considered necessary to stabilize the stock of global emissions at current levels (World Resources Institute 1990). What is the optimal policy instrument to achieve this objective? A carbon tax will result in some uncertainty about the magnitude of reductions but less uncertainty about the cost of reductions. Under a regime with tradeable permits the magnitude of emissicn reductions will be known, but there may be great uncertainty about the total cost of reductions. This is an important distinction between the two instruments in the case of global warming. Oates and Portney (1991) make this distinction when comparing a carbon tax with tradeable permits. If there is great uncertainty regarding the costs of emission reductions, a tax is preferred in order to avoid potentially large unexpected costs. (This is particularly important if the marginal costs of reduction are rising steeply after some initial reductions have been achieved.) However, if the costs of global warming are believed to be unacceptably high or there is a threshold effect, it becomes very important to limit total emissions to an upper bound. In this case, tradeable permits are preferred to a tax even though there will be great uncertainty regarding the costs of emission reductions. At this point in time, we do not know whether there is a threshold with respect to the stock of carbon emissions beyond which temperatures would rise exponentially. Furthermore, we know little about the economic costs and environmental consequences of global warming. Given present ignorance regarding the global warming phenomenon, one might currently argue for a carbon tax in order to limit unexpected costs of emission reductions. When, or if, future research reveals more about possible threshold effects and the costs of warming, tradeable permits may become the appropriate instrument. A global tradeable permit (or carbon tax) regime poses an additional problem in terms of initial permit allocations (or redistribution of tax revenues). Larsen and Shah (1992b) evaluate alternative allocations, such as allocations relative to GDP or population, and conclude that neither of the two are likely to induce participation from significant groups of countries. They propose an alternative allocation, based on willingness to pay for carbon reductions, that may induce broader participation in an international treaty. 7.0 Sunmarv and Conclusions This paper has evaluated the case for carbon taxes on national interest grounds. As a background to this discussion, it has also reviewed current energy pricing regimes in developing countries and their implications for greenhouse gas emissions (Larsen and Shah, 1992). The following conclusions emerge from the analysis: 31 * A global carbon tax raises difficult issues of tax administration, compliance and international resource transfer, and is therefore unlikely to be implemented in the near future. * National carbon taxes can raise significant amounts of revenue in a cost effective manner and, in developing countries, are not likely to have as regressive an impact as commonly perceived. Such 2 iax also fares quite well in efficiency terms if introduced in a revenue-neutral manner as a partial replacement for corporate income taxes. In general, the welfare costs of carbon taxes vary directly with the existing level of energy taxes and therefore a carbon tax should be the instrument of choice for countries with no or low levels of energy taxation, such as Indonesia and India. * A carbon tax also has significant benefits in terms of local pollutant reductions in addition to CO2 reductions. The cost-benefit analysis for selected countries presented in this paper suggests that countries with low or non-existent energy taxes can receive substantial net gains from a carbon tax, not just in efficiency terms, but on grounds of local environmental considerations alone. - A carbon tax of US $10/ton results in very small output losses for the Pakistani industries analyzed in this paper. The estimated effects are somewhat lower than comparable estimates for the U.S. obtained by Jorgenson-Wilcoxen (1990). The value added losses are, however, offset by the health benefits associated with reductions in NOx, SO2 a,nd particulate matter (PM) emissions, even if reductions of global externalities associated with the curtailment of CO% emissions are ignored. o Tradeable permits represent a preferred alternative to carbon taxes should there be a known critical threshold in the stock of carbon emissions beyond which temperatures would rise exponentially. Given our current lack of knowledge about the costs of carbon emission reductions and the threshold effect, a carbon tax appears to be a superior and more flexible instrument that avoids potentially large and unexpected costs. Thus, while a universal case for national carbon taxes cannot be made, even ignoring global externalities, such taxes make eminent sense for a large number of developing countries in terms of efficiency, equity and local environmental externality considerations. 32 APPENDIX: Measurement of Differential Welfare Costs of Carbon Taxes Welfare costs L of a tax system T, = (T11,TI2,...Tl) introduced at a non-distorted equilibrium with prices po = (p1,pO2,...pOn) is defined as the difference in the expenditure level E necessary to maintain a utility level U in the presence of T and the expenditure level required to sustain U in the absence of T, minus the tax revenues R: L (p,PO,) = E(p1,U) - E(po,U) - R (pj,1poU) (1) with P, = Po + T1. The expenditure funcdons can be approximated by a second order Taylor expansion in prices. Thus in general the welfare costs of taxes introduced at a non-distorted initial equilibrium is L =-l E E SIM Tl T1J (2) where S,4 = 6X1I/8pli, the cross-derivative of the compensated demand function and Ti is the unit tax of good i. In the presence of existing taxes, welfare costs of changes in the tax system is not simply L. An intermediate step becomes necessary (Feldstein 1978). Consider a revenue neutral tax policy change such that P2 = pi + T2, with T2 a vector of additional taxes. The total welfare costs of the tax system T1 + T2 is L'(p2,po0,) = E(p2,U) - E(po,U) - R(p2,pO,U) (3) or in general L = -lh E l: S21 (Tri+T29(T1+T2)) (4) The additional welfare costs of the revenue neutral tax change is LN = L - L = E(p,U) - E(p2,0) (5) since R(p2,po,U) = R(pj,pO,U) because of revenue neutrality. 33 Case A: Welfare Costs of Carbon Taxes That Displace Equal Yield Personal Income taxes. Consider the case of two goods (x, 1-H), where x is fossil fuels and 1-H is leisure (H is supply of labor). Prices of fossil fuels and leisure is P0 W0) in an initial non-distorted equilibrium. The welfare cost of pre-existing taxes on fossil fuels and labor income CrT, T,1), before introducing a carbon tax, is given by (2): L 1 lx T,x T,K - 1 OX ETx TH 18(1 (- 10 Tx H 2 apTlT 2 awTlr 2 a(pH)TX L ~~ 2 op; T1T11 I28w with T,H = WI - Wo < 0, T1x = p, - po > 0. Writing L with compensated elasticities, (5) becomes b~~~E I 4(T px - 2 exW( p-) (_Lw-)p1X , 1 e ( T1 ) 2p(.X1 e,-l) 2 pi WI 'I + P1 W _ 2 V(_L1 ) 2WIHl with the elasticities evaluated at (pi, xI) and (wl, H1). Suppose that carbon taxes are levied on fossil fuels T21= p2 - p, > O, in addition to existing taxes Tl1, and that labor income taxes are reduced in a revenue neutral manner T2K = W2 - W, > 0. The welfare cost of the tax systen C(1 + T2Z, T1" + T2M) is given by (4): 34 L= - e (T2+Ti)a - 12/X( Ti+.T?2 ( T2 2~~ p2 ~ 1 2 xv HT 2 1I N2 + ' 1 E + e' (+ 2)2wH2 2-2 X A W2 ) W with the elasticities now evaluated at (P2, X2) and (W2, H2). The change in welfare costs of the revenue neutral tax change is t_(TJ) (T3 4Ta)2]pX pi2 ex, TIT -TI (Tlx+ T2X) ( T1*V Ti.") 1,]p x1 2-6v P1wl (8) 2 PlElW 1 l t(TH)2- (T1 + TH)2JWH where O = share of energy in total consumption and by noting that I (T_+ T 2P)X2 = xe( Tp 2plX+ and similarly for the other elasticities. The indirect terms are multiplied by the expenditure share of fossil fuels, O, to account for the fact that in reality there are more goods than leisure and fossil fuels. The first term in (8) is 'lie direct effect on consumption of fosil fuels of higher fossil fuel prices. The last term is the direct effect on labor supply of higher after tax wages. The 35 two middle terms are the indirect effects (cross effects) on fossil fuel consumption of higher after tax wages and on labor supply of higher fossil fuel prices (lower real wages). L N > 0 would imply a welfare gain from the revenue neutral tax change. We would like to express the two indirect effects in terms of eNw which can be accomplished in two steps. The first step is to express the third term in (8) in terms of 'x by noting that Op8w 2(9) from the symmetry in the two-by-two matrix in the second order Taylor expansion of the expenditure function. The negative sign in (9) comes from the use of leisure as 1 - H. Given that aE = x and aE = H anv are the compensated demand functions, it follows that ax= aH aw Op which is (9), and tierefore We i H (10) Thus, the two indirect tems can be expressed as 36 - e,|,[ T - (Tl + T2 ) (T1' + T211) 1pwl *o (1 - e,~i (11)T T +T The second step is to express the compensated elasticity exw in terms of eII. Let U=f(x,l -H) (12) The total differential of (12) (letting au = 0 is afax + 8f a(1-H) = 0 (13) Ox 7( 1 -H) From the first order conditions of utility maximization 8f/Ox =p (14) aflaO(1-H) w By (13) and (14) and dividing through by dW: ax aH = 0 (15) This gives pxexh* = Hejw (16) To quantify LN, T2H is derived from the revenue neutrality condition aR = T1xX1 - T2RH1 = o (17) for small changes in the tax system. From (17) we get T2 - T2-H (18) H1 37 With (11), (16) and (18) we have LM 1 -2. (Tl )2 - (TI + T2) 2 (XT" TX + T2X) (Tf + T2X-L (Tl I * (T H1 2 ehW[ P,WI ]VIHM, (19) (T1H) 2 - (T2M + TX X1 ) 2 + I 1HG W 2 ] I A Note that the elasticity values applied to (19) are uncompensated elasticities rather than the theoretically correct compensated elasticities. The difference in terms of welfare cost is quite small (Willig 1976), approximately 10% with an income elasticity of 1 and 0 = 0.05 given our uncompensated elasticity values. This result may be derived from the Slutsky decompositions of the substitution and income effect. Thus welfare costs are slightly overstated here. Case B: Welfare Costs of Carbon Taxes That Displace Equal Yield Corporate Income Taxes. Coxporate income may be regarded as return on savings (Feldstein, 1978), i.e. on assets or sharholders equity. Consider the case of two goods (x,R) in a two period model where x is fossil fuel consumption in the first period and R is second penod consumption of first period savings, both in real terms. Prices of fossil fuels (x) and second period consumption (R) are (p.p03) in an initial non- distorted equilibrium. In the presence of existing unit taxes on fossil fuels and second perod consumption (T1,TIR), welfare costs are given by (20): 38 L ' 1 axTxTx 1 aX T7XTR - 1O2RTfTR (20) 1 OR "R"R With T1R = pIR - poR > O T1X = plx _ poX > 0. If ro is the rate of return on savings in the corporate sector, then po e -r0" is the discounted (current) price of consumption in period T+ 1 in the case of no tax on corporate income. Similarly, = e 1 is the after tax discounted price, with r, = (1 - t) r. and t is the corporate income tax rate. Thus corporate income taxes reduces the real value of period one savings since PI P > ° Writing L with compensated elasticities, (20) becomes .9r~ Tx TmR x _ ( T1 2plX1 w(p) R)P (21) TX R 2 p-R i(-p) (R)p1RR 2 p ( L)2P1 with the elasticities evaluated at (p1,xj) and (piR,RI) Suppose that carbon taxes are levied on fossil fuels T2= p2 - p1 > O in addition to existing taxes T11, and that corporate income taxes are reduced in a revenue neutral manner T2R = p2R plR < 0. The welfare costs of the tax system (T + T2x, T1R + T2R) is 39 L/=- 2/( 1Ta 2P2X2 ,/ R( 12) P TA+ R 2~ p2 2' -W 2 P2 (22) s Ti'+Ta ) ( T2 " TzR) pRR2 - 1 e 1 TlR+Tz ) 2pi?RR tax is LN ~ ~ a . -let(Ti')2 - (Tlx T2") 2j X LN ~ ~~2xR = L -T + 22 )~ (P1+)2R) 1 T~~~1P (23) a sl=lR -(Tlx + T2") (T1 + Ta) ]p11RR 0 (T ) - (pTR + T22R 2 , R( (P R) 2 2 where O = expenditure share of fossil fuels in total consumption and by noting that ( ( Ti_+_Ta") 2p=X_ e ~( T T + T2 ) 2 X1 and similarly for the other elasticities. The first term in (23) is the direct effect of higher fossil fuel prices on consumption of fossil fuels. The last term is the direct effect of lower prices on second period consumption. The lower taxc on corporate savings reduces the inter temporal inefficienlcy. The two middle 40 terms are the indirect effects (cross effects) on fossil fuel consumption of lower prices on second period consumption and on savings from higher prices on fossil fuels. L N > 0 would imply a welfare gain from the revenue neutral tax change. We would like to express the two indirect effects in terms of @A which can be accomplished in two steps. The first step is to express the third term in (23) in terms of6Ps by noting that 82E = 2E (24) ap 'lap apap R1 from the symmetry in the two-by-two matrix in the second order Taylor expansion of the expenditure function. Given that IE = X and - = i yp ap R are the compensated demand functions, it follows that Bx _ R A pR ap R yp which is (24), and therefore AR , e RtX (25) Thus, the two indirect terms can be expressed as 41 ERj [ TlXl (TI T2 )R +T T2 ) P1. (26) P1pf The second step is to express the compensated elasticity '> in terms of 'R60' Let U = f (x,R) (27) The total differential of (27) (lettng 8U=O) is afax + fR = O (28) From the first order conditions of utility maximization -f/Rx = p (29) By (28) and (29) and dividing through by OpR: i,;. jk pR RR - o (30) a,PR OP R This gives p Xe Rw =_ pit R 'ERpRi (31) By substituting (26) and (31) into (23): 42 L = _ 1 l (T1X)2 - (TX + X)2 R E (TI TIR -(lx + T2 ) (TJ + T2:) p3RR1 . (32) pipf --6ie Rs ( T12 - (TR + TR) 2 2 ~~~~(PiR) 2 It remains to express EAP in terms of the elasticity of savings with respect to the after tax rate of return for which elasticity alternatives are available. Note that P R = " .R + q (33) where n is the uncompensated elasticity and or is the marginal propensity to save out of exogenous income (Feldstein, 1978). Given that savings is S = p'R, we have R as . 'pR =a (P RR) p lsp 3pR S ap R S p + aR (pR)2 (34) S a8PR S By (33) and (34), R = Sp R - (1-a) (35) Recal that the discounted price of period T + I consumption isp = e with r the after tax rate of return on period 1 savings. Thus, 43 as R CIS PR 1 =f p 1$pR apR S ar S -Te-zT ar S -. R (36) as r 1 rT°181/rT because R = - Te7a ar. It follows that Rp = - (11,,/ZT + 1 - I ) (37) To quantify LN, T2R is derived from the revenue neutrality condition. With I being total tax revenues, I =T2xXl + T2 =O (38) for small changes in the tax system. Thus, $A= _ Ta.X TR A(39) With (32), (37) and (39), we have 2 ( (x) 2 xx Lg =_ 1er tT )_(T1 +T2)]pX T3TR- ( Tx+T )( -(118e/r/T+1-a) t RlPR FlJPIRIO (40) (TR) a_ R x " 44- Case C: Welfare Costs of a Revenue Enhancing Carbon Tax with No Change In Existing Taxes Consider the case of two goods (x, 1-H) as in Case A. The welfare cost LN may be derived directly from (19) by noting that with nu other changes in thc tax system than the carbon tax on fossil fuels, T2H=O i. e. T2XX- o in (19)) H1 Thus, L,N = - 1 (,) 2 vp1X1 (41) - eaT[ T1N W1H10 as the last term vanishes. The case of a revenue increasing carbon tax may alternatively be considered by recognizing the indirect effect on corporate savings instead of the indirect effect on labor supply. In this case L can be derived for (40) instead of (19). The first term will remain the same and the last term will vanish, but the indirect effect will in this case be unambiguously positive since TIR > 0. This is because higher prices on current period consumption induces a substitution to second period consumption, i.e. savings will increase. Thus the welfare loss will be slightly smaller than the direct effect on fossil fuel consumption, contrary to the case previously considered with indirect effect on labor supply. 45 Case D: Welfare Effects of a Revenue Enhancing Carbon Tax with No Change In Exlsitng taxes but Accounting for Subsidies The ' Sulation will only include the direct effect on fossil fuel consumption from a carbon tax, i.e. the first term in (19) or (40). Fossil fuels are disaggregated as petroleum products/natural gas (x) and coal (y). Notation is the same as beiore. Interfuel substitution effects are ignored in order to be consistent with calculations in Case A-c. The expression for welfare cost becomes TI_ 2e (T) (1 T2 )2]px2 2 ~~~(pir) 2 (42) - 1 E {'(T, )2 - (TY + 1) 2 2 ~~~(pY) 2 Note ftat the second term is positive (welfare gain) if T2i < 2 1 Tly I . This is because T1? < o is a subsidy. 46 REFERENCES Anderson, Dennis (1991) "Energy and The Environment". Special Briefing Paper, The Wealth of Nations Foundation. Angell, V. and Srinivasan, S. (1991) "Economic Issues Related To Greenhouse Gas Emissions Abatement in Asia". Discussion Paper (draft), Asia Technical Department, Environment Division. World Bank. Asian Development Bank (1989): Energy Indicators of Developing Member Countries of ADB. Ballard, Charles L. (1990) "Marginal Welfare Cost Calculadons - Differential Analysis vs. Balanced-Budget Analysis". ourmal of Public Economics, 41:263-276. Bates, R. and Moore, E (1991) "Commercial Energy Efficiency and The Environment". Background Paper No. 5. World Development Report 1992. The World Bank. Washington, D.C. Bemow, S. and Marron, D. (1990) "Valuation of Environmental Externalities for Energy Planning and Operations, 1990 Update. Cambridge, Massachusetts: Tellus Institute. Bernstein, Jeffrey and Shah, Anwar (1991b) "Taxes, Incentives and Production: The Case of Pakistan". Processed. Birdsall, Nancy and John Dixon (1991). Some Economics of Global Climate Change: The Developing Countries. Processed. Browning, Edgar K. (1987). On the Marginal Welfare Costs of Taxation. The Amcan Economic Review, March, pp. 11-23. Burgess, Joanne C. (1990) "The Contribution of Efficient Energy Pricing to Reducing Carbon Dioxide Emissions". En=g Polijy, 18(5):449-55. Chishti, Salim and Fakhre Mahmood (1991). The Energy Demand in the Industral Sector of Pakdstan. The PaldgtaDMvMment Revie, 30:1 (Spring 1991), pp.83-88. Churchill, A.A. and Saunders, R.J. (1991) "Global Warming and The Developing World". Finance and Development, June. Diwan, I. and Shafik, N. (1991) "Towards a North-South Deal on the Environment". The World Bank. Draft. 47 Eckaus, Richard S. (1991) "Economic Issues in Greenhouse Warming". Presented at the World Bank, WDR Seminar, May 1991. Energy Information Administration: International Energy Annual 1989. DOE/EIA-0219 (89). Epstein, J.M. and Gupta, R. (1990) "Controlling the Greenhouse Effect - Five Global Regimes Compared"; The Brookings Institution, Brookings Occasional Papers. Washington, D.C. Feldstein, Martin (1978) "The Welfare Cost of Capital Income Taxation". Jourmal of Political EQonomy, 86(2):29-51. Glomsrod, S., Vennemo H. and Johnsen, T. (1990) "Stabilization of Emissions of C02: A Computable General Equilibrium Assessment". Discussion Paper No. 48. Central Bureau of Statistics, Oslo, Norway. Goulder, Lawrence H. (1991) "Effects of Carbon Taxes in an Economy with Prior Tax Distortions: An Intertemporal General Equilibrium Analysis for the U.S.". Stanford University and NBER. Draft. Green, J.R. and Sheshinski, E. (1979) "Approximating The Efficiency Gain of Tax Reforms". Journal of Public Economics, 11:179-95. Haugland, T. and Roland, K. (1990) "Energy, Environment and Development in China". The Fridtjof Nansen Institute. Report 1990/17. Hoel, Michael (1990). C02 and the Greenhouse Effect: A Game Theoretic Exploration. University of Oslo, Preliminary Draft. Hoel, Michael (1991) "Efficient International Agreements for Reducing Emissions of C02". The Energy Journal, 12(2):93-107. International Energy Agency: Energy Prices and Taxes 4/1990. OECD Paris 1991. International Monetary Fund (1991): Intemational Financial Statistics. Iqbal, Mahmood (1986) "Substitution of Labor, Capital and Energy in the Manufacturing Sector of Pakistan". Empirical Economics, 11(2):81-95. Jeetun, A. (1978) "Tax Shiffing in Pakistan: A Case Study of Excise Duty, Sales Tax, and Import Duty". Discussion Paper 30. Applied Economics Research Centre, University of Karachi. Processed. 48 Jorgenson, D.W. and Wilcoxen, P.J. (1990) "The Cost of Controlling U.S. Carbon Dioxide Emissions". Presented at Workshop on Economic/Energy/Environmental Modeling for Climate Policy Analysis, Washington, D.C., October 22-23, 1990. Kosmo, Mark (1989) "Commercial Energy Subsidies in Developing Countries". EnergyQoicy, June:44-53. Larsen, Bjom aaud Shah, Anwar (1992a) 'World Energy Subsidies and Global Carbon Emissions." Background paper No. 25. World Development Report 1992. The World Bank. Washington, D.C. larsen, Bjom and Shah, Anwar (1992b) "Tradeable Carbon Emissions Permits and International Transfers." Presented at the 15th Annual International Conference of the International Association for Energy Economics, Tours, France, May t8-20, 1992. Maler, K.G. (1989) "The Acid Rain Game" in H. Folmer and E. van Ierland (eds): Valuation Methods and Policy Makdng in Environmental Economics. Elsevier, Amsterdam. Manne, A.S. and Richels, R.G. (1991) "Global C02 Emission Reductions - the Impacts of Rising Energy Costs". The Enrgy Journal, 12(1):87-107. Moore, Edwin A. (1991) "Energy Efficiency and The Environment'. Report for Environmental Policy and Research Division, Environment Departnent, World Bank. Morey, Edward R. (1984) "Confuser Suiplus". De Ameican lconomic Review, 74(1):163-73. Nordhaus, William D. (1991a) "The Cost of Slowing Climate Change: a Survey". The Eynr JgMal, 12(1):37-64. Nordhaus, William D. (1991b) "To Slow Or Not To Slow: The Economics of The Greenhouse Effect". The Economic Joumal, 0lI(July):920-37. Oates, Wallace (1991). Pollution Charges as a Source of Public Revenue. Processed. Oates, Wallace and Paul R. Portney (1990). Policies For the Regulation of Global Carbon Emissions. Processed. OECD (1989): Energy and The Environment: Policy Overview. OECD (1991): Estimation of greenhouse gas emissions and sinks. (Draft). Pearce, David (1991) "The Role of Carbon Taxes in Adjusting to Global Warming". 3k Econom Journal, 10(July):938-48. 49 Pindyck, R.S. (1979) "Interfuel Substitution and the Industrial Demand for Energy: An International Comparison". Review of Economics and Statistics, 61:169-79. Pitt, Mark (1985) "Estimating Industry Energy Demand with Firm-Level Data: The Case of Indonesia". The Energy Journal, 6(2):25-39. Postel, S. and Flavin, C. (1991) "Reshaping the Global Economy" in Lester R. Brown ed. Sa of the World, pp. 170-88. Poterba, James M. (1991) "Tax Policy to Combat Global Warming: On Designing a Carbon Tax." In R. Dombusch and J.M. Poterba (eds.): Global Warming: Economic Policy Response. M1T Press, Cambridge, Massachusetts. Pourgerami, A. and von Hirschhausen, R. (1989) "Aggregate Demand for Energy and Dynamics of Energy Demand Elasticities in Non-Oil Developing Countries". The Journal of Energy and Dcevopment, 14(2):237-52. Radian Corporation (1990) "Emission and Cost Estimates for Globally Significant Anthropogenic Combustion Sources of NOX, N20, CH4, CO and C02". Prepared for Office of Research and Development, U.S. Environmental Protection Agency. Repetto, Robert (1990) "Environmental Productivity and Why It Is So Important." Challenge, Sep-Oct 1990. Shah, Anwar and Baffes, John (1991) "Do Tax Policies Stimulate Investment in Physical and Research and Development Capital?". PRE Working Paper No. 689. The World Bank. Shah, Anwar and Larsen, Bjorn (1991) "Taxing Choices for Global Environmental Security". Research Proposal, Public Economics Division, Country Economics Department, World Bank Shah, Anwar and Larsen, Bjom (1992) "Global Warming, Carbon Taxes and Developing Countries." Presented at the 1992 Annual Meetings of the American Economic Association, January 3, 1992, New Orleans, USA. Shah, Anwar and Whalley, John (1991) :Tax Incidence Analysis of Developing Countries: An Alternative View". The World Bank Economic Review, 5(3):535-52. Sterner, Thomas (1989) "Oil Products in Latin America: The Politics of Energy Pricing". 1k Eo=y Journal, 10(2):25-45. Summers, Lawrence H. (1991) "The Case For Corrective Taxation". The National Tax Jounmal, September. Vol. XLIV, No. 3: 289-292. 50 Summers, Robert and Heston, Alan (1991) "The Penn World Table (Mark 5): An Expanded Set of International Comparisons, 1950-1988". Quartly Jo.numal of Economics, 106:327-68. Tietenberg, T. (1985) "Emission Trading: An Exercise in Reforming Pollution Policy". Resources for the Future, Washington, D.C. The World Bank (1985) "Domestic Coal Pricing: Suggested Principles and Present Policies in Selected Countries". Energy Departnent Paper No. 23. The World Bank (1990) "Review of Electricity Tariffs in Developing Countries During the 1980's". Industry and Energy Department Working Paper Energy Series Paper No. 32. The World Resources Institute (1990): World Resources 1990-91. Uri, Noel D. (1979) "Energy Demand and Interfuel Substitution in India". Er== EoS Rve, 12(2):181-90. Vashist, D.C. (1984) "Substititdon Possibilities and Price Sensitivity of Energy Demand in Indian Manufacturing". The Indian Economic Journal, 32(2):84-97. Wiliams, M. and Laumas, P. (1981) "The Relation Between Energy and Non-Energy Inputs in India's Manufacturing Industries". The Journal of Industrial Economics, 30(2):113-22. Willig, Robert D. (1976) "Consumer's Surplus Without Apology". The Amercan Economic mview, 66(4):589-97. 51 Policy Research Working Paper Series Contact Title Author Date for paper WPS934 Public Hospital Costs and Quality Maureen A Lewis July 1992 P. Trapani in the Dominican Republic Margaret B. Sulvetta 31947 Gerard M. LaForgia WPS935 The Precautionary Demand for Boum-Jong Choe July 1992 S. Lipscomb Commodity Stocks 33718 WPS936 Taxation, Information Asymmetries, Andrew Lyon July 1992 C. Jones and a Firm's Financing Choices 37699 WPS937 How Soft is the Budget Constraint Evan Kraft July 1992 CECSE for Yugoslav Firms? Milan Vodopivec 37178 WPS938 Health, Govemment. and the Poor: Nancy Birdsall July 1992 S. Rothschild The Case for the Private Sector Estelle James 37460 WPS939 How Macroeconomic Policies Affect Daniel Kaufmann July 1992 D. Kaufmann Project Performance In the Social Yan Wang 37305 Sectors WPS940 Private Sector Approaches to Karen G. Foreit August 1992 0. Nadora Effective Family Planning 31091 WPS941 Projecting the Demographic Impact Rodoifo A. Bulatao August 1992 0. Nadora of AIDS Eduard Bos 31091 WPS942 EfHicient Environmental Regulation: Arik Levinson August 1992 WDR Case Studies of Urban Air Pollution Sudhir Shetty 31393 (Los Angeles, Mexico City, Cubatao, and Ankara) WPS943 Burden-sharing among Official and Asli Demirg0c-Kunt August 1992 K. Waelti Private Creditors Eduardo Fernindez-Arias 37664 WPS944 How Public Sector Pay and Gail Stevenson August 1992 PHREE Employment Affect Labor Markets: 33680 Research Issues WPS945 Managing the Civil Service: What Barbara Nunberg August 1992 P. Infente LDCs Can Learn from Developed 37642 Country Reforms WPS946 Retraining Displaced Workers: What Duane E. Leigh August 1992 PHREE Can Developing Countries Learn from 33680 OECD Nations? WPS947 Strategies for Creating Transitional Stephen L Mangum August 1992 PHREE Jobs during Structural Adjustment Garth L Mangum 33680 Janine Bowen Policy Research Working Paper Series Contact Title Author Date for paper WPS948 Factors Affecting Private Financial Mohua Mukherjee July 1992 R. Lynn Flows to Eastern Europe, 1989-91 32169 WPS949 The Impact of Formal Finance on the Hans Binswanger August 1992 H. Binswanger Rural Economy of India Shahidur Khandker 31871 WPS950 Service: The New Focus in Hans Jl)rgen Peters August 1992 A. Elcock International Manufacturing and Trade 33743 WPS951 Piecemeal Trade Reform in Partially Glenn W. Harrison August 1992 D. Ballantyne Liberalized Economies: Thomas F. Rutherford 38004 An Evaluation for Turkey David G. Tarr WPS952 Unit Costs, Cost-Effectiveness, and Susan Horton August 1992 0. Nadora Financing of Nutrition Interventions 31091 WPS953 The 'Pedigree' of IEC Conversion Michael Hee August 1992 E. Zamora Factors for Per Capita GNP Computations 33706 for the World Bank's Operational Guidelines and Atlas WPS954 How OECD Policies Affected Latin Chris Alien August 1992 T. G. Srinivasan America In the 1980s David Currie 31288 T. G. Srinlvasan David Vines WPS955 OECD Fiscal Policies and the George Alogoskoufis August 1992 D. Gustafson Relative Prices of Primary Panos Varangis 33714 Commodities WPS956 Regression Estimates of Per Capita Sultan Ahmad August 1992 E. O'Reilly-Campbell GDP Based on Purchasing Power Parities 33707 WPS957 Carbon Taxes, the Greenhouse Anwar Shah August 1992 WDR Office Effect, and Developing Countries Bjorn Larsen 31393