Assessment of Global Mitigation Progress: A Decomposition of CO 2 Emissions for the World s Top Emitting Countries

Transcription

1 1 Assessment of Global Mitigation Progress: A Decomposition of CO 2 Emissions for the World s Top Emitting Countries by Dirk C. Böhm, Robert Bosch GmbH / University of Hohenheim, Langemarckstr. 9, D Rastatt, Germany Phone , dirkcboehm@web.de Abstract Although declining global CO 2 emissions per unit of GDP could be observed within the last decades, strong economic growth especially in China and India has led to a worrying rise in global CO 2 emissions in the last years. This work provides a decomposition of the change in carbon dioxide emissions in order to analyze the relationship between emission growth and changes in underlying factors using the LMDI1 method. It covers the biggest carbon dioxide emitting countries and regions that together account for over 80% of total emissions worldwide in the period from 1971 to The results show that GDP growth is by far the biggest contributor to global emissions followed by an increasing population, while decreasing energy intensity was and still is the most important factor to reduce emissions. 1. Introduction and Literature Review According to the International Panel on Climate Change Fourth Assessment report (IPCC 2007), most of the observed rise in global average temperatures since the mid-20th century is very likely 1 due to the increase in anthropogenic GHG concentrations. The combustion of fossil fuels is the largest contributor to greenhouse gas (GHG) emissions and carbon dioxide (CO 2 ) is responsible for about 95% of the energy-related emissions. Thus fossil fuel combustion is the single largest human influence on climate. While emissions have doubled in the period between 1971 and 2005, real GDP has reached three times the value of the base year (see fig. 1). But although declining global CO 2 emissions per unit of gross domestic product (GDP) could be observed, strong economic growth especially in China and India has led to a worrying rise in global CO 2 emissions in the last years. 1 > 90% probability

2 =100 CO2 Emissions GDP Fig. 1: World CO 2 emissions from fuel combustion and real GDP 2 For analytic purposes emissions can be represented by a function of several contributing factors. For example, emissions at a national level can be expressed as a function of population, income (GDP per capita), and intensity (emissions per unit of GDP): CO Emissions=Population (1) In this case if a country s population and emission intensity remain constant while its per-capita income increases, emissions likewise will rise. The CO 2 intensity itself can be expressed as a function of energy intensity (energy per unit of GDP) and fuel mix (emissions per unit of energy): = (2) The term energy intensity includes economic efficiency, energy conservation, and the overall economic structure, whereas fuel mix represents the carbon intensity of the fuels that an economy uses to produce energy. 2 world GDP in 2000-USD using Purchasing Power Parities (PPP)

3 3 By combining equations (1) and (2) CO 2 emissions can be formulated as a function known as the Kaya identity 3 : CO Emissions=Population (3) These four factors can have exacerbating or mitigating effects on total emissions. For example, relatively small increases in income and population can result in huge increases in total emissions. On the other hand, improved energy intensity or a change in the fuel mix towards lower carbon emitting fuels can offset large increases in income growth (Herzog, Baumert et al. 2006). In general, historical changes in economic, environmental or other socio-economic indicators can be analyzed by assessing the underlying forces or determinants that are responsible for these changes. This can be done with two different methods, the structural decomposition analysis and the index decomposition analysis. The main difference between the structural and the index decomposition method is that the former uses input-output tables while the latter uses only aggregate sector information (Hoekstra and van der Bergh 2003). As index decomposition analysis requires less data which in addition is often available on an annual basis, it can be conducted to study different time periods and it allows cross-country comparisons. Usually data from two periods are used to examine which determinant changes have contributed most to a change in the indicator. Furthermore, the index decomposition technique can be divided into methods linked to the Laspeyres index as well as methods linked to the Divisia index. Methods linked to the Laspeyres index are based on a percentage change and the impact of a determinant is computed through letting that determinant to change while holding all the others at their base year values. In contrast, the Divisia index methods use a weighted sum of logarithmic growth rates, where the weights are the determinants shares in the total value. 3 The Kaya identity was developed by the Japanese energy economist Yoichi Kaya in 1992 and since then plays a major role in the development of emission scenarios in the IPCC reports.

4 4 Decomposition analysis Index decomposition analysis (IDA) structural decomposition analysis (SDA) methods linked to Divisia index methods linked to Laspeyres index Multiplicative decomposition Additive decomposition Multiplicative decomposition Additive decomposition LMDI 1 LMDI AMDI AMDI Fig. 2: Methods in decomposition analysis As exhibited in figure 2, decomposition can be performed multiplicatively, where the ratio change of an aggregate is decomposed, or additively, where the difference change is decomposed. Howarth et al. (1991) compared the results between the Laspeyres index and the Divisia index in an investigation of manufacturing energy use in eight OECD countries between 1973 and 1987 and found only minor differences between the two techniques. Ang and Lee (1994) looked at five different methods and found that the adaptive weighting and the simple average Divisia index methods tended to yield smaller residuals in decomposition. A refined Divisia index method was introduced by Ang et al. (1998). This logarithmic mean Divisia index (LMDI) approach was able to get perfect decomposition, handle the zero values in the dataset and study the decomposition of a differential change. Sun (1998) presented a complete decomposition model to solve residual terms with a decomposition of the changes in world energy consumption. Ang and Liu (2001) then presented the log-mean Divisia method I (LMDI I), which had the desirable characteristics of perfect decomposition and consistency in aggregation. According to Ang (2004) the desirability of a decomposition method can be evaluated by four factors: (a) the theoretical foundation, (b) the adaptability, (c) the ease of use and (d) the ease

5 5 of result interpretation. On this basis he recommends the log mean Divisia index methods LMDI 1 (additive or multiplicative) as the most preferred methods for empirical studies. The most common application areas for the index decomposition analysis are energy demand and supply analysis as well as energy-related greenhouse gas (GHG) emissions. In empirical papers covering energy demand and supply issues, the relative contributions of the impacts of structural change and energy intensity change are studied. Industrial energy demand is the most prominent area, but also demand analysis for the total economy and other sectors, such as transport or residential are conducted. These studies analyze the relative contribution of different factors affecting changes in energy use, and thus help to understand changing energy consumption patterns and to predict future energy demand. One of the early contributions in this context was the paper of Jenne and Cattell (1983) who examined the change in the ratio of energy consumed to industrial production in the UK and concluded that economic growth is the major key to rising efficiency and the second biggest effect is structural change, i.e. a shift away from the heavy energy users within a sector. Howarth and et al. (1991) used the Laspeyres index method to look at trends in manufacturing energy use in eight OECD countries between 1973 and 1987 by decomposing the changes in energy consumption into an activity effect, a structure effect and an energy intensity effect. They compared the results to those obtained by using the Divisia index method and found minor differences between the Laspeyres index and the Divisia index calculations. Another study that assessed the influence of developments in energy efficiency and economic structure on the total primary energy consumption is that of Farla and Blok (2000). Park (1992) used structural change, energy intensity and output level, to decompose the industrial energy demand in Korea from 1973 to In the field of energy-related GHG emissions, mostly changes in carbon dioxide emission are analyzed. In addition to structural and energy intensity changes, sectoral fuel shares and emission coefficients are relevant factors for changes in emissions on a national level. Ang and Zhang (1999) used a five factor decomposition for several groups of countries in 1993 to decompose the differences

6 6 of emissions from fossil fuel use between the regions. The large disparities found between regions were mainly due to differences in GDP and energy intensity. Greening et al. (1998) applied the adaptive weighted Divisia index to analyze energy consumption and carbon intensity of the freight sector of 10 OECD countries. A four factor decomposition for Turkey for the period was conducted by Lise (2006). This study identified a structural (composition) effect for the changing shares of sectors in GDP. The GDP effect, structure effect, and carbon intensity effect were all associated with substantial increases in emissions, while the energy intensity effect led to a small reduction in emissions. Lee and Oh (2006) produced a five factor decomposition for 15 APEC countries between 1980 and This group included high, middle and lower income countries. Although GDP and population were strong factors associated with an increase in emissions in all cases, in the high income countries falls in energy intensity and the share of fossil fuels, as well as a change in the fossil fuel mix all contributed to partially offsetting the impacts of the growth in the economies. The group of lower income countries was dominated by China which, in this period, experienced a large fall in energy intensity, offsetting nearly half the impacts of the increase in income and population. This empirical study uses the LMDI 1 methodology to decompose the changes in carbon dioxide emissions between 1971 and 2005 to five effects, namely the emissions per unit of fossil fuel, the change in the share of fossil fuels in total energy, the change in energy intensity, the change in GDP per capita and the change in population. It covers the biggest carbon dioxide emitting countries and regions that together account for over 80% of total emissions worldwide. 2. Methodology Changes in CO 2 emissions E may be studied by quantifying the contributions from changes in four different factors, the fossil fuel energy supply FPES, the total primary energy supply TPES, the Gross Domestic Product GDP and the population POP. If this is the case, then emissions in a country i can be expressed as the product of these factors:

7 7 = =, (4) where C (=E/FPES) is the CO 2 emissions coefficient, S (=FPES/TPES) is the share of fossil fuel consumption in total energy consumption, I (=TPES/GDP) is the energy intensity, G (=GDP/POP) is the per capita income, and P (=POP) is the population. The change in a country s carbon dioxide emissions ( E i ) between a base year 0 and a target or end year T can equally be decomposed into the following effects: E i E i T E i 0 C eff + S eff + I eff + G eff + P eff, (5) which are: (i) (ii) (iii) (iv) (v) the coefficient effect C eff, i.e. the change in the emissions per unit of fossil fuel the substitution effect S eff, i.e. the change concerning the share of fossil fuels in total energy, the energy intensity effect I eff, i.e. the change in the ratio of total primary energy supply to GDP the income effect G eff, i.e. the change in GDP per capita and the population effect P eff, i.e. the change in population. These effects can be calculated from the following formulae using the Log Mean Divisia Index method (LMDI 1) with data on all the variables for a common start and end year: C eff = [E T i E 0 i ]{ln [C T i / C 0 i ] / ln [E T i / E 0 i ]} (6) A positive coefficient effect in the sense that carbon emissions per unit of fossil fuel supply rise can be observed if there is a relative shift to higher emitting fuels, e.g. the share of coal rises relative to the share of gas.

8 8 S eff = [E i T E i 0 ]{ln [S i T / S i 0 ] / ln [E i T / E i 0 ]} (7) The ratio of fossil fuels to total energy supplied changes if the share of hydro, nuclear and renewables rises or falls relative to the share of fossil fuels. I eff = [E T i E 0 i ]{ln [I T i / I 0 i ] / ln [E T i / E 0 i ]} (8) Emissions fall due to an energy intensity effect if the use of energy increases more slowly than the level of GDP. This is the case either if the sector structure of GDP changes towards sectors that are less energy intensive or if energy efficiency increased in one or more sectors, without any structural shifts. The other two factors, the income and population effects, are straightforward and can be expressed as follows: Y eff = [E T i E 0 i ]{ln [Y T i / Y 0 i ] / ln [E T i / E 0 i ]} (9) P eff = [E T i E 0 i ]{ln [P T i / P 0 i ] / ln [E T i / E 0 i ]} (10) 3. Data The data used for this empirical investigation contains the following indicators: E = CO 2 emissions from the consumption of fossil fuels (Mt of CO 2, sectoral approach 4 ) FPES = TPES = fossil fuel share of primary energy supply (mtoe) total primary energy supply (mtoe) GDP = gross domestic product (billion 2000 US$ using PPPs) POP = population (millions) The annual data used in this study is obtained from the International Energy Agency Online Data Services database. The time frame observed reaches from 1971 to 2005 for the top CO 2 emitters. All 4 See International Energy Agency (2007) pp. 41 ff. for details on sectoral and reference approaches for the calculation of aggregate CO 2 emissions

9 9 time-series downloaded are also available from the annual IEA print publications Energy Balances of OECD Countries, Energy Balances of Non-OECD Countries and CO 2 Emissions from Fuel Combustion. Table 1: world indicators Indicator unit % change CO2 emissions (Mt of CO2) ,3% Total Primary Energy Supply (million toe) ,7% Share of Fossil Fuels (%) 86% 81% GDP (bn USD PPP) ,8% Population (millions) ,8% CO2 per capita (t of CO2 p.c.) 3,75 4,20 11,9% GDP per capita (000 USD PPP) ,6% While global GDP has more than tripled in the period between 1971 and 2005, Total Primary Energy Supply has doubled within the same time frame (see table 1). This downward energy intensity trend has led to corresponding decrease in emission intensity (CO 2 emissions per unit of GDP). On the other hand emissions per capita have increased from 3,75 to 4,2 tons of CO 2 as the change of carbon dioxide output from fossil fuel consumption was larger than the increase in the world s population. With a share of 81% - compared to 86% in fossil fuels still satisfy most of the world s energy supply. This relatively small step was mainly achieved by the rise of nuclear energy in power generation, not by a shift towards renewable energy sources. The countries and regions included in this study summed up to almost 85% of the world s CO 2 emissions and GDP in 2005 (and approx. 67% of the world population). Looking at table 2 it becomes obvious that there are vast differences between the biggest CO 2 emitting regions in terms of per capita income and emissions. The largest emissions per capita can be found in North America (the United States and Canada), Australia and Saudi Arabia as the biggest oil exporting country. All these countries also have a high GDP per capita. On the other side, India, Indonesia, Brazil and China have the lowest emissions and incomes per head; but with enormous populations also pose the biggest threat to the climate in the future.

10 10 Table 2: Indicators for the Top CO 2 emitters CO2 emissions Total Primary Energy Supply Share of Fossil Fuels GDP Population CO2 per capita GDP per capita (Mt of CO2) (million toe) (%) (bn USD PPP) (millions) (t of CO2 p.c.) (000 USD PPP) USA 5.817, ,1 86,2% ,8 296,7 19, ,2 CHN 5.059, ,5 84,2% 7.842, ,5 3, ,7 EU , ,5 78,8% ,4 492,0 8, ,9 FSUREG 2.302, ,8 90,1% 2.099,3 284,9 8, ,2 JPN 1.214, ,2 81,9% 3.473,8 127,8 9, ,7 IND 1.147, ,9 68,0% 3.362, ,6 1, ,5 CAN 548, ,0 75,7% 990,5 32,3 17, ,8 KOR 448, ,5 80,9% 957,9 48,3 9, ,1 IRN 407, ,6 98,7% 483,8 68,3 6, ,7 MEX 389, ,6 88,8% 982,7 105,3 3, ,3 AUS 376, ,7 94,5% 616,7 20,5 18, ,5 IDN 341, ,3 67,9% 754,1 220,6 1, ,0 ZAF 330, ,2 87,1% 463,5 46,9 7, ,5 BRA 329, ,7 56,7% 1.393,4 186,4 1, ,3 SAU 319, ,3 100,0% 323,2 23,1 13, ,8 World , ,1 80,8% , ,7 4, ,4 If all the regions listed in table 2 would emit the same amount of CO 2 per person as the United States, then the world carbon dioxide output would almost be four times higher than at present. The European Union as a whole is the third biggest emitter in the world. With per capita emissions of 8,1 tons of CO 2 it lies close to the middle of the range, but immense disparities within the Union make a closer look at its members indispensable. Due to the availability and composition of the aggregate time series, only 23 countries of the European Union (EUR23) are studied. Estonia, Latvia and Lithuania are assigned to the Former Soviet Union (FSUREG). Data for Slovenia is not available for the whole period under observation. Together the 23 EU members make up 98,7% of the total EU27 emissions and 99% of GDP. Apart from the Eastern European members Romania, Bulgaria and Hungary, also Portugal and Malta are characterized by low per capita income and emissions. France and Sweden on the other hand are rich countries with fairly low per capita emissions due to the small share of fossil fuels in total primary energy supply. 5 Apart from FSUREG (Former Soviet Union Region) all country codes used in tables and figures are officially assigned ISO alpha-3 codes, with country names being English short country names officially used by the ISO 3166 Maintenance Agency

11 11 Table 3: EU Indicators 2005 CO2 emissions Total Primary Energy Supply Share of Fossil Fuels GDP Population CO2 per capita GDP per capita (Mt of CO2) (million toe) (%) (bn USD PPP) (millions) (t of CO2 p.c.) (000 USD PPP) DEU 813, ,5 82,9% 2.169,4 82,5 9, GBR 529, ,2 88,6% 1.699,5 60,2 8, ITA 454, ,4 91,2% 1.521,6 58,5 7, FRA 388, ,6 53,2% 1.696,0 62,7 6, ESP 341, ,0 83,8% 995,5 43,4 7, POL 295, ,9 95,7% 473,4 38,2 7, NLD 183, ,1 93,3% 478,8 16,3 11, CZE 118, ,2 83,8% 182,2 10,2 11, BEL 111, ,3 74,1% 293,7 10,5 10, GRC 95, ,4 93,6% 282,6 11,1 8, ROU 91, ,0 83,6% 174,4 21,6 4, AUT 77, ,6 78,1% 247,3 8,2 9, PRT 63, ,5 84,7% 194,1 10,5 6, HUN 57, ,0 80,6% 155,8 10,1 5, FIN 55, ,7 55,0% 152,8 5,2 10, SWE 51, ,3 35,1% 270,9 9,0 5, DNK 47, ,9 83,2% 164,4 5,4 8, BGR 46, ,4 73,2% 62,2 7,7 6, IRL 43, ,2 96,3% 141,3 4,1 10, SVK 38, ,3 72,1% 73,4 5,4 7, LUX 11, ,4 92,6% 25,9 0,5 24, CYP 7, ,4 97,8% 15,5 0,8 9, MLT 2,6 949,2 100,0% 6,9 0,4 6, EU , ,5 78,8% ,4 492,0 8, If we take a look at the changes in emissions between 1971 and 2005 for the ten biggest emitters in 2005 (see fig. 3), different regional and national developments become apparent % +533% % % +477% -17% +61% -15% +54% +781% USA CHN FSU JPN IND DEU CAN GBR ITA KOR Fig. 3: Top 10 emitters 2005 in comparison to 1971 Of all countries, China, the United States and India have made the biggest leaps in absolute terms; Korea, China and India in relative terms. Under the top ten emitters, Germany and the United Kingdom were the only countries able to cut their emissions from 1971 to 2005.

12 12 4. Empirical Results On the following pages the results of the decomposition analysis are mostly exemplified for the complete time frame ( ), but in the graphical illustrations (see appendix) and in some cases within the text the outcome is also divided in two separate time frames ( and ) to better assess the target achievements of the Kyoto protocol. When looking at the aggregated results for the countries analyzed (see fig. 4) which together represent 85% of global carbon dioxide emissions it becomes evident that the largest contributor to the rise in emissions between 1971 and 2005 is the change in GDP per capita. This income effect G eff is almost three times higher than the effect arising from a growing world population P eff. aggregated CO2 emissions aggregated CO2 emissions aggregated CO2 emissions Fig. 4: aggregated decomposition results for the periods , and The positive income and population effects (in a numerical sense) can only partly be offset by a negative change in energy intensity, I eff. On this aggregate level neither the change in emissions per unit of fossil fuel the coefficient effect C eff nor the changing share of fossil fuels in total energy supply the substitution effect S eff could add significantly to a mitigation of global CO 2 emissions. On a regional level the picture looks similar. A strong positive income and population effect is partly compensated by energy productivity gains. Only Korea, Iran, Mexico, South Africa and Saudi Arabia exhibit a positive energy intensity trend, out of which the effect is significantly high for the energy producers Iran and Saudi Arabia.

13 13 Table 4: decomposition results for the top CO2 emitters C eff S eff I eff G eff P eff USA 1520,19 111,51-521, , , ,55 CHN 4259,92 62,47 784, , , ,53 EU23 177,61-372,99-660, , ,62 395,53 FSUREG 308,64-24,47-133,83-29,20 174,20 322,11 JPN 471,13-20,22-158,11-283,05 741,88 190,47 IND 948,50-43,18 325,68-261,41 564,73 362,53 CAN 208,12-29,89-46,19-173,07 289,40 167,91 KOR 397,95-27,35-37,48 48,08 344,40 70,30 IRN 365,66 20,15 2,69 175,45 32,42 135,02 MEX 292,08-16,53 13,29 40,34 97,46 157,42 AUS 233,37 20,02 8,35-50,54 149,46 106,08 IDN 315,87-3,62 125,44-42,25 163,00 73,31 ZAF 156,49-88,90-6,98 56,42 18,03 177,89 BRA 238,59-14,22 48,90-23,97 109,56 118,30 SAU 306,56 28,07 0,02 143,57 5,28 129,64 In Europe a shift to relatively lower emitting fuels can be observed (negative coefficient effect C eff ), whereas in the United States and China this effect is positive due to the rising share of coal in energy supply. The change in the ratio of fossil fuels to total energy supplied (substitution effect S eff ) is positive for the low income countries China, India and Indonesia as the share of combustible renewables and waste (e.g. biomass for cooking) declines with the rising level of GDP per capita and a decreasing importance of the agricultural sector. The European countries together have been able to compensate the income effect in emissions through large cuts in energy intensity but also negative substitution and coefficient effects. Table 5: decomposition results for 23 EU members C eff S eff I eff G eff P eff DEU -170,17-121,33-150,21-547,88 603,26 45,90 GBR -96,61-106,97-49,07-395,21 411,92 42,63 ITA 158,78-7,63-11,10-111,18 259,42 29,21 FRA -46,99-48,65-216,91-116,21 260,98 73,79 ESP 220,95-15,75-21,33 38,18 170,12 49,78 POL -2,06-15,88-8,22-214,76 191,89 44,94 NLD 52,51-9,31-10,69-55,43 95,04 32,97 CZE -33,03-8,27-23,48-77,75 71,07 5,38 BEL -5,89-11,52-34,43-48,96 79,76 9,26 GRC 70,26 5,04 0,57 17,69 35,73 11,23 ROU -24,18 0,02-14,64-107,69 92,46 5,67 AUT 28,23 0,45-9,02-15,99 47,06 5,77 PRT 48,41-0,30 1,51 12,91 28,01 6,27 HUN -4,70-17,24-9,93-27,07 51,20-1,63 FIN 15,33-2,35-12,64-14,97 39,21 6,09 SWE -32,35-2,26-53,63-23,20 39,58 7,17 DNK -8,07-0,03-9,01-36,08 32,55 4,52 BGR -16,95-3,65-15,56-47,28 54,84-5,30 IRL 21,92-1,41-1,02-30,16 44,03 10,46 SVK -0,87-0,54-11,16-16,30 20,66 6,46 LUX -4,12-5,60-0,50-16,03 14,16 3,86 CYP 5,24-0,01-0,03-0,87 5,38 0,77 MLT 1,97 0,20 0,00-0,85 2,31 0,31

14 14 Strong energy productivity gains can be observed for the top emitters Germany, the UK, Italy and France, but also for the Eastern European countries Poland, the Czech Republic, Rumania, Hungary and Bulgaria. Unsurprisingly in these countries the negative energy intensity effect I eff was particularly strong in the period after the fall of the iron curtain in the period between 1990 and 2005 as the inefficient power generation and production sites were closed or replaced. Eastern European Countries Eastern European Countries Eastern European Countries Fig. 5: results for Eastern European Countries for the periods , and The coefficient effect is particularly strong in Germany, Great Britain and France. In Germany this is mainly due to the fuel changes after reunification in the post 1990 period. In the UK the dominance of oil in energy supply has been reduced by the substitution of oil and coal with natural gas in heating and electricity generation. Most of the enormous substitution effect in France was achieved by a rising share of nuclear energy in power generation, which has also lead to a relative low emission per capita ratio. This holds also true for Belgium and Sweden, where hydro and nuclear energy together account for 64% of total primary energy demand. 5. Conclusion and Policy Implications This paper analyzes the development of carbon dioxide emissions and underlying factors for the world s top CO2 emitters between 1971 and 2005 using the log mean Divisia index (LMDI 1) approach of decomposition. The developments in CO 2 emissions from fuel combustion illustrate the need for a more sustainable energy system by implementing measures of energy efficiency and reducing carbon intensity of energy supply by shifting from fossil fuels to renewable energy sources.

15 15 The results show that GDP growth is by far the biggest contributor to global emissions followed by an increasing population. Decreasing energy intensity was and still is the most important factor to reduce emissions. Following the oil price shocks of the 1970s industrialized countries witnessed a rapid reduction in emissions per unit of GDP in the period between 1973 and 1990 through a decline in their energy intensity. The substitution of carbon-intensive fuels like coal through lower emitting fossil fuels (e.g. natural gas) or renewable energy sources and nuclear energy has only been successful in some countries but has not yet played a major role on a global scale. The world s CO 2 intensity of the fuel mix has remained almost constant between 1971 and 2005, as fossil fuels still continue to dominate the global energy system. A shift away from coal towards lower emitting fossil fuels in countries like China and India will be a challenging task because large coal reserves are accessible to those countries, whereas reserves of other energy sources are limited. Coal generates around twice the CO 2 emissions of gas, despite having an equivalent share in the world energy supply (IEA 2007), which emphasizes the need for clean coal technologies. In future climate regimes developing countries have to be included with measures that suit their financial abilities and their need for economic and social development. Industrialized nations are responsible for a large part of cumulative emissions and many of them - like the United States, Canada and Australia - have the highest per capita output of GHGs. These countries in particular have a responsibility to mitigate their GHG emissions and play a leading role in global actions against climate change. Although the European countries have made some effort since 1990, there is still a long way to go in order to fulfill their own climate change targets for These include a reduction of at least 20% in greenhouse gases rising to 30% if there is an international agreement committing other developed countries to "comparable emission reductions and economically more advanced developing countries to contributing adequately according to their responsibilities and respective capabilities and a 20% share of renewable energies in EU energy consumption by 2020 (European Commission, 2008).

16 16 How countries fare in terms of emission levels is generally related to their availability of natural resources (i.e. coal reserves, oil and gas infrastructure, access to nuclear energy and hydroelectricity) and their economic structure (i.e. share of agricultural, industrial and service sector). However, whether or not global and regional climate change targets will be met is mostly dependent on the development of the world s economy in the years to come. As long as renewable energy sources are not available at comparable price levels, transport is still relying on oil as the major energy source and with no clean coal technologies being available on a global scale, emissions will continue to rise with global GDP. The question whether carbon dioxide emissions decrease beyond some level of income per capita, i.e. if at high-income levels economic growth leads to environmental improvement (Environmental Kuznets Curve) is heavily debated and an issue not yet solved. Nevertheless developing countries like China and India will emit enormous amounts of CO 2 and other GHG until they reach an income per capita level where this trend reverses. A worldwide recession within the next years could slow down the acceleration rate of CO 2 emissions in the short run. But falling oil prices and rising unemployment rates could also lead to a shift of priorities away from a sound environmental policy to unburden households and industry. This would have devastating medium and long term effects for the climate. As the change in energy intensity is the factor that contributed most to decelerate carbon dioxide emissions, changes in energy consumptions have to be analyzed in terms of efficiency and structural effects. To deepen the understanding of the underlying mechanisms, this should be done on a sectoral level (residential, industry and transport). A change towards a sustainable global fuel mix is still not observable. Rising prices of fossil fuels and additional regulatory measures could speed up the development of renewable energy sources.

17 17 References Ang, B. W. (2004). "Decomposition analysis for policymaking in energy: which is the preferred method?" Energy Policy 32(9): Ang, B. W. and S. Y. Lee (1994). "Decomposition of Industrial Energy Consumption: Some Methodological and Application Issues." Energy Economics 16(2): Ang, B. W. and F. L. Liu (2001). "A new energy decomposition method: perfect in decomposition and consistent in aggregation " Energy 26(6): Ang, B. W. and F. Q. Zhang (1999). "Inter-regional comparisons of energy-related CO2 emissions using the decomposition technique." Energy 24(4): Ang, B. W., F. Q. Zhang, et al. (1998). "Factorizing changes in energy and environmental indicators through decomposition " Energy 23(6): European Commission (2008). "20 20 by Europe's climate change opportunity". Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions. Brussels. Farla, J. C. M. and K. Blok (2000). "Energy Efficiency and Structural Change in the Netherlands, 1980â 1995." Journal of Industrial Ecology 4(1): Greening, L. A., W. B. Davis, et al. (1998). "Decomposition of Aggregate Carbon Intensity for the Manufacturing Sector: Comparison of Declining Trends from Ten OECD Countries for the Period " Energy Economics 20(1): Herzog, T., K. A. Baumert, et al. (2006). Target: Intensity - An Analysis of Greenhouse Gas Intensity Targets, World Resources Institute (WRI). Hoekstra, R. and J. C. J. M. van der Bergh (2003). "Comparing Structural and Index Decomposition Analysis." Energy Economics 25(1): Howarth, R. B. and et al. (1991). "Manufacturing Energy Use in Eight OECD Countries: Decomposing the Impacts of Changes in Output, Industry Structure and Energy Intensity." Energy Economics 13(2): IEA, I. E. A. (2007). CO2 Emissions from Fuel Combustion Paris, OECD. IPCC (2007). IPCC Fourth Assessment Report, UNEP.

18 18 Jenne, C. A. and R. K. Cattell (1983). "Structural Change and Energy Efficiency in Industry." Energy Economics 5(2): Lee, K. and W. Oh (2006). "Analysis of CO2 emissions in APEC countries: A time-series and a cross-sectional decomposition using the log mean Divisia method." Energy Policy 34(17): Lise, W. (2006). "Decomposition of CO2 emissions over in Turkey." Energy Policy 34(14): Park, S.-H. (1992). "Decomposition of Industrial Energy Consumption: An Alternative Method." Energy Economics 14(4): Sun, J. W. (1998). "Changes in Energy Consumption and Energy Intensity: A Complete Decomposition Model." Energy Economics 20(1):

19 19 Appendix graphical illustration of country results a) USA CHN EUR FSU JPN IND CAN KOR IRN MEX AUS IDN ZAF BRA SAU b) USA CHN EUR FSU JPN IND CAN KOR IRN MEX AUS IDN ZAF BRA SAU

20 20 c) USA CHN EUR FSU JPN IND CAN KOR IRN MEX AUS IDN ZAF BRA SAU