41760 Paper number 113 E N V I R O N M E N T D E P A R T M E N T P A P E R S Climate Change Series Growth and CO2 Emissions How Do Different Countries Fare? Robert W. Bacon Soma Bhattacharya November 2007 Sustainable Development Vice Presidency THE WORLD BANK ENVIRONMENT DEPARTMENT Growth and CO2 Emissions How Do Different Countries Fare? Robert W. Bacon Soma Bhattacharya November 2007 Papers in this series are not formal publications of the World Bank. They are circulated to encourage thought and discussion. The use and citation of this paper should take this into account. The views expressed are those of the authors and should not be attributed to the World Bank. Copies are available from the Environment Department of the World Bank by calling 202-473-3641. © The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. Manufactured in the United States of America First printing November 2007 Design: Jim Cantrell Cover photo: Gas flaring, natural gas being burned to CO2 during oil production: Simone D. McCourtie/World Bank Photo Library Contents Acknowledgments v Abbreviations and Acronyms vii Executive Summary 1 Chapter 1 Background 5 Chapter 2 The Technique of Decomposition Analysis 7 Chapter 3 Previous Studies on the Decomposition of CO2 Emissions from Fossil Fuels 11 Chapter 4 Variables and Data Sources 13 Chapter 5 Emissions Levels and Decomposition of Emissions Changes Between 1994 and 2004 15 Chapter 6 Conclusions 31 Appendix 1 Decomposition of CO2 Emissions for India Comparing EIA and Government of India Data 33 Appendix 2 Decomposition of Emissions During Subperiods 35 RefeRences 39 Climate Change Series iii Growth and CO2 Emissions -- How do Different Countries Fare? figuRes Figure 1. Top 20 Countries Ranked by Percentage Growth in Emissions Between 1994 and 2004 17 Figure 2. Emissions per Unit of GDP Measured at PPP and at MER, 2004 20 Figure 3. Emissions per Capita and GDP per Capita, PPP, 2004 22 Figure 4. Emissions per Capita and GDP per Capita, MER, 2004 22 Figure 5. Decomposition for "Negative Offsetting" Countries, 1994­2004 27 Tables Table 1. CO2 Emissions, 2004 16 Table 2. The Ratio of Fossil Fuel CO2 Emissions to Total GHG Emissions, 2000 18 Table 3. Emissions per Unit of GDP and GDP per Capita, 2004 19 Table 4. Emissions per Capita and GDP per Capita, 2004 21 Table 5. Decomposition of the Change in CO2 Emissions Between 1994 and 2004 23 Table 6. Offsetting Coefficients for Decomposition of Emissions, 1994­2004 25 Table 7. Offsetting Coefficients and Changes in Emissions, 1994­99 and 1999­2004 28 iv Environment Department Papers Acknowledgments T his report was prepared by a World Bank team Bank, and Joe Roop of Pacific Northwest Labs comprising Robert W. Bacon (lead author), (United States). The report was funded in part by the Soma Bhattacharya, Richard Damania, Masami Energy Sector Management Assistance Program, a Kojima, and Kseniya Lvovsky. Valuable joint program of the United Nations Development comments were provided by Stene Jorgensen, Charles Programme and the World Bank. The authors are Di Leva, Laszlo Lovei, Alan Miller, Sachiko Morito, grateful to James Warren Evans, Director, Environment Govinda Timilsina, and Vivien Foster of the World Department, for support and guidance. Climate Change Series v Abbreviations and Acronyms Btu British thermal units CO2 carbon dioxide EIA Energy Information Administration GDP gross domestic product GHG greenhouse gas IPCC Intergovernmental Panel on Climate Change LMDI logarithmic mean Divisia index MER market exchange rate PPP purchasing power parity Climate Change Series vii Executive Summary T he recent Intergovernmental Panel on Climate the Energy Information Administration (EIA) Change (IPCC) Fourth Assessment report of the US Department of Energy and the World (2007) has deepened understanding of the Resources Institute. magnitude of climate change, its consequences, · Provides a decomposition of the change in fossil and possible policies to reduce and mitigate its fuel CO2 emissions between 1994 and 2004 into effects. The combustion of fossil fuels is the largest changes in five factors: the average emissions per single contributor to carbon dioxide (CO2) and total unit of fossil fuel consumed (the carbon intensity greenhouse gas (GHG) emissions and, of all major of fossil fuel); the share of fossil fuel consumption sources, has grown the most rapidly over the period in total energy consumption (including fossil fuel 1970 to the present. Furthermore, the IPCC 2007 consumption and non­fossil fuel power produc- report shows that a long-observed trend in declining tion, but not solar heating of water and the like); global CO2 emission intensity per unit of GDP reversed total energy consumption per unit of GDP (energy around 2000. This means that, with world economic intensity); GDP per capita; and population. This growth the strongest it has been in decades, global CO2 analysis excludes the use of biomass, for which emissions are growing faster than at any time since there is no comparable data available 1970. · Carries out this decomposition for the top 70 countries measured by 2004 emissions of CO2 The growth of CO2 emissions over time has shown from fossil fuels. Half of these are developing substantial variation among countries, and there is countries. This provides a much wider coverage of considerable interest about which countries have emissions and their decomposition than given in recently experienced rapid growth of emissions and previous studies. why. The relationship between this growth and changes · Bases the decomposition on changes in emissions in various structural factors in the economies, such as and the related factors between 1994 and 2004, energy intensity and the share of fossil fuels in total providing a more up-to-date study than others cur- energy consumption, as well as between this growth rently available. This update is important because and the growth of the economies themselves, has been of significant changes observed in the growth of the focus of a number of studies. CO2 emissions for a number of countries around This report 2000. · Splits the decade into two five-year subperiods · Provides comparative data on a number of (1994­99, 1999­2004) to identify how the relative measures of emissions for a wide range of coun- importance of the different factors changed during tries, allowing comparisons to be made among the period. the different measures. The main data sources are Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? The principal findings of the report follow: 4. The ranking of countries by emissions intensity changed substantially relative to the ranking by 1. The top 70 countries ranked by CO2 emissions total CO2 emissions. Among large developing from fossil fuels in 2004 accounted for 95 percent economies, Brazil, India, and Mexico ranked of the global total. The top 20 were mostly drawn very much lower on emissions intensity than on from higher income developed countries, but also total emissions among the group of 70 countries included several major developing economies. In studied. Several middle-income countries were addition, several large oil-producing countries among those with the highest rankings. were in the top 30. During the period 1994­2004, 5. Emissions per capita were positively but only some 15 countries experienced a fall in CO2 moderately correlated with GDP per capita and emissions from fossil fuel consumption, and this showed no evidence of an eventual decline in group comprised mainly Eastern European and emissions per capita at higher per capita income Central Asian countries. The percentage growth in (the Environmental Kuznets Curve phenomenon). emissions during the decade showed considerable The rankings relative to those for total emissions, variation among countries, with some large emit- or emissions per unit of GDP, were quite different. ters experiencing large growth and others relatively Some countries with large populations, such as low growth. India and Indonesia, moved almost to the bottom 2. The ratio of CO2 emissions from consumption of of the ranking based on this measure, while some fossil fuels to total GHG emissions is more than countries with small populations moved to the top 50 percent for 55 of the countries, and rises to of the rankings. more than 80 percent for 14 countries. The group 6. The decomposition analysis related the change of countries where the ratio is particularly low in emissions during the decade to changes in the includes those where the most effective policies five factors. For the group of countries as a whole, to reduce GHG emissions are likely to be distinct GDP per capita was the dominant variable linked from those designed to reduce CO2 emissions from to the growth in total emissions, with population fossil fuels. being only one-half as important. However, the 3. The intercountry distribution of emissions per decrease in energy intensity was so large that it unit of GDP (emissions intensity) was measured offset about 40 percent of the combined influence using GDP at purchasing power parity (PPP) and of these two factors. The fossil fuel mix showed a at market exchange rates (MER). The emissions small negative effect on the change in emissions, intensity was, for most countries, very much lower while the share of fossil fuels in total energy on the former measure. The ranking of countries consumption contributed a small increase to the by emissions intensity also changed substantially change in emissions. relative to the ranking by total emissions, and 7. The relative importance of the three factors that was not systematically related to GDP per capita. could be directly linked to policies to slow down Because the ratio of the measure of GDP at PPP the growth of emissions (fossil fuel mix, share of to GDP at MER stayed constant throughout the fossil fuels in total energy, and energy intensity decade, the percentage change in emissions per of GDP) to the positive effects of GDP growth is unit of GDP was identical for the two measures, measured by the "offsetting" coefficient. For the and this analysis of changes in emissions uses the decade as a whole, 15 countries more than offset PPP measure. the combined impacts of GDP and population growth. This group consisted mainly of former 2 Environment Department Papers Executive Summary Soviet Union and Eastern European countries, but Six countries moved from positive to negative also included two Scandinavian countries, whose offsetting between the two subperiods, indicating policies toward emissions may give important clues that they had actually seen emissions rise faster for other countries wanting to slow the growth of relative to GDP. The experiences of the countries emissions. There were also 19 countries where the with largest absolute emissions across the two changes in the potential offsetting factors actually subperiods were quite different. The United States led to a further increase in emissions beyond that and India both experienced substantial offsetting which would have been expected from the growth throughout the period, while China and Russia of GDP ("negative offsetting"). The group included both experienced high offsetting in the first period, several oil producers, but also some non-oil-pro- but much reduced offsetting in the second. For the ducing, high-income developed countries. group as a whole, there was no apparent correla- 8. Among countries with the largest absolute emis- tion between the improvement in offsetting and sions, the United States was able to offset a sizeable the level of GDP, indicating that during the period fraction of its growth in emissions but, given the studied higher income countries had not per- size of the economy, the increment was still large. formed better in slowing the growth of emissions By contrast, Japan, with relatively low income and relative to GDP. population growth, experienced negative offset- ting, but the total increment was relatively small. However, a number of countries did experience The Russian Federation fully offset the growth in improved performance of emissions relative to emissions related to income and population during GDP, suggesting that there need not be a negative this period but, given that population decline and trade-off between slowing the growth of emissions massive changes in sectoral structure and energy and maintaining high growth rates of the economy. efficiency were stimulated by the political changes Countries that were particularly successful in achieving during this period, a similar performance may be high offsetting could well serve as case studies for difficult to maintain in the coming decade. China how this might be achieved. However, the experience and India, while enjoying rapid economic growth of several countries also makes it clear that, without during the decade, were also able to offset signifi- active policies to curb the emissions intensity of the cant portions of the potential growth in emissions. economy, emissions can actually increase faster than 9. When the data was split into two five-year GDP, even when GDP has reached a high level. More subperiods, it became apparent that for some detailed analysis could shed further light on the factors countries large shifts in structure had taken place determining aggregate energy intensity for an economy, in the course of the decade. For one group of because this factor appears to have accounted for the countries, the offsetting coefficient increased largest differences in performance between economies, substantially--this group included several develop- once the growth of GDP is allowed for. In particular, ing countries. An important group of 13 countries the distinction between changes resulting from sectoral experienced negative offsetting in both subperiods, composition and changes resulting from improvements indicating that this was a longer term trend--the in energy efficiency may be helpful in giving clues about group included some major oil producers, and the future course of emissions, and the possibilities of some high-income, non-oil-producing countries. finding policies that could make a substantial difference to global CO2 emissions. Climate Change Series 1 Background T he recent Intergovernmental Panel on Climate country-by-country investigation, and to provide Change (IPCC) Fourth Assessment report focus on the key factors that are amenable to policy (2007) has deepened understanding of the interventions. magnitude of climate change, its consequences, and possible policies to reduce and mitigate its The principal tool for describing the relationship effects. The combustion of fossil fuels is the largest between the growth of emissions and changes in various single contributor to carbon dioxide (CO2) and total related factors is "decomposition analysis." Although greenhouse gas (GHG) emissions and, of all major there is substantial literature that applies this technique sources, has grown the most rapidly over the period to fossil fuel emissions of CO2, much of it concentrates 1970 to the present. Furthermore, the IPCC 2007 on high-income countries, with one or two large report shows that a long-observed trend in declining developing countries also being covered. Studies global CO2 emission intensity per unit of gross covering a slightly wider range of countries used data domestic product (GDP) reversed around 2000, from 2002 or earlier, so recent important developments meaning that, with the strongest economic growth of are not reflected in these studies. the world economy in decades, global CO2 emissions are growing faster than at any time since 1970. This report is designed to serve several purposes: The growth of CO2 emissions over time has shown · It provides a statistical decomposition, or substantial variation among countries, and there is breakdown, of fossil fuel CO2 emissions into five considerable interest about which countries have factors: the average emissions per unit of fossil fuel recently experienced rapid growth of emissions. The consumed (the carbon intensity of fossil fuel); the relationship between this growth and changes in various share of fossil fuel consumption in total energy factors, such as the energy intensity and the share of consumption (including fossil fuel consumption fossil fuels in total energy consumption, as well as the and non­fossil fuel power production, but not growth of the economy itself, has been the focus of a solar heating of water and the like); total energy number of studies. The primary purpose of this report consumption per unit of GDP (energy intensity); is to provide a comparative baseline analysis of the GDP per capita; and population. This analysis change in fossil fuel emissions of a large number of excludes the use of biomass, for which there is no countries that can serve as a starting point for detailed comparable data available. Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? · The decomposition is carried out for the top 70 The analysis compares annual emissions at these countries measured by 2004 emissions of CO2 two points in time and relates the change in these from fossil fuels. Half of these are developing emissions to changes in a number of variables over countries. This provides a much wider coverage of the same period. emissions and their decomposition than has been analyzed in previous studies. The report begins with an account of the technique of statistical decomposition analysis and briefly mentions · The decomposition is based on changes in some other studies that have used it. This is followed emissions and the related factors between 1994 by a discussion of the model used in this report and the and 2004, providing a more up-to-date study than data sources. Countries in the study are then ranked others currently available. This update is important by various measures of emissions including total CO2 because of significant changes observed in the fossil fuel emissions, emissions per capita, and emissions growth of CO2 emissions for a number of countries per unit of GDP. This is followed by the five-factor around 2000, which is explored through splitting decomposition and discussion of the results. the decomposition into two five-year subperiods. Environment Department Papers 2 The Technique of Decomposition Analysis T he decomposition of fossil fuel CO2 emissions Kaya (1990) was influential in proposing an identity into related factors dates back to a series of around which a decomposition of emissions related to studies undertaken in the 1980s, mainly at the four factors could be based: industry level for a single industrialized country. CO2emissions from energy CO2 emissions per unit of energy consumed × energy consumed per unit of GDP × GDP per capita × population (1) This has subsequently been expanded: CO2 emissions from energy CO2 emissions per unit of fossil fuel consumed × fossil fuel consumed per unit of energy consumed × energy consumed per unit of GDP × GDP per capita × population (2) These identities focus on CO2 emissions from the be expressed as the sum of absolute changes in the five combustion of fossil fuels (oil, gas, and coal). Although factors. Various solutions to providing a satisfactory and these are identities that must always be satisfied by the complete decomposition of the changes in emissions, data, and are not based on an estimated model of causal related to the sum of a measure of changes of the links between the variables, the movements of the factors, have been reviewed by Ang (2004) and a widely components provide an important guide to changes in used solution is based on the so called logarithmic factors influencing CO2 emissions from energy use. mean Divisia index (LMDI 1) as explained by Ang (2005). Because the variable of interest--emissions from the consumption of energy--is related to the product of According to Lee and Oh (2006), equation (2) can be several factors, the change in emissions cannot simply rewritten as follows: Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? E = the amount of CO2 emissions from the consumption of fossil fuel FEC = the amount of fossil fuel consumption TEC = the total primary energy consumption GDP = gross domestic product POP = population. Hence, emissions in country i can be expressed as Ei (Ei / FECi) × (FECi /TECi) × (TECi / GDPi) × (GDPi / POPi) × (POPi) (3) Ci Si Ii Gi Pi (4) The change in a country's emissions (Ei) between a energy, termed the substitution effect, Seff); (c) the base year 0 and an end year T can be decomposed into change in E (the energy intensity effect, Ieff); (d) the the effects of (a) the change in C (the emissions per change in GDP per capita (Geff); and (e) the change in unit of fossil fuel, termed the coefficient effect, Ceff) population (Peff). ; (b) the change in S (the share of fossil fuels in total Ei Ei(T) ­ Ei(0) Ceff + Seff + Ieff + Geff + Peff (5) The effects, in turn, can be calculated from the following formula using LMDI1: Ceff = [Ei(T) ­ Ei(0)] × { ln [ Ci(T) / Ci(0) ] / ln [ Ei(T) / Ei(0) ] } (6) Other effects (Seff, Ieff, Geff, Peff) can also be derived from similar formulae. With data on all the variables for a common base year The ratio of fossil fuels consumed to total energy and terminal year, the decomposition of the change in consumed will rise if the share of non­fossil fuels emissions can be calculated according to equation (6). (hydro, nuclear, and renewables) falls relative to the share of fossil fuels consumed. The energy intensity The change in emissions will reflect changes in the five of the economy will fall if the use of energy increases factors because of the nature of the identity linking more slowly than the level of GDP. This can occur for them. To interpret these links, it is necessary to consider two main reasons. First, if the sector structure of GDP the circumstances under which the factors that can be changes toward sectors that are less energy intensive, directly influenced by emissions-related policies will without any other changes, the average use of energy change. in total GDP would fall. Second, if energy efficiency The carbon emissions per unit of total fossil fuel increased in one or more sectors, without any structural consumption will rise if there is a relative shift to shifts, the overall energy intensity would fall. The higher-emitting fuels (for example, the share of coal aggregate form of decomposition analysis used in this rising relative to the share of gas). This could occur paper does not distinguish between these two effects. even if the total quantity of fossil fuels consumed stayed constant. Environment Department Papers The Technique of Decomposition Analysis The exclusion of the use of biomass fuel outside the small establishments switch from biomass, commercial power sector limits the generality of analysis, especially or otherwise, to fossil fuels and electricity--and because its use varies greatly among countries. However, households are most certain to do so with increasing its omission is unlikely to have a large effect on the income--fossil fuel intensity and CO2 emissions may current level of total CO2 emissions because the use of both rise more than what might be projected based on biomass is generally, but not always, carbon neutral. past trends. Including biomass consumption outside the power sector would produce an increase in measured energy intensity and decrease the share of fossil fuels in total Note energy use relative to what is reported in this paper. 1. For small changes in the factors, this formula The differences between the results obtained with and apportions the total change in emissions according without full inclusion of biomass would obviously to the ratio of the growth rate of each factor to the depend on the amount of biomass consumed outside growth rate of emissions. the power sector. In the future, if households and Climate Change Series Previous Studies on the 3 Decomposition of CO2 Emissions from Fossil Fuels A large number of studies using decomposition Paul and Bhattacharya (2004) used a four-factor analysis now exist. Those of the greatest interest decomposition for the four major sectors (agriculture, in relation to this report are ones that make a industry, residential, and other) for India between 1980 comparative analysis using several countries, and 1996. They separated the energy intensity effect especially those that include a wide range of income from a structural effect by analyzing the emissions of levels. Some of the most recent and related studies each sector and the GDP of each sector. Allowing for are briefly mentioned below. These studies do not this refinement, they showed that the GDP effect was all use the logarithmic mean Divisia index (LMDI) dominant, but that the structural effect was associated decomposition but, because the principles are similar in with an increase in emissions, as the economy shifted each case, this aspect is not reported. relatively into sectors with higher emissions per unit of GDP. Changes in energy intensity for each sector were Ang and Zhang (1999) used the five-factor of relatively low importance once structural changes decomposition for groupings of countries in 1993 had been factored into the decomposition. to decompose the differences between regions of the emissions from fossil fuel use. The regions included Nag and Parikh (2005) analyzed the decomposition of three Organisation for Economic Co-operation and total emissions from the Indian power sector between Development groupings, the former Soviet Union 1974 and 1998 using a sector-based identity similar to and Central and Eastern Europe, and the rest of the that based at an economy level as explained above in world. Differences in GDP and energy intensity were Paul and Bhattacharya (2004). They produced scenarios the dominant factors in explaining the large differences of emissions until 2015 based on assumptions about found between regions. changes in the various coefficients identified in the decomposition. Viguier (1999) used a four-factor decomposition for the periods 1980­90 and 1990­94 for six countries Wang, Chen, and Zou (2005) carried out a five-factor (France, Hungary, Poland, the former Soviet Union, decomposition analysis for China for the periods 1957­ the United Kingdom, and the United States). The 79 and 1979­2000. They showed that the change in study covered emissions of NOx (nitrogen oxides) and economic policy in 1979 was associated with a large SO2 (sulfur dioxide) as well as CO2. The reduction change in the importance of the different factors, with of emissions intensity in Western countries was the reduction in energy intensity after 1979 playing a accompanied by a reduction in energy intensity, large role in holding back the growth in emissions. while in the other countries the energy intensity rose. However, changes in fuel mix, especially in the Russian Lise (2006) used a four-factor decomposition for Federation, provided some offset to the increase in Turkey for the period 1980­2003. Again this study energy intensity. identified a structural (composition) effect for the Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? changing shares of sectors in GDP. The GDP (scale) in energy intensity and the share of fossil fuels, and a effect, structure effect, and carbon-intensity effect change in the fossil fuel mix, all contributed to partially (the coefficient effect plus substitution effect) were all offsetting the impacts of growth in the economies. The associated with substantial increases in emissions, while group of lower income countries was dominated by energy intensity was associated with a small reduction China, which, in this period, experienced a large drop in emissions. in energy intensity, offsetting nearly half the impacts of the increase in income and population. Ebohon and Ikeme (2006) derived a three-factor decomposition for nine non-oil-producing and six Herzog, Baumert, and Pershing (2006) produced a oil-producing countries in Africa between 1971 and four-factor decomposition for a selection of eight 1981. The results were very different for oil producers countries (Brazil, China, France, the Republic of and non­oil producers, with much larger effects Korea, Thailand, Ukraine, the United Kingdom, and observed for non­oil producers. The aggregate results the United States) between 1990 and 2002. In most were dominated by Nigeria in the oil-producing group countries, the changing fuel mix was associated with and South Africa in the non-oil-producing group. In a reduction in emissions, as was the decline in energy both groups several countries experienced a decrease in intensity, but GDP and population effects outweighed energy intensity during this period. these in all cases except for the United Kingdom and Ukraine. Lee and Oh (2006) produced a five-factor decomposition for 15 countries in the Asia Pacific Liu and Ang (2007) reviewed a large number of studies Economic Cooperation region between 1980 and 1998. using various approaches to decomposition analysis This group includes high-, middle-, and lower-income and discussed their strengths and weaknesses. They representatives. Although GDP and population were concluded that recently the LMDI 1 methodology has strong factors associated with increases in emissions become the most widely accepted approach. in all cases, in the high-income countries decreases 2 Environment Department Papers 4 Variables and Data Sources T o undertake a comparative analysis of a large each fossil fuel. Different fuels require an individual number of countries it is important to work conversion factor from energy available to emissions with a common data source for each series that produced, and the conversion factor for coal further covers all the countries in the sample to ensure depends on the quality of coal consumed. Details of that a common methodology has been followed in these conversion factors are not provided on the US compiling the data. This may minimize the chance that Department of Energy Web site. Data in million metric larger variations between observations are introduced tons of CO2 emitted are provided for a list of 163 solely because of the use of different assumptions in countries for every year in our data period and from compiling the data. The period 1994 to 2004 was these, the 70 countries with the largest CO2 emissions chosen to permit wide coverage of countries (data in 2004 were selected for further study. coverage is not as full for earlier years) while focusing on the most recent data available for a sufficiently The Consumption of Fossil Fuels lengthy period to identify important changes in the (quadrillion Btu) relationship between emissions and the various factors used. The definitions and sources are explained below. The EIA Web site provides consumption data for the three fossil fuels measured in Btu, and these are aggregated to provide total fossil fuel consumption. The Emission of CO2from Fossil Fuel The ratio of fossil fuel emissions to the consumption Consumption (million metric tons) of fossil fuels provides the value of C (the fossil fuel The data source for emissions of fossil fuel consumption coefficient of emissions; see chapter 2). was the US Department of Energy Web site,1 produced by the Energy Information Administration (EIA). To The Consumption of Primary Commercial arrive at the total emissions from all fossil fuels, it is Energy (quadrillion Btu) necessary to aggregate the emissions from the three fuels involved (coal, oil, and natural gas) into common The EIA figures for the consumption of primary units. This is provided by the EIA through a two stage commercial energy includes oil, dry gas, coal, net calculation. First, consumption of each fuel measured hydro production, net nuclear production, and in physical units is converted into British thermal units renewable sources of electricity supplied to the grid (Btu) as a common energy unit that can be aggregated (net geothermal, solar, wind, and wood and waste). It to provide a measure of total fossil fuel consumption.2 does not include biomass or solar consumed outside The second step is to derive emissions of CO2 for of the power sector. The ratio of the consumption of Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? fossil fuels to the total consumption of primary energy recent census data available. The ratio of GDP to provides the value of S (the share of fossil fuels in total population (G) provides the measure of GDP per energy). capita. Although the data are taken from well-established The Level of GDP (2000 US$ at purchasing sources, where updating and regular publication are power parity) observed, there are clearly possibilities of differences Data on the level of GDP is taken from the World between the common source and government statistics. Development Indicators (World Bank various years). Differences in definition, differences in assumptions The data chosen was in constant 2000 US dollars, about conversion factors to energy equivalents and valued according to purchasing power parity (PPP). CO2 emissions, and lags in publication can all lead to Data valued at market exchange rates (MER) in divergences in figures given for the decomposition and constant US$ was also investigated. Data valued in the basic data used. Inferences made about individual at MER showed, in virtually every case, the same country results are limited by the accuracy of the data percentage increase between 1994 and 2004 as the data available. Although it was not possible to undertake in PPP. This indicated that for each country, the ratio a country-by-country cross check on data, this was between exchange rate-based and PPP-based figures possible for India as part of the World Bank's India Low was constant throughout this period because of the Carbon Growth Study, and the results are discussed in lack of updating of the PPP calculations. The ranking appendix 1. of countries by energy intensity (energy consumption per unit of GDP) does change markedly when the Notes calculation is shifted from PPP to MER, but this does not affect the decomposition analysis, which focuses 1. http://www.eia.doe.gov/emeu/international/con- on changes in emissions during the period. The ratio tents.html of total energy consumption to GDP (I) provides the 2. For coal, which comes in various qualities ranging measure of energy intensity. from lignite to hard coal, the conversion factor depends strongly on the type and quality of the coal consumed and can be a source of differences The Population of the Country (millions) among various estimates of emissions. Data on population is taken from the UN Population Fund and is based on extrapolations from the most 4 Environment Department Papers Emissions Levels and 5 Decomposition of Emissions Changes Between 1994 and 2004 T his chapter begins with a description of the Federation, Germany, and the United Kingdom) level of emissions in each country, measured as while other large emitters experienced substantial total emissions, emissions per unit of GDP, and growth in emissions (China and India). The group emissions per capita. This is followed by the of oil producers also included several with very decomposition of changes in emissions between 1994 rapid increases in the level of emissions, with the and 2004, and then a breakdown of the decomposition exception of Nigeria, where emissions actually fell. into two subperiods, 1994­99 and 1999­2004. The top 20 countries ranked by percentage growth in emissions are shown in figure 1. Although this The Level of Emissions group includes some countries with very low levels of emissions, it also includes several whose levels of Table 1 shows the level of CO2 emissions from fossil emissions are substantial, so that continuation of fuel combustion in 2004 by country, ranked by total this trend could move them rapidly up the rank- emissions. The absolute and percentage changes ings of emitters. The group of mid-level emitters compared with emissions in 1994 are also shown. with rapid growth (such as the Republic of Korea, the Islamic Republic of Iran, Australia, and Saudi Several features are immediately apparent: Arabia) is of particular interest in this respect. · The top 70 emitters accounted for 95 percent of · The majority of countries with the highest levels of emissions in 2004 are high-income, developed global CO2 emissions from fossil fuels, while the top 30 accounted for 86 percent, and the top 10 countries, but several large developing economies accounted for 68 percent. Emissions by country are also in the top 20. are extremely unequal, with the Gini coefficient for · A notable group of countries that appear in the top 30 emitters are large oil producers, some of which the country level of inequality of CO2 emissions among the 70 countries being 0.72.1 have small populations or relatively low per capita · The selection of a cutoff point at the 70th largest incomes. emitter has the effect of omitting countries with · Fifteen countries experienced absolute decreases in emissions during the period; these include several less than 20 million metric tons of CO2 emitted per year. The largest emitter in 2004, the United Eastern European and Central Asian countries, States, emitted nearly 300 times this level. whose economies underwent major transforma- tions during the period. · The percentage growth of emissions between 1994 To place the country data on the emissions of CO2 from the commercial use of fossil fuels in the wider and 2004 showed considerable variation among context of total GHG emissions, data measured in countries. Some large emitters saw their total emissions remain essentially the same (the Russian million tons of CO2 equivalent in 2000 is available Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? Table 1. CO2 Emissions, 2004 (million metric tons) Percent Percent Increase increase Increase increase Emissions since since Emissions since since Country Rank in 2004 1994 1994 Country Rank in 2004 1994 1994 United States 1 5,912 674 13 Greece 36 106 22 26 China 2 4,707 1,911 68 Romania 37 95 -21 -18 Russian Fed. 3 1,685 -5 0 Nigeria 38 94 -1 -1 Japan 4 1,262 174 16 Algeria 39 77 -7 -8 India 5 1,113 384 53 Philippines 40 75 22 42 Germany 6 862 -5 -1 Austria 41 70 13 23 Canada 7 588 95 19 Israel 42 66 18 38 United Kingdom 8 580 12 2 Portugal 43 63 17 38 Korea, Rep. of 9 497 143 40 Chile 44 62 26 70 Italy 10 485 85 21 Finland 45 61 4 7 South Africa 11 430 86 25 Sweden 46 59 0 0 France 12 406 46 13 Vietnam 47 57 30 108 Iran, Islamic 13 402 153 62 Hungary 48 56 -2 -4 Rep. of Australia 14 386 107 38 Denmark 49 56 -9 -13 Mexico 15 385 52 15 Belarus 50 55 -10 -16 Saudi Arabia 16 365 127 53 Colombia 51 55 2 4 Ukraine 17 364 -76 -17 Syrian Arab Rep. 52 53 12 29 Spain 18 362 128 55 Norway 53 51 15 43 Brazil 19 337 69 26 Bulgaria 54 47 -3 -7 Indonesia 20 308 99 48 Switzerland 55 45 3 6 Poland 21 288 -32 -10 Ireland 56 42 13 46 Netherlands 22 267 46 21 Slovak Rep. 57 38 -2 -5 Thailand 23 219 92 72 Bangladesh 58 38 19 97 Turkey 24 212 73 53 New Zealand 59 38 7 22 Kazakhstan 25 172 18 12 Azerbaijan 60 37 -9 -20 Malaysia 26 154 65 73 Trinidad and 61 33 11 50 Tobago Belgium 27 148 20 16 Morocco 62 29 2 7 Egypt, Arab Rep. 28 147 50 51 Peru 63 27 4 19 of Venezuela, R. B. 29 143 24 20 Oman 64 23 8 56 de Argentina 30 142 27 23 Bahrain 65 23 7 44 United Arab 31 141 47 51 Ecuador 66 23 5 27 Emirates Singapore 32 129 48 60 Croatia 67 22 4 23 Uzbekistan 33 121 23 24 Tunisia 68 21 5 30 Czech Rep. 34 112 -7 -6 Dominican Rep. 69 20 11 107 Pakistan 35 106 22 26 Angola 70 20 12 169 Source: World Bank calculations. Note: These numbers are taken from a table posted on the EIA Web site before September 18, 2007. The revised figures for 2004 posted on September 18 differ slightly from those shown in this table. from the Climate Analysis Indicators Tools database CO2 emissions from fossil fuel combustion to total (World Resources Institute 2006). The data on total GHG emissions (including those from land use change GHG emissions include the six main gases--CO2, and international bunkers) shown in table 2 indicates CH4 (methane), N2O (nitrous oxide), PFCs the relative importance of fossil fuels globally and in (perfluorocarbons), HFCs (hydrofluorocarbons), and specific countries. SF6 (sulfur hexafluoride)--from all sources, including land use changes and international bunkers. Although The ratio of fossil fuel CO2 emissions to total GHG the derivation of CO2emissions from fossil fuel emissions varies substantially among countries, but combustion may not be exactly the same as that used by for the majority in the sample, fossil fuel CO2 is the EIA, and 2000 is the latest year currently available responsible for more than 50 percent of the total. For for the range of countries in this study, the ratio of certain countries, such as Indonesia and Brazil, where Environment Department Papers Emissions Levels and Decomposition of Emissions Changes Between 4 and 2004 Figure 1. Top 20 Countries Ranked by Percentage Growth in Emissions Between 1994 and 2004 180 160 140 120 100 Percent 80 60 40 20 0 of of AngolaVietnamRepublic Chile India BangladeshMalaysiaThailand ChinaRep.Singapore Oman SpainArabia Turkey Rep.EmiratesTobago Ireland Arab and Indonesia Islamic Saudi Arab Dominican Iran, Egypt,UnitedTrinidad Source: World Bank calculations. other sources of GHG emissions are predominant, The data in table 3 indicate that the ranking of policies to reduce total emissions need to focus more countries according to emissions per unit of GDP at intensively on non­fossil fuel sources. PPP is different from that for total emissions--with some low population countries appearing at the top of A different perspective is provided by tabulating CO2 the ranking. Oil producers are again well represented emissions per unit of GDP (emissions intensity). Table in the top group. Among large developing economies, 3 provides the information based on GDP measured China, and especially India, Brazil, and Mexico, move at purchasing power parity (PPP) and at market well down the list. Among high-income countries, exchange rates (MER).2 The values of GDP per capita the two largest contributors to global CO2 emissions, are also provided, and the list of countries is ranked the United States and Japan, also move down the list by emissions per unit of GDP at PPP. Although the when ranked by emissions per unit of GDP. Higher correlation between the two measures of GDP per and lower income countries are scattered throughout capita is extremely high (R2 = 0.92), most countries the table, suggesting that there is little evidence of a have rather different values on the two measures. For systematic relationship between emissions per unit of virtually all, the value at PPP is much higher than at GDP and the level of GDP. The relationships between MER, while just a few high-income countries have emissions intensity and the measures of GDP per GDP at MER slightly higher than at PPP. Accordingly, capita are negative--the squared correlation at MER is emissions per unit of GDP are generally very much statistically significant at 21 percent, but that at PPP is lower when measured at PPP. Figure 2 shows a scatter statistically insignificant at 4 percent. plot of emissions per unit of GDP in 2004 based on the two measures. Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? Table 2. The Ratio of Fossil Fuel CO2 Emissions to Total GHG Emissions, 2000 Percent CO2 from fossil Percent CO2 from fossil fuel combustion to total fuel combustion to total Country GHG emissions Country GHG emissions Indonesia 9.3 Oman 66.3 Peru 10.3 Iran 66.5 Malaysia 12.5 Greece 66.8 Angola 13.6 Sweden 68.5 Brazil 13.8 Croatia 68.5 Nigeria 19.1 Denmark 68.9 Ecuador 19.2 South Africa 69.4 Colombia 21.4 Syria 69.7 Bangladesh 23.4 Canada 69.7 Philippines 30.4 Spain 70.0 Pakistan 30.5 Romania 70.1 Singapore 31.5 Belgium 70.2 Venezuela 35.6 France 70.7 New Zealand 38.8 Kazakhstan 71.6 Argentina 39.1 Bulgaria 71.8 Morocco 47.0 Israel 72.0 Thailand 48.8 Hungary 73.1 Tunisia 51.8 Saudi Arabia 73.4 Vietnam 52.0 Azerbaijan 74.2 Chile 53.1 Slovak Republic 75.4 India 53.7 Russia 75.8 Turkey 53.8 Trinidad and Tobago 75.9 United Arab Emirates 55.4 Switzerland 76.6 Egypt 57.2 United Kingdom 76.9 Mexico 58.5 Finland 77.1 Dominican Republic 59.0 Portugal 77.2 Ukraine 59.5 Italy 77.8 China 59.5 Austria 78.5 Algeria 60.9 Poland 78.6 Ireland 61.4 Korea, Republic of 79.0 Uzbekistan 63.6 Germany 80.1 Norway 64.4 Czech Republic 82.3 Belarus 64.6 Bahrain 82.5 Netherlands 64.9 Japan 83.3 Australia 65.2 United States 86.8 Source: World Resources Institute 2006. Table 4 shows emissions per capita for each country in where some authors have found an inverted U shape the list and GDP per capita, with countries ranked by between them, suggesting that initially as income per emissions per capita. The relationship between these capita increases, emissions per capita rise, but then two variables for individual countries over time is as income per capita increases further the level of explored in the Environmental Kuznets Curve literature Environment Department Papers Emissions Levels and Decomposition of Emissions Changes Between 4 and 2004 Table 3. Emissions per Unit of GDP and GDP per Capita, 2004 Emissions/ Emissions/ Emissions/ Emissions/ GDP GDP GDP GDP (metric (metric Per Per (metric (metric Per Per tons per tons per Capita Capita tons per tons per Capita Capita million million GDP, GDP, million million GDP, GDP, Country US$, PPP) US$, MER) PPP MER Country US$, PPP) US$, MER) PPP MER Uzbekistan 2,686 7,204 1,712 639 Israel 437 548 22,950 18,319 Trinidad and 2,054 2,938 12,181 8,516 Croatia 437 973 10,890 4,891 Tobago Kazakhstan 1,785 6,309 6,504 1,840 Indonesia 431 1,564 3,245 894 Bahrain 1,749 2,292 18,148 13,852 New Zealand 423 614 22,423 15,425 United Arab 1,488 1,470 22,135 22,405 422 923 6,951 3,175 Emirates Turkey Ukraine 1,300 8,254 5,949 937 Finland 418 465 28,078 25,239 Russian Fed. 1,298 5,124 9,018 2,285 Germany 403 442 25,905 23,627 Azerbaijan 1,240 4,683 3,551 940 Mexico 402 624 9,061 5,847 Saudi Arabia 1,204 1,698 12,661 8,977 Algeria 392 1,161 6,058 2,046 Singapore 1,202 1,227 25,209 24,689 Chile 378 707 10,168 5,436 Venezuela, 995 1,188 5,457 4,568 Hungary 366 1,017 15,228 5,474 R. B. de South Africa 972 2,821 9,362 3,226 Japan 364 259 27,080 38,041 Belarus 877 3,308 6,425 1,704 India 362 1,887 2,831 542 Iran, Islamic 867 3,163 6,738 1,847 Spain 357 552 23,782 15,372 Rep. of Syrian Arab 861 2,484 3,304 1,145 Denmark 350 334 29,338 30,685 Rep. Bulgaria 799 3,101 7,577 1,953 Pakistan 349 1,235 1,969 556 Nigeria 762 1,663 959 439 Dominican Rep. 341 948 6,786 2,442 Angola 719 1,594 1,772 799 Portugal 332 548 18,278 11,096 Australia 690 848 28,049 22,846 United Kingdom 331 364 29,406 26,741 Oman 663 1,027 13,881 8,961 Italy 326 428 25,641 19,527 China 661 2,745 5,441 1,311 Argentina 316 495 11,750 7,486 Malaysia 658 1,437 9,374 4,296 Ireland 314 359 33,102 28,981 Czech Rep. 638 1,758 17,233 6,251 Norway 307 283 36,234 39,302 Poland 632 1,497 11,797 4,983 Tunisia 290 897 7,170 2,322 Canada 631 747 29,164 24,618 Austria 288 340 29,675 25,087 Netherlands 569 670 28,918 24,560 Vietnam 274 1,394 2,520 496 Romania 569 2,030 7,688 2,154 France 249 287 26,989 23,456 United States 552 552 36,234 36,234 Brazil 247 514 7,406 3,564 Korea, Rep. 546 810 19,108 12,879 Morocco 243 727 3,875 1,297 of Egypt, Arab 541 1,286 3,747 1,577 Sweden 232 224 28,226 29,219 Rep. of Slovak Rep. 536 1,589 13,282 4,479 Philippines 207 836 4,431 1,094 Belgium 499 599 28,437 23,681 Colombia 196 586 6,275 2,099 Greece 476 774 20,077 12,354 Switzerland 194 177 31,958 35,060 Ecuador 463 1,154 3,740 1,501 Peru 194 445 5,122 2,227 Thailand 460 1,457 7,453 2,356 Bangladesh 155 655 1,756 416 Source: World Bank calculations. emissions per capita declines. This literature has been global Gini coefficient for emissions per capita, where reviewed by Stern (2003). all individuals within a country are assumed to share emissions equally, is 0.53. This is almost identical to On a per capita basis, a group of low population the measure of inequality of per capita incomes for the countries heads the ranking, but the United States same group of countries. also has a high value. China, and especially India and Indonesia, all with very large populations, move toward Plots of emissions per capita against income per the bottom of the list. Higher income countries are capita for both PPP and MER measures are shown in found predominantly in the top half of the rankings, figures 3 and 4. The former shows a modest squared and lower income countries in the bottom half. The correlation of 34 percent between the two series and Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? Figure 2. Emissions per unit of GDP Measured at PPP and at MER, 2004 (tons per million 2000 US$) 9,000 8,000 7,000 MER 6,000 at 5,000 GDP/ 4,000 3,000 Emissions2,000 1,000 0 0 500 1,000 1,500 2,000 2,500 3,000 Emissions / GDP at PPP Source: World Bank calculations. the latter 30 percent, indicating that there is a tendency approximation of the decomposition for the global for CO2 emissions per capita to increase as GDP per economy. capita increases. From this cross-section data there is little evidence for the downturn of an Environmental The pattern of results is fairly consistent across Kuznets Curve at higher income levels. countries during this period: · At the global level, increases in GDP per capita The Decomposition of Emissions Between (Geff) were the largest single factor associated with 1994 and 2004 growth of emissions. Population increases (Peff) The decomposition of the change in CO2 emissions were associated with an effect almost half as large between 1994 and 2004 is presented in table 5, where as GDP per capita. A large decrease in energy countries are ranked by decreasing GDP per capita. intensity (Ieff) offset 40 percent of the combined The decomposition is based on GDP measured in PPP effects of these two factors. The effect of the fossil and, as mentioned above, the results are identical for fuel mix (Ceff) was associated with a small decline virtually all countries when GDP in MER is used. The in aggregate emissions, and the share of fossil fuels table also includes the decomposition for the aggregate in total energy (Seff) with a small increase in total of the 70 countries; this can be taken as a close emissions. 20 Environment Department Papers Emissions Levels and Decomposition of Emissions Changes Between 4 and 2004 Table 4. Emissions per Capita and GDP per Capita, 2004 Emissions per GDP per Emissions per capita capita capita GDP per (tons per (2000 US$, (tons per capita (2000 Country person) PPP) Country person) US$, PPP) United Arab Emirates 32.94 22,135 Malaysia 6.17 9,374 Bahrain 31.75 18,148 Portugal 6.08 18,278 Singapore 30.30 25,209 Bulgaria 6.06 7,577 Trinidad and Tobago 25.02 12,181 Iran, Islamic Rep. of 5.84 6,738 United States 20.01 36,234 Belarus 5.64 6,425 Australia 19.36 28,049 Hungary 5.57 15,228 Canada 18.40 29,164 Venezuela, R. B. de 5.43 5,457 Netherlands 16.45 28,918 Croatia 4.76 10,890 Saudi Arabia 15.24 12,661 Uzbekistan 4.60 1,712 Belgium 14.20 28,437 Azerbaijan 4.40 3,551 Finland 11.74 28,078 Romania 4.37 7,688 Russian Fed. 11.71 9,018 Chile 3.84 10,168 Kazakhstan 11.61 6,504 Argentina 3.71 11,750 Norway 11.12 36,234 Mexico 3.65 9,061 Czech Rep. 10.99 17,233 China 3.60 5,441 Germany 10.43 25,905 Thailand 3.43 7,453 Korea, Rep. of 10.43 19,108 Turkey 2.93 6,951 Ireland 10.41 33,102 Syrian Arab Rep. 2.85 3,304 Denmark 10.26 29,338 Algeria 2.38 6,058 Israel 10.04 22,950 Dominican Rep. 2.32 6,786 Japan 9.87 27,080 Tunisia 2.08 7,170 United Kingdom 9.75 29,406 Egypt, Arab Rep. of 2.03 3,747 Greece 9.56 20,077 Brazil 1.83 7,406 New Zealand 9.47 22,423 Ecuador 1.73 3,740 Oman 9.20 13,881 Indonesia 1.40 3,245 South Africa 9.10 9,362 Angola 1.27 1,772 Austria 8.54 29,675 Colombia 1.23 6,275 Spain 8.49 23,782 India 1.02 2,831 Italy 8.36 25,641 Peru 0.99 5,122 Ukraine 7.74 5,949 Morocco 0.94 3,875 Poland 7.46 11,797 Philippines 0.92 4,431 Slovak Rep. 7.12 13,282 Nigeria 0.73 959 France 6.73 26,989 Vietnam 0.69 2,520 Sweden 6.56 28,226 Pakistan 0.69 1,969 Switzerland 6.20 31,958 Bangladesh 0.27 1,756 Source: World Bank calculations. · Fifteen countries experienced an overall decrease in · In the majority of countries, the substitution effect emissions. These countries were predominantly in (the change in the share of fossil fuels in total Eastern Europe and the former Soviet Union. energy consumption), whether positive or nega- · Fifty-six countries experienced a negative value tive, was small. Norway and Canada experienced for the coefficient effect--that is, the emissions increases in the share of fossil fuels in total energy per unit of fossil fuel used declined. This was consumed, which made a large contribution to the probably due to a faster increase in the use of gas overall increase of emissions, while Germany and (especially) and oil than of coal. Nigeria and Brazil Ukraine experienced declines in the share of fossil are examples where this effect was dominant in the fuel that were substantial in relation to the total total change of emissions, while Japan was one of change in emissions. the few countries where an increase in this factor · Forty-nine countries experienced a reduction in was a major contributor to the overall growth in energy intensity, and in several cases this made emissions. a significant contribution in relation to the size Climate Change Series 2 Growth and CO2 Emissions -- How do Different Countries Fare? Figure 3. Emissions per Capita and GDP per Capita, PPP, 2004 35 30 atipacrepsnoissimE 25 )nosrep/ 20 OC 2 15 (ton 10 5 0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 Per Capita GDP (2000 US$, PPP) Source: World Bank calculations. Figure 4. Emissions per Capita and GDP per Capita, MER, 2004 35 30 )nosrep/ 25 OCnot(atipacrepsnoissimE2 20 15 10 5 0 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45,000 Per Capita GDP (2000 US$, MER) Source: World Bank calculations. 22 Environment Department Papers Emissions Levels and Decomposition of Emissions Changes Between 4 and 2004 Table 5. Decomposition of the Change in CO2 Emissions Between 1994 and 2004 (million metric tons) 2004 GDP per capita (2000 US$, Country Ceff Seff Ieff Geff Peff E PPP) Norway -2.4 11.6 -6.7 10.3 2.6 15.4 36,234 United States 6.1 14.6 -1,126.1 1,210.7 569.0 674.5 36,234 Ireland -0.9 -0.2 -12.3 22.3 4.6 13.4 33,102 Switzerland -0.2 1.3 -4.4 4.2 1.6 2.6 31,958 Austria 0.2 2.0 -3.1 12.5 1.3 13.0 29,675 United Kingdom -25.5 8.3 -133.7 143.7 19.6 12.4 29,406 Denmark -1.3 -6.2 -13.5 10.1 2.3 -8.6 29,338 Canada -9.4 38.6 -109.9 124.1 52.2 95.5 29,164 Netherlands 11.1 -2.9 -21.1 46.0 13.3 46.4 28,918 Belgium -6.8 -0.2 -2.5 25.6 4.0 20.1 28,437 Sweden -1.1 -2.3 -13.3 15.2 1.5 -0.1 28,226 Finland -1.2 -2.5 -13.3 19.3 1.7 4.0 28,078 Australia 6.3 3.3 -22.0 80.7 38.8 107.2 28,049 Japan 41.8 -0.4 0.6 106.1 26.0 174.1 27,080 France -2.5 -4.1 -32.0 69.9 15.0 46.3 26,989 Germany -22.2 -23.7 -84.7 111.6 14.1 -4.9 25,905 Italy -6.6 3.2 20.0 61.6 6.5 84.8 25,641 Singapore -4.1 0.0 -0.5 28.6 24.4 48.4 25,209 Spain -4.9 8.7 22.3 82.3 20.2 128.5 23,782 Israel -1.6 0.2 0.7 5.3 13.5 18.1 22,950 New Zealand 1.7 1.3 -7.2 7.7 3.4 6.9 22,423 United Arab -4.6 0.0 -21.6 2.0 71.7 47.4 22,135 Emirates Greece -2.9 -1.8 -8.1 30.3 4.7 22.2 20,077 Korea, Rep. of. -11.0 -21.9 -28.0 175.9 27.9 142.8 19,108 Portugal -1.1 1.3 2.6 12.4 2.3 17.4 18,278 Bahrain -0.3 0.0 -1.7 4.4 4.5 7.0 18,148 Czech Rep. -6.9 -6.1 -23.2 30.3 -1.2 -7.1 17,233 Hungary -4.1 0.0 -20.0 23.2 -1.2 -2.1 15,228 Oman -0.3 0.0 2.0 3.1 3.4 8.3 13,881 Slovak Rep. -2.8 -1.4 -14.5 16.1 0.4 -2.2 13,282 Saudi Arabia -6.6 0.0 70.9 -18.7 81.2 126.8 12,661 Trinidad and -6.3 0.0 1.5 14.7 1.0 10.9 12,181 Tobago Poland -20.0 1.0 -148.0 134.8 0.2 -31.9 11,797 Argentina -4.6 2.8 14.4 0.0 14.1 26.6 11,750 Croatia 0.8 0.5 -5.5 8.7 -0.5 4.0 10,890 of the total change in emissions. This group · In several of the oil-producing countries, energy included some high-income countries (notably the intensity increased during the period, adding to the United States, the United Kingdom, Canada, and effects of growth in income and population. Germany); Russia, Ukraine, and Eastern European · To understand the decomposition for an individual countries (Hungary, Poland) benefiting from the country, it would be necessary to undertake a more restructuring of their economies; and China3 and detailed analysis of policy and structural changes India, where substantial contributions from a during the period. However, the broad patterns reduction in energy intensity in relation to the size established give a guide to the relative importance of the total change in emissions were experienced. of the five factors used in the decomposition. Climate Change Series 2 Growth and CO2 Emissions -- How do Different Countries Fare? Table 5. Decomposition of the Change in CO2 Emissions Between 1994 and 2004 (million metric tons) (continued) 2004 GDP per capita (2000 US$, Country Ceff Seff Ieff Geff Peff E PPP) Chile -3.5 3.4 3.7 15.8 6.2 25.6 10,168 Malaysia -5.2 2.7 8.6 31.9 26.8 64.9 9,374 South Africa -0.3 -1.0 -30.1 62.2 55.0 85.8 9,362 Mexico -9.5 -4.0 -28.9 40.1 54.1 51.7 9,061 Russian Fed. -34.5 -17.7 -414.3 514.2 -52.4 -4.6 9,018 Romania 0.3 -7.8 -39.3 30.4 -4.9 -21.3 7,688 Bulgaria -2.4 -1.6 -8.0 12.1 -3.6 -3.4 7,577 Thailand -9.7 -0.8 49.7 35.7 16.8 91.8 7,453 Brazil -27.2 2.0 22.3 27.6 43.9 68.6 7,406 Tunisia -1.3 -0.1 -2.5 6.4 2.2 4.9 7,170 Turkey -8.4 2.1 9.7 41.9 27.7 73.0 6,951 Dominican Rep. 0.3 -0.2 3.2 5.1 2.1 10.5 6,786 Iran, Islamic Rep. -27.3 1.3 35.2 107.5 36.6 153.1 6,738 of Kazakhstan 3.3 2.9 -59.9 84.3 -12.8 17.8 6,504 Belarus -3.8 -0.1 -34.6 31.0 -2.8 -10.3 6,425 Colombia -2.1 -0.5 -7.1 2.3 9.3 1.9 6,275 Algeria -6.7 -0.3 -30.4 18.2 12.4 -6.9 6,058 Ukraine -11.3 -16.6 -93.0 84.0 -38.8 -75.7 5,949 Venezuela, R. B. 2.2 -0.2 8.9 -12.3 25.4 24.1 5,457 de China -108.3 -35.8 -1,155.4 2,917.7 293.0 1,911.2 5,441 Peru -0.1 -1.7 -2.6 4.6 4.1 4.3 5,122 Philippines -0.3 -1.2 -2.1 13.2 12.5 22.0 4,431 Morocco -0.9 -0.9 -4.6 3.9 4.4 1.9 3,875 Egypt, Arab Rep. -11.4 2.6 5.8 29.8 22.9 49.7 3,747 of Ecuador 0.6 0.2 -1.4 2.3 3.1 4.7 3,740 Azerbaijan -1.7 -1.1 -31.4 21.8 3.4 -9.0 3,551 Syrian Arab Rep. -1.2 0.7 -2.3 2.6 12.0 11.9 3,304 Indonesia -1.6 -3.3 30.1 40.6 33.6 99.4 3,245 India -31.7 20.0 -152.4 394.4 153.2 383.6 2,831 Vietnam -5.9 1.4 5.5 22.9 5.9 29.8 2,520 Pakistan -3.7 1.0 -10.4 13.3 21.7 22.0 1,969 Angola 5.4 0.2 -2.2 5.7 3.3 12.4 1,772 Bangladesh 0.2 0.2 4.2 8.5 5.5 18.7 1,756 Uzbekistan -3.9 1.2 -14.7 24.1 16.7 23.3 1,712 Nigeria -29.8 -0.5 -6.6 13.3 22.7 -1.0 959 Aggregate -156.4 150.4 -3,389.3 5,735.0 2,664.6 5,004.3 9,099 Source: World Bank calculations. Note: See the discussion of equation (5) in chapter 2 for definitions of the decomposition factors. One useful way to summarize the decomposition been expected to moderate emissions, and the three analysis is to compare the impacts from income and other factors, which had the potential through policy population growth, which are virtually always positive,4 interventions to reduce emissions. To this end, an and are outside direct interventions that might have "offsetting" coefficient is defined: 24 Environment Department Papers Emissions Levels and Decomposition of Emissions Changes Between 4 and 2004 Offsetting coefficient = - [ sum of changes from emissions per unit of fossil fuel, fossil fuel consumption relative to total energy consumption, and energy intensity of GDP] / [changes from GDP per capita and population] (7) A score of 100 percent indicates that the total increase For the aggregate of 70 countries, 40 percent of in emissions attributed to GDP per capita growth and the potential growth in emissions from GDP and population growth was exactly offset by improvements population growth was offset by the three factors. For in the three other factors. A negative score indicates individual countries, the offsetting coefficients, ranked that emissions increased faster than would have been by size of coefficient, are shown in table 6. accounted for by the growth of GDP and population. Table 6. Offsetting Coefficients for Decomposition of Emissions, 1994­2004 Country Offsetting coefficient Country Offsetting coefficient Ukraine 267.4 Belgium 32.0 Romania 183.6 Trinidad and Tobago 30.4 Denmark 169.1 India 30.0 Bulgaria 140.3 Korea, Rep. of 29.9 Belarus 136.4 South Africa 26.8 Azerbaijan 135.5 Netherlands 21.7 Czech Republic 124.4 Bahrain 21.5 Poland 123.7 Syrian Arab Rep. 19.0 Algeria 122.6 Philippines 14.2 Slovak Republic 113.5 Ecuador 12.4 Hungary 109.5 Australia 10.3 Germany 103.9 Singapore 8.8 Nigeria 102.7 Austria 6.0 Russian Fed. 101.0 Egypt, Arab Rep. of 5.7 Sweden 100.4 Brazil 4.0 United Kingdom 92.4 Israel 3.5 Colombia 83.6 Vietnam -3.3 Finland 80.7 Turkey -4.9 Morocco 77.3 Iran, Islamic Rep. of -6.3 Kazakhstan 75.0 Malaysia -10.5 United States 62.1 Chile -16.2 Switzerland 56.7 Portugal -18.6 Croatia 51.5 Norway -19.5 Peru 50.6 Italy -24.4 Ireland 50.2 Spain -25.4 Canada 45.8 Oman -26.9 France 45.5 Japan -31.9 Mexico 45.1 Bangladesh -33.2 Tunisia 43.8 Indonesia -33.9 Uzbekistan 42.8 Angola -38.6 China 40.5 Dominican Republic -45.6 New Zealand 37.9 Thailand -74.6 Pakistan 37.1 Venezuela, R. B. de -83.8 Greece 36.5 Argentina -89.8 United Arab Emirates 35.7 Saudi Arabia -102.8 Source: World Bank calculations. Climate Change Series 2 Growth and CO2 Emissions -- How do Different Countries Fare? The table of offsetting coefficients highlights a number decomposition for the group of countries with of features: negative offsetting are shown in figure 5. · Among countries with the largest absolute emis- · The range of offsetting experienced during the sions, the United States was able to offset a sizeable period was wide, with 15 countries more than fraction of its growth in emissions but, given the fully offsetting the potential increase in emissions size of the economy, the increment was still large. from income and population growth. A group of By contrast, Japan, with relatively low income 19 countries experienced "negative offsetting" in and population growth, experienced negative that movements in the three factors (emissions per offsetting, but the total increment was relatively unit of fossil fuel, fossil fuel consumption relative small. Russia fully offset the growth in emissions to total energy consumption, and energy intensity related to income and population but, given that of GDP) added to the increase in emissions from population decline and massive changes in sectoral income and population growth. structure and energy efficiency were stimulated by · The group of countries with full offsetting included the political changes during this period, a similar several Eastern European and former Soviet performance may be difficult to maintain in the Union countries, where massive changes in sector coming decade. China and India, while enjoying structure may have been responsible for shifting rapid growth during the decade, were also able to the economies to lower energy intensity. However, offset significant portions of the potential growth it also included Denmark and Sweden, where in emissions. market forces alone would have been unlikely to · Looking across the list of countries as a whole, it is bring about large-scale sectoral adjustment and noticeable that the offsetting coefficient is not cor- improvements in energy efficiency. The detailed related with income levels. Although high-income routes by which Denmark and Sweden achieved countries may have been best able to carry out this performance may be instructive for other high- policies that would have reduced emissions growth, income countries considering how to reduce the it does not appear that during the decade studied total level of their CO2 emissions. The group also there was a systematic worldwide link between included Algeria and Nigeria, both of which are the degree of offsetting and the level of per capita major hydrocarbon producers. Algeria experienced income. a major decline in energy intensity, while Nigeria experienced a large drop in emissions per unit of The Changing Pattern of Decomposition fossil fuel consumed. Between 1994 and 2004 · The group of countries with negative offsetting are mixed in type, including high-income developed Many economies experienced important changes countries such as Japan, Italy, and Spain; several oil during the decade analyzed, particularly those of producers (including Saudi Arabia, Indonesia, and Eastern Europe and the former Soviet Union where República Bolivariana de Venezuela) that may have the first years overlapped or immediately succeeded adopted very energy-intensive growth strategies; important political changes. In addition, the increasing and some low-income countries (Dominican price of oil and other fossil fuels, as well as heightened Republic, Bangladesh). There does not appear to awareness of global warming, may have led more be a common link between these countries, except recently to intensification of policies to slow the growth for the group of oil producers, and country-level of emissions. By splitting the data into two subperiods studies would be needed to understand the causes of equal duration, some important shifts in behavior of their performance. The components of the can be detected. The offsetting coefficient and changes 2 Environment Department Papers Emissions Levels and Decomposition of Emissions Changes Between 4 and 2004 Figure 5. Decomposition for"Negative Offsetting"Countries, 1994­2004 Offset | Contribution Vietnam Turkey Iran Malaysia Chile Portugal Norway Italy Spain Oman Japan Bangladesh Indonesia Angola Dominican Republic Thailand Venezuela Argentina Saudi Arabia -30 0 30 60 90 120 Emissions Intensity of Fossil Fuels Share of Fossil Fuels Energy Intensity GDP Per Capita Population Source: World Bank calculations. Climate Change Series 2 Growth and CO2 Emissions -- How do Different Countries Fare? in emissions for the two subperiods, ranked by the decompositions during the two subperiods are given in magnitude of offsetting coefficients for the whole appendix 2. period, are shown in table 7. The detailed tables for the Table 7. Offsetting Coefficients and Changes in Emissions, 1994­99 and 1999­2004 Offsetting coefficient Offsetting coefficient Country E (1994­99) (1994­99) E (1999­2004) (1999­2004) Ukraine -115.4 -6.1 39.7 71.1 Romania -25.3 -1,461.9 4.0 83.5 Denmark -7.0 185.2 -1.6 138.7 Bulgaria -6.3 -101.4 2.9 73.6 Belarus -7.8 191.0 -2.4 113.1 Azerbaijan 2.8 46.0 -11.8 155.2 Czech Rep. -15.9 249.5 8.8 47.5 Poland 4.7 94.9 -36.7 177.1 Algeria 0.3 97.6 -7.2 142.7 Slovak Rep. -2.9 133.4 0.6 91.8 Hungary -0.9 109.4 -1.2 109.5 Germany -35.6 148.5 30.7 38.4 Nigeria -10.4 194.5 9.5 58.9 Russian Fed. -153.1 -65.1 148.4 72.2 Sweden -1.1 111.8 1.0 86.4 United -17.1 120.6 29.5 61.9 Kingdom Colombia 3.9 -3.3 -1.9 123.5 Finland -9.9 188.9 13.9 -83.4 Morocco 3.0 -6.1 -1.1 118.6 Kazakhstan -23.3 -190.9 41.2 44.8 United States 419.2 59.8 255.2 65.5 Switzerland 3.2 -7.4 -0.7 122.7 Croatia 2.3 40.3 1.7 62.2 Peru 3.4 22.4 0.9 79.7 Ireland 9.1 41.4 4.3 64.1 Canada 65.9 30.0 29.5 65.1 France 41.5 9.2 4.8 88.4 Mexico 26.7 43.9 25.0 46.3 Tunisia 3.0 31.3 1.9 57.9 Uzbekistan 5.2 63.6 18.1 30.1 China 105.4 91.5 1,805.9 -10.2 New Zealand 3.2 32.8 3.7 42.0 2 Environment Department Papers Emissions Levels and Decomposition of Emissions Changes Between 4 and 2004 Table 7. Offsetting Coefficients and Changes in Emissions, 1994­99 and 1999­2004 (continued) Offsetting Offsetting coefficients coefficients Country E (1994­99) (1994­99) E (1999­2004) (1999­2004) Pakistan 18.4 -18.5 3.6 83.1 Greece 10.8 17.5 11.4 49.0 United Arab Emirates 23.8 19.3 23.6 48.6 Belgium 10.5 32.2 9.6 31.7 Trinidad and Tobago 4.5 15.0 6.4 39.8 India 205.6 21.4 178.0 39.4 Korea, Rep. of 72.2 15.8 70.6 41.7 South Africa 24.7 45.4 61.1 13.4 Netherlands 15.8 61.5 30.6 -88.5 Bahrain 4.4 -24.2 2.6 55.1 Syrian Arab Rep. 9.4 -45.3 2.4 72.3 Philippines 16.0 -47.2 6.0 63.2 Ecuador 1.6 -128.7 3.2 34.7 Australia 71.8 -7.9 35.4 36.1 Singapore 22.8 14.2 25.6 1.8 Austria 6.7 14.8 6.3 -6.9 Egypt, Arab Rep. of 16.8 35.5 32.9 -32.9 Brazil 65.5 -100.1 3.1 92.9 Israel 11.4 -0.6 6.7 10.9 Vietnam 13.4 -9.2 16.4 2.8 Turkey 41.3 -35.7 31.8 23.0 Iran, Islamic Rep. of 66.1 -35.3 87.0 11.9 Malaysia 16.9 28.3 48.0 -47.6 Chile 22.5 -84.6 3.0 74.4 Portugal 16.8 -52.6 0.6 85.3 Norway 7.9 -4.0 7.5 -46.9 Italy 36.7 -0.7 48.1 -55.2 Spain 62.5 -32.4 66.0 -18.1 Oman 5.4 -91.3 2.9 29.2 Japan 61.4 -44.2 112.7 -25.5 Bangladesh 7.5 -34.6 11.2 -31.9 Indonesia 55.1 -248.1 44.3 30.5 Angola 6.1 -59.3 6.3 -15.1 Dominican Rep 4.4 -8.9 6.1 -125.4 Thailand 43.2 -355.2 48.6 -1.6 Venezuela, R. B. de 13.7 -174.6 10.4 -24.2 Argentina 23.5 -73.8 3.1 -704.4 Saudi Arabia 23.7 -61.7 103.1 -118.6 For a number of countries, the degree of offsetting related factors were making a further contribution is substantially different between the two periods to reducing emissions in that period. In the second indicating that significant structural changes took place subperiod, no country experienced negative GDP between the two subperiods: growth. The very large offsetting coefficient for Romania in the second subperiod was caused by · For a small group of countries (Ukraine, Romania, the very small size of the denominator, indicating Bulgaria, Russia, and Kazakhstan) the apparent the slight decrease in emissions that could be negative offsetting in the first subperiod occurred related to the change in GDP. because the net effect of GDP and population · For a large number of countries, similar offsetting changes reduced emissions, and the three policy- coefficients were experienced, indicating that Climate Change Series 2 Growth and CO2 Emissions -- How do Different Countries Fare? the balance of factors did not change markedly the first subperiod, partly thanks to high offsetting, but between the two periods. in the second subperiod experienced a very large growth · For a third group of countries, the offsetting coef- in emissions, when the effects of very rapid growth ficient increased substantially, indicating a move were slightly compounded by negative offsetting. Japan toward slowing down the growth or even reducing experienced modest increases in emissions during both the level of emissions relative to the growth of periods, partly related to negative offsetting in both. GDP and population. This group of countries Russia experienced a substantial decline in emissions includes a number of lower income countries that during the first period, but then experienced a sizeable experienced negative offsetting during the first growth in emissions during the second period, despite subperiod, but moved to partial positive offsetting a high degree of offsetting. Finally, India experienced in the second period. This latter group consists of similar growth of emissions during the two subperiods, Pakistan, Syria, the Philippines, Ecuador, Australia, with the slightly lower increase in the second period Brazil, Turkey, the Islamic Republic of Iran, and being associated with higher offsetting. Indonesia. · A group of 13 countries experienced negative offsetting in both subperiods. In addition to some Notes large oil producers (Norway, Angola, República 1. More detailed analysis of the inequality of CO2 Bolivariana de Venezuela, and Saudi Arabia), the emissions among countries is given by Heil and group includes some high-income, non-oil produc- Wodon (2000) and Duro and Padilla (2006). ers (Italy, Spain, and Japan). 2. The choice between using GDP at MER and PPP · Six countries moved from partial positive offsetting to calculate emissions intensity is discussed by in the first subperiod to negative offsetting in the Stern (2006) and by Holtsmark and Alfsen (2004). second subperiod, indicating that the policy-re- lated factors had ameliorated the effects of growth 3. The very large fluctuation in China's energy inten- earlier, but added to it in the latter period. This sity, which declined until 2002 but then started group includes Colombia, China, the Netherlands, to increase, has been the stimulus for a detailed Austria, Egypt, and Malaysia. analysis of changes in the country's economic structure. See Ma and Stern (2006); Liao, Fan, and The experience of countries with the largest absolute Wei (2007); and Hoffman and Labar (2007). emissions during the two subperiods was quite 4. The exceptions are Saudi Arabia, where per capita different. The United States experienced growth in income fell during the period, and Russia and emissions during both periods, but in both offset a other countries in the former Soviet Union and substantial fraction of the effects of the growth in Eastern Europe, where changed political boundar- income. China experienced little growth in emissions in ies led to declines in population. 0 Environment Department Papers 6 Conclusions T he report focuses on the 70 top ranking countries Brazil, India, and Mexico ranked much lower on based on CO2 emissions from fossil fuels in emissions intensity than on total emissions among the 2004, covering 95 percent of the global total group of 70 countries studied. Several middle-income CO2 emissions. The top 20 contributors to global countries were among those with the highest rankings CO2 emissions from fossil fuels were mainly drawn of emissions intensity. from higher income developed countries, but they also included several large developing economies. In Emissions per capita were positively but only addition, several large oil-producing countries were moderately correlated with GDP per capita and showed in the top 30. During the period 1994­2004, some no evidence of an eventual decline in emissions per 15 countries experienced a decline in emissions; this capita at higher per capita income (the Environmental group comprised mainly former Soviet Union and Kuznets Curve phenomenon). The global per capita measure of inequality of emissions was much lower Eastern European countries that had undergone major than that for the country level inequality, and was very structural change during the period. The percentage similar to the measure of income inequality. Some growth in emissions during the decade showed countries with large populations, such as India and considerable variation among countries, with some Indonesia, moved almost to the bottom of the ranking large emitters experiencing significant growth, and based on this measure. others relatively low growth. The degree of inequality between countries with respect to total CO2 emissions The decomposition analysis related the change is extremely high, with a Gini coefficient of 0.72. in emissions during the decade to changes in five factors (emissions per unit of fossil fuel, fossil fuel The intercountry distribution of emissions per unit of consumption relative to total energy consumption, GDP (emissions intensity) was measured using GDP at energy intensity of GDP, GDP per capita, and purchasing power parity and at market exchange rates. population). For the group of countries as a whole, The emissions intensity for most countries was very GDP per capita was the dominant variable linked to much lower using the former measure. The ranking the growth in total emissions, with population being of countries by emissions intensity also changed only one-half as important. However, the decrease in substantially relative to the ranking by total emissions. energy intensity was so large that it offset about 40 Emissions intensity was not systematically related to percent of the combined influence of these two factors. GDP per capita. The fossil fuel mix showed a small negative effect on the change in emissions, while the share of fossil fuels in The ranking of countries by emissions intensity total energy consumption contributed a small increase changed substantially relative to the ranking by total to the change in emissions. CO2 emissions. Among large developing economies, Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? The relative importance of the three factors that could in the first period, but much reduced offsetting in the be directly linked to policies to slow down the growth second. For the group as a whole there was no apparent of emissions (fossil fuel mix, share of fossil fuels in total correlation between the improvement in offsetting and energy, and energy intensity of GDP) to the positive the level of GDP, indicating that during the period effects of GDP growth is measured by the "offsetting" studied higher income countries had not performed coefficient. For the decade as a whole, 15 countries better in slowing the growth of emissions relative to more than offset the combined impacts of GDP and GDP. population growth. This group consisted mainly of former Soviet Union and Eastern European countries, However, several countries did experience improved but also included two Scandinavian countries, whose performance of emissions relative to GDP, suggesting policies toward emissions may give important clues that there need not be a negative trade-off between for other countries wanting to slow the growth of slowing the growth of emissions and maintaining high emissions. There were also 19 countries where the growth rates of the economy. Countries that were potential offsetting factors actually led to a further particularly successful in achieving high offsetting increase in emissions beyond that which would have could well serve as case studies for how this might be been expected from the growth of GDP ("negative achieved. However, the experience of several countries offsetting"). The group included several oil producers, also makes it clear that, without active policies to curb but also some non-oil-producing, high-income the emissions intensity of the economy, emissions developed countries. Two developing countries with can actually increase faster than GDP, even when the largest populations and CO2 emissions, China and GDP has reached a high level. Additional and more India, had offsetting coefficients near the global average detailed analysis could shed further light on the factors for the period. determining aggregate energy intensity for an economy, because this factor appears to have accounted for the When the data was split into two five-year subperiods it largest differences in performance between economies, become apparent that for some countries large changes once the growth of GDP is allowed for. In particular, had taken place in the course of the decade. For one the distinction between changes resulting from sectoral group of countries the offsetting coefficient increased composition and energy efficiency may be helpful in substantially; this group included several developing giving clues about the future course of emissions, and countries. An important group of 13 countries the possibilities of finding policies that could make a experienced negative offsetting in both subperiods, substantial difference to global CO2 emissions. indicating that this was a longer term trend; the group included some major oil producers, and some high- The magnitude of the decline in energy intensity income, non-oil-producing countries. Six countries in many countries suggests that further work moved from positive to negative offsetting between the to understand why this has happened would be two subperiods, indicating that they had actually seen important. In particular, a further decomposition into emissions rise faster relative to GDP. The experiences changes in sector structure (the shares of agriculture, of countries with the largest absolute emissions across manufacturing, and services) and changes in energy the two subperiods were quite different. The United intensity at sector levels would provide information on States and India both experienced substantial offsetting the extent to which changes in the sector mix that are throughout the entire period, while China and the related to economic growth and development have been Russian Federation both experienced high offsetting responsible for changes in overall emissions. 2 Environment Department Papers Appendix 1-- Decomposition of CO2 Emissions for India Comparing EIA and Government of India Data A s part of the India Low Carbon Growth Study, in the thermal content and in the emissions per physical a separate exercise on CO2 emissions data unit. Data on GDP at 2000 US$ in PPP continued was carried out using the most recent data to be taken from the World Development Indicators on fuel use and emissions provided by the (WDI; World Bank various years), as the sole source for government of India. These data can be taken as the comparable PPP data; population data continued to be most authoritative currently available, and provide a taken from UN Population Fund (UNFPA) as the most useful comparison with that published by the Energy widely used estimate of population between censuses. Information Administration (EIA) in June 2006, which Data from the two sources is shown to illustrate the was based on earlier published government sources. The sources of differences, and these are then used to data from the government of India takes account of the provide alternative decompositions. specific quality of the fuels consumed, which is reflected Consumption of Fossil Fuel Data in Physical Units Government Government EIA Government of India EIA Government of India Energy source 1994 Government 2004 Government Petroleum (thousand barrels a day) 1,413 EIA of1994 of1994 2,450 EIA of2004 of2004 2,419 India 2004 EIA India 1994 1,412 India Energy source 1994 EIA India Energy source Dry gas (billion cubic feet) Petroleum (thousand barrels a day) 1,413 1994 1,088 2,450 2004 1,185 2,419 2004 Petroleum (thousand barrels a day) 1,413 594 1,412 1,412 612 Coal (million short tons) Dry gas (billion cubic feet) 314 261 2,450 1,088 478 2,419 1,185 397 Dry gas (billion cubic feet) 594 612 Coal (million short tons) 594 612 1,088 1,185 Coal (million short tons) 314 314 261 261 478 478 397 397 Consumption of Energy in Thermal Units (quadrillion British thermal units) Government Government EIA Government of India EIA Government of India Energy source 1994 Government 2004 Government Petroleum EIA of1994 of2.88 India 5.02 EIA of2004 of2004 2004 4.92 India Energy source 1994 EIA 1994 India 2004 EIA India Energy source 1994 2.94 Dry gas Petroleum 0.62 2.88 1994 2004 Petroleum 0.68 2.94 1.13 5.02 1.26 4.92 Coal Dry gas 2.94 5.45 2.88 5.05 5.02 8.11 4.92 7.19 Dry gas 0.68 0.62 1.13 1.26 All primary energy including nonfossil fuels Coal 0.68 0.62 15.42 1.13 13.84 1.26 Coal 9.97 5.45 8.89 5.05 8.11 7.19 All primary energy including nonfossil fuels 5.45 5.05 15.42 8.11 13.84 7.19 All primary energy including nonfossil fuels 9.97 9.97 8.89 8.89 15.42 13.84 Emission of CO2 from Fossil Fuel Consumption (million metric tons) Government Government EIA Government of India EIA Government of India Energy source 1994 Government 2004 Government Petroleum EIA of1994 of1994 EIA of2004 of342 India 2004 EIA 2004 India 1994 200 India Energy source 1994 EIA India Energy source 1994 193 2004 306 Petroleum 342 2004 Dry gas Petroleum 193 193 36 200 200 33 306 306 64 342 67 Dry gas Coal Dry gas 497 36 36 461 33 33 741 64 64 656 67 67 Coal All fossil fuels including gas flaring Coal 497 729 497 461 693 461 1,113 741 741 1,064 656 656 All fossil fuels including gas flaring All fossil fuels including gas flaring 729 729 693 693 1,113 1,113 1,064 1,064 WDI 1994 WDI 2004 GDP WDI 1994 WDI 1994 323 WDI 2004 WDI 2004 590 Climate Change Series GDP GDP 323 323 590 590 EIA Government of India EIA Government of India Energy source 1994 Government 2004 Government Petroleum EIA of1994 of1994 200 India EIA of2004 of342 India Energy source 1994 EIA 193 India 2004 EIA 306 2004 India Energy source Petroleum Dry gas 1994 1994 2004 342 2004 Petroleum 193 193 36 200 200 33 306 306 64 342 67 Dry gas Coal Dry gas 497 36 36 461 33 33 741 64 64 656 67 67 Growth and CO2 Emissions -- How do Different Countries Fare? All fossil fuels including gas flaring Coal Coal 729 497 497 693 461 461 1,113 741 741 1,064 656 656 All fossil fuels including gas flaring All fossil fuels including gas flaring 729 729 693 693 1,113 1,113 1,064 1,064 Gross Domestic Product (billionWDI US$ at PPP) WDI 1994 20002004 GDP WDI 1994 323 WDI 2004 590 GDP WDI 1994 WDI 2004 GDP 323 323 590 590 Population (millions) UNFPA 1994 UNFPA 2004 Population UNFPA 1994 918 UNFPA 2004 1,087 Population UNFPA 1994 UNFPA 2004 Population 918 918 1,087 1,087 Decomposition of Changes in Emissions Between 1994 and 2004 (million metric tons) Source Ceff Seff Ieff Geff Peff E EIA data Source -31.7 20.0 -152.4 394.4 153.2 383.6 Source Government of -15.9 Ceff Seff 376.0 Geff 146.0 Peff 371.0 E EIA data -31.7 Ceff 20.0 S3.9 eff -139.1 -152.4 IIeff eff 394.4 Geff 153.2 Peff 383.6 E EIA data India data Government of -31.7 20.0 -152.4 394.4 153.2 383.6 Government of -15.9 3.9 -139.1 376.0 146.0 371.0 India data -15.9 3.9 -139.1 376.0 146.0 371.0 India data The comparison of the EIA data and the recent in 2004 are about 5 percent lower than those given by government of India data shows some important the EIA. Thus, the change in emissions to be used in differences, particularly with respect to the the decomposition analysis is slightly lower. The shares consumption of coal, where the government figures of the five factors do change but the main conclusions show lower total consumption and lower emissions. As about the relative importance of the different factors are a result, the total levels of CO2 emissions in 1994 and similar. 4 Environment Department Papers Appendix 2 -- Decomposition of Emissions During Subperiods Decomposition of Emissions Between 1994 and 1999 (million metric tons) Country Ceff Seff Ieff Geff Peff E Norway 0.0 3.3 -3.1 6.3 1.3 7.9 United States -3.2 -19.2 -601.2 753.6 289.4 419.2 Ireland -0.6 0.1 -5.8 14.0 1.5 9.1 Switzerland -0.3 0.9 -0.3 2.0 1.0 3.2 Austria 0.3 -0.3 -1.2 7.2 0.7 6.7 United Kingdom -26.0 -4.1 -69.5 73.1 9.5 -17.1 Denmark -5.4 -2.6 -7.3 7.0 1.3 -7.0 Canada -3.4 31.5 -56.4 69.6 24.6 65.9 Netherlands 7.1 -2.5 -29.8 34.3 6.6 15.8 Belgium -5.8 -0.9 1.7 13.3 2.2 10.5 Sweden -0.4 -3.1 -6.6 8.6 0.5 -1.1 Finland -6.3 -4.8 -10.0 10.3 0.8 -9.9 Australia 4.5 2.3 -1.6 47.3 19.2 71.8 Japan -10.9 -23.7 53.5 28.0 14.5 61.4 France -0.5 7.5 -11.2 38.4 7.3 41.5 Germany -28.1 -13.3 -67.6 63.5 9.9 -35.6 Italy -10.4 1.0 9.6 33.1 3.3 36.7 Singapore 0.7 0.0 -4.4 12.6 13.9 22.8 Spain -3.4 10.3 8.4 43.1 4.1 62.5 Israel -0.4 0.2 0.3 4.1 7.2 11.4 New Zealand -0.2 1.7 -3.1 3.2 1.5 3.2 United Arab 0.3 0.0 -6.0 0.5 29.0 23.8 Emirates Greece -1.6 -1.2 0.5 10.0 3.0 10.8 Korea, Rep. of -5.8 -16.5 8.8 69.6 16.1 72.2 Portugal -0.1 4.1 1.8 10.1 0.9 16.8 Bahrain 0.3 0.0 0.6 0.9 2.7 4.4 Czech Rep. -6.9 -0.3 -19.3 11.2 -0.5 -15.9 Hungary -2.3 0.3 -8.2 9.8 -0.5 -0.9 Oman 1.2 0.0 1.4 0.6 2.3 5.4 Slovak Rep. -4.4 0.4 -7.4 8.2 0.4 -2.9 Saudi Arabia -17.7 0.0 26.7 -19.5 34.1 23.7 Trinidad and -1.4 0.0 0.6 4.7 0.5 4.5 Tobago Poland -9.2 0.2 -79.5 92.2 1.0 4.7 Argentina -2.3 6.4 5.9 5.9 7.6 23.5 Croatia 0.7 -0.3 -1.9 4.4 -0.5 2.3 Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? Decomposition of Emissions Between 1994 and 1999 (million metric tons) (continued) Country Ceff Seff Ieff Geff Peff E Chile -0.5 6.6 4.2 8.8 3.4 22.5 Malaysia -2.8 0.3 -4.2 11.4 12.1 16.9 South Africa -0.6 -3.1 -16.9 11.4 33.9 24.7 Mexico 5.2 -10.1 -16.1 19.4 28.2 26.7 Russian Fed. -8.6 -17.3 -34.5 -77.3 -15.4 -153.1 Romania -0.1 -9.0 -14.6 1.1 -2.8 -25.3 Bulgaria 0.1 -2.7 -0.7 -1.3 -1.9 -6.3 Thailand -0.8 0.9 33.5 1.5 8.0 43.2 Brazil -4.7 4.2 33.3 10.3 22.4 65.5 Tunisia -0.4 0.0 -0.9 3.2 1.2 3.0 Turkey -3.5 2.9 11.5 16.5 13.9 41.3 Dominican Rep. -0.1 -0.2 0.6 3.2 0.9 4.4 Iran, Islamic Rep. of -13.2 3.0 27.4 29.8 19.1 66.1 Kazakhstan 2.0 2.1 -19.5 0.1 -8.1 -23.3 Belarus -1.8 -0.8 -13.8 9.8 -1.2 -7.8 Colombia -0.9 0.8 0.2 -1.3 5.0 3.9 Algeria -0.3 -0.3 -13.3 7.6 6.7 0.3 Ukraine -9.1 -16.1 18.5 -93.0 -15.7 -115.4 Venezuela, R. B. de 1.0 -1.6 9.3 -7.8 12.8 13.7 China -102.7 -23.9 -1,010.1 1,111.9 130.2 105.4 Peru 0.2 0.0 -1.2 2.2 2.1 3.4 Philippines 0.7 -2.6 7.0 4.6 6.3 16.0 Morocco 0.1 0.1 0.0 0.5 2.3 3.0 Egypt, Arab Rep. of -2.5 -1.3 -5.4 16.1 9.9 16.8 Ecuador 0.4 -0.1 0.5 -0.8 1.5 1.6 Azerbaijan 17.3 -0.3 -19.4 2.8 2.4 2.8 Syrian Arab Rep. -0.1 0.4 2.6 0.5 6.0 9.4 Indonesia 4.0 -2.9 38.1 -0.2 16.0 55.1 India -13.7 13.0 -55.2 187.4 74.0 205.6 Vietnam -3.2 0.7 3.6 9.7 2.6 13.4 Pakistan -0.1 4.9 -1.9 4.0 11.6 18.4 Angola 4.6 0.2 -2.5 2.6 1.2 6.1 Bangladesh 0.1 0.2 1.6 3.2 2.4 7.5 Uzbekistan -2.2 1.1 -7.9 6.2 8.1 5.2 Nigeria -19.0 0.0 -2.5 -0.3 11.3 -10.4 Source: World Bank calculations. Environment Department Papers Appendix 2 -- Decomposition of Emissions During Subperiods Decomposition of Emissions Between 1999 and 2004 (million metric tons) Country Ceff Seff Ieff Geff Peff E Norway -2.7 8.8 -3.7 3.8 1.3 7.5 United States 9.8 35.6 -531.0 457.1 283.7 255.2 Ireland -0.3 -0.3 -7.0 8.5 3.4 4.3 Switzerland 0.1 0.4 -4.2 2.4 0.6 -0.7 Austria -0.1 2.5 -2.0 5.3 0.7 6.3 United Kingdom 1.2 12.3 -61.4 67.7 9.7 29.5 Denmark 3.8 -3.4 -6.0 3.0 1.0 -1.6 Canada -6.3 6.6 -55.3 55.9 28.6 29.5 Netherlands 3.7 -0.2 10.9 9.8 6.5 30.6 Belgium -0.8 0.8 -4.4 12.2 1.7 9.6 Sweden -0.7 0.9 -6.5 6.5 0.9 1.0 Finland 5.4 2.7 -1.8 6.9 0.7 13.9 Australia 1.8 1.0 -22.7 34.5 20.8 35.4 Japan 54.7 25.2 -57.0 78.7 11.0 112.7 France -2.1 -12.3 -21.9 33.0 8.1 4.8 Germany 6.3 -9.9 -15.5 45.9 3.9 30.7 Italy 4.6 2.2 10.3 27.9 3.2 48.1 Singapore -5.5 0.0 5.0 16.2 9.9 25.6 Spain -1.2 -3.1 14.5 38.4 17.5 66.0 Israel -1.2 0.0 0.4 1.0 6.5 6.7 New Zealand 2.0 -0.6 -4.1 4.5 1.9 3.7 United Arab Emirates -5.6 0.0 -16.7 1.7 44.2 23.6 Greece -1.4 -0.5 -9.1 20.8 1.6 11.4 Korea, Rep. of -5.2 -4.3 -41.0 109.8 11.4 70.6 Portugal -1.2 -3.3 1.0 2.6 1.6 0.6 Bahrain -0.6 0.0 -2.5 3.9 1.8 2.6 Czech Rep. 0.3 -5.4 -2.9 17.3 -0.5 8.8 Hungary -1.8 -0.3 -11.8 13.4 -0.7 -1.2 Oman -1.9 0.0 0.7 2.9 1.2 2.9 Slovak Rep. 1.6 -1.8 -6.8 7.6 0.0 0.6 Saudi Arabia 15.1 0.0 40.9 4.7 42.5 103.1 Trinidad and Tobago -5.2 0.0 0.9 10.2 0.5 6.4 Poland -11.4 0.8 -73.7 48.4 -0.8 -36.7 Argentina -2.5 -4.0 9.2 -6.6 7.0 3.1 Croatia 0.1 0.8 -3.7 4.4 0.0 1.7 Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? Decomposition of Emissions Between 1999 and 2004 (million metric tons) (continued) Country Ceff Seff Ieff Geff Peff E Chile -3.9 -4.3 -0.7 8.4 3.5 3.0 Malaysia -1.9 2.5 14.8 19.5 13.0 48.0 South Africa 0.3 2.4 -12.2 51.6 18.9 61.1 Mexico -15.5 6.7 -12.8 20.8 25.8 25.0 Russian Fed. -24.3 0.4 -360.8 567.7 -34.6 148.4 Romania 0.3 1.2 -21.6 25.8 -1.8 4.0 Bulgaria -2.4 1.1 -6.8 12.5 -1.6 2.9 Thailand -10.1 -2.1 13.0 39.0 8.8 48.6 Brazil -24.9 -2.5 -12.5 19.2 23.7 3.1 Tunisia -0.9 -0.1 -1.6 3.3 1.1 1.9 Turkey -5.2 -1.1 -3.2 27.1 14.2 31.8 Dominican Rep. 0.5 -0.1 2.9 1.4 1.3 6.1 Iran, Islamic Rep. of -13.8 -2.4 4.5 82.2 16.6 87.0 Kazakhstan 0.9 0.4 -34.7 77.8 -3.2 41.2 Belarus -2.0 0.7 -19.7 20.1 -1.5 -2.4 Colombia -1.3 -1.3 -7.6 3.7 4.6 -1.9 Algeria -6.4 -0.1 -17.7 11.0 6.0 -7.2 Ukraine -1.5 0.3 -96.5 156.4 -19.0 39.7 Venezuela, R. B. de 1.2 1.6 -0.8 -4.4 12.8 10.4 China 24.4 -5.1 148.2 1,510.9 127.4 1,805.9 Peru -0.4 -1.9 -1.4 2.5 2.1 0.9 Philippines -1.2 1.7 -10.8 9.5 6.7 6.0 Morocco -1.0 -1.1 -4.9 3.6 2.2 -1.1 Egypt, Arab Rep. of -9.2 4.4 13.0 12.3 12.5 32.9 Ecuador 0.1 0.3 -2.1 3.4 1.5 3.2 Azerbaijan -17.4 -0.9 -14.9 20.0 1.4 -11.8 Syrian Arab Rep. -1.2 0.3 -5.5 2.4 6.5 2.4 Indonesia -6.7 -0.2 -12.5 45.6 18.2 44.3 India -18.7 6.5 -103.4 212.6 81.0 178.0 Vietnam -2.5 0.7 1.3 13.5 3.4 16.4 Pakistan -3.9 -4.4 -9.3 10.2 11.0 3.6 Angola -0.3 0.0 1.1 3.2 2.3 6.3 Bangladesh 0.0 0.0 2.6 5.4 3.1 11.2 Uzbekistan -1.6 0.0 -6.2 17.8 8.1 18.1 Nigeria -9.3 -0.5 -3.8 12.9 10.2 9.5 Source: World Bank calculations. Environment Department Papers References Ang, B. W. 2004. "Decomposition Analysis for China Working Paper Series No. 6, World Bank, Policymaking in Energy: Which is the Preferred Washington, DC. Method?" Energy Policy 32: 1131­39. Holtsmark, B. J., and K. H. Alfsen. 2004. "PPP ------. 2005. "The LMDI Approach to Correction of the IPCC Emission Scenarios ­ Does Decomposition Analysis: A Practical Guide." Energy It Matter?" Discussion Paper No. 366, Research Policy 33: 867­71. Department, Statistics Norway, Oslo. Ang, B. W., and F. Q. Zhang. 1999. "Inter-Regional IPCC (Intergovernmental Panel on Climate Change). Comparisons of Energy-Related CO2 Emissions Using 2007. "Climate Change 2007: Mitigation of Climate the Decomposition Technique." Energy 24 (4): 297­ Change." Working Group III Contribution to the 304. IPCC Fourth Assessment Report: Summary for Policymakers. UNEP. Duro, J. A., and E. Padilla. 2006. "International Inequalities in per Capita CO2 Emissions: A Kaya, Y. 1990. "Impact of Carbon Dioxide Emission Decomposition Methodology by Kaya Factors." Energy Control on GNP Growth: Interpretation of Proposed Economics 28: 170­87. Scenarios." Paper presented to IPCC Energy and Industry Subgroup, Response Strategies Working EIA (Energy Information Administration). 2007. Group. "International data." http://www.eia.doe.gov/emeu/ international/contents.html. Lee, K., and W. Oh. 2006. "Analysis of CO2 Emissions in APEC Countries: A Time-Series and a Cross- Ebohon, O. J., and A. J. Ikeme. 2006. "Decomposition Sectional Decomposition Using the Log Mean Divisia Analysis of CO2 Emission Intensity Between Oil- Method." Energy Policy 34: 2779­87. Producing and Non-Oil Producing Sub-Saharan African Liao, H., Y. Fan, and Y. Wei. 2007. "What Induced Countries." Energy Policy 34: 3599­3611. China's Energy Intensity to Fluctuate: 1997­2006?" Heil, M. T., and Q. T. Wodon. 2000. "Future Energy Policy 35 (9): 4640­4649. Inequality in CO2 Emissions and the Impact of Lise, W. 2006. "Decomposition of CO2 Emissions over Abatement Proposals." Environmental and Resource 1980­2003 in Turkey." Energy Policy 34: 1841­52. Economics 17: 163­81. Liu, N., and B. Ang. 2007. "Factors Shaping Aggregate Herzog, T., K. A. Baumert, and J. Pershing. 2006. Energy Intensity Trend for Industry: Energy Intensity Target Intensity. An Analysis of Greenhouse Gas versus Product Mix." Energy Economics 29: 609­35. Intensity Targets. Washington, DC: World Resources Institute. Ma, C., and D. I. Stern. 2006. "China's Changing Energy Intensity Trend: A Decomposition Analysis." Hoffman, B., and K. Labar. 2007. "Structural Change Working Papers in Economics No. 0615, Rensselaer and Energy Use: Evidence from China's Provinces." Polytechnic Institute, Troy, NY. Climate Change Series Growth and CO2 Emissions -- How do Different Countries Fare? Nag, B., and J. K. Parikh. 2005. "Carbon Emission Viguier, L. 1999. "Emissions of SO2, NOx and CO2 Coefficient of Power Consumption in India: Baseline in Transition Economies: Emissions Inventories and Determination from the Demand Side." Energy Policy Divisia Index Analysis." The Energy Journal 20: 59­87. 33: 777­86. Wang, C., J. Chen, and J. Zou. 2005. "Decomposition Paul, S., and R. N. Bhattacharya. 2004. "CO2 of Energy-Related CO2 Emission in China: 1957­ 2000." Energy 30: 73­83. Emission from Energy Use in India: A Decomposition Analysis." Energy Policy 32: 585­93. World Bank. various years. World Development Indicators. Washington, DC: World Bank. Stern, D. I. 2003. "The Environmental Kuznets Curve." Internet Encyclopaedia of Ecological Economics. WRI (World Resources Institute). 2006. Climate http://www.ecoeco.org/pdf/stern.pdf. Analysis Indicators Tool. Version 3.0. Washington, DC. http://cait.wri.org. Stern, N. 2006. The Economics of Climate Change: The Stern Review. Cambridge University Press. 40 Environment Department Papers Environment Department T H E W O R L D B A N K 1818 H Street, NW Washington, D.C. 20433 Telephone: 202-473-3641 Facsimile: 202-477-0565