Elsevier

Energy Policy

Volume 37, Issue 11, November 2009, Pages 4689-4699
Energy Policy

A decomposition analysis of CO2 emissions from energy use: Turkish case

https://doi.org/10.1016/j.enpol.2009.06.019Get rights and content

Abstract

Environmental problems, especially “climate change” due to significant increase in anthropogenic greenhouse gases, have been on the agenda since 1980s. Among the greenhouse gases, carbon dioxide (CO2) is the most important one and is responsible for more than 60% of the greenhouse effect. The objective of this study is to identify the factors that contribute to changes in CO2 emissions for the Turkish economy by utilizing Log Mean Divisia Index (LMDI) method developed by Ang (2005) [Ang, B.W., 2005. The LMDI approach to decomposition analysis: a practical guide. Energy Policy 33, 867–871]. Turkish economy is divided into three aggregated sectors, namely agriculture, industry and services, and energy sources used by these sectors are aggregated into four groups: solid fuels, petroleum, natural gas and electricity. This study covers the period 1970–2006, which enables us to investigate the effects of different macroeconomic policies on carbon dioxide emissions through changes in shares of industries and use of different energy sources. Our analysis shows that the main component that determines the changes in CO2 emissions of the Turkish economy is the economic activity. Even though important changes in the structure of the economy during 1970–2006 period are observed, structure effect is not a significant factor in changes in CO2 emissions, however intensity effect is.

Introduction

Environmental problems, especially “climate change”, have been on the agenda since 1980s. Intensive use of fossil fuels and destruction of forests can be cited among the main reasons of the significant increase in anthropogenic greenhouse gases that lead to climate change. Among those gases, carbon dioxide (CO2) is the most important one and is responsible for most of the greenhouse effect. Enacted in February 2005, Kyoto Protocol is the first agreement which tries to limit greenhouse gas emissions and requires a timetable for realization of those reductions. The predecessor of Kyoto Protocol, UN Conference on Environment and Development was held in 1992 and a voluntary non-binding mechanism, UNFCCC (United Nations Framework Convention on Climate Change) was enacted in 1994. Turkey has joined UNFCCC in May 2004 and is considered as an Annex I Party, and eventually ratified the Kyoto Protocol in February 2009.

Turkey's contribution to CO2 emissions is quite low; in 2003 per capita CO2 emissions were 2.87 tons, much lower than the OECD average of 11.08 tons and also her share was 1.59% in total OECD emissions and 0.81% in world emissions (WECTNC, 2006a, WECTNC, 2006b). However, a rapid increase in CO2 emissions is observed in recent years. CO2 emissions excluding emissions/removals from land use, land-use change and forestry (LULUCF) (including emissions/removals from LULUCF) amounted to 140 (96) million tons in 1990 and 242 (168) million tons in 2004 which corresponds to a 73.3% (74.7%) increase between these years. This increase corresponds to one of the highest increases in the CO2 emissions among Annex I parties (UNFCCC, 2007).

Even though Turkey tries to comply with international standards, implementation of them is weak. This is mainly because of the priority given to the “industrial development strategy” over the environmental concerns. Although, the concept of “sustainable development” takes its place in the latest development plans, implementations in the environmental arena contradict with this theme.

Objective of this study is to identify the factors that contribute to changes in CO2 emissions for the Turkish economy by utilizing “decomposition analysis”. Turkish economy is divided into three aggregated sectors, namely agriculture, industry and, services. This study covers the period 1970–2006, during which significant policy changes have been observed in the economy. While Turkey had pursued an import-substitution development strategy until 1980, after this year she has adopted an export oriented strategy, gradually liberalizing the economy. At the same time, Turkish economy encountered several economic crises during the study period and adopted many stabilization programs under the auspices of IMF. Therefore, this study period enables us to observe the effects of different macroeconomic policies on carbon dioxide emissions through changes in shares of industries and the use of energy sources.

Decomposition analysis is one of the widely employed approaches in energy use and CO2 emission analyses in the literature. Among different methods of decomposition, Log Mean Divisia Index (LMDI) method developed by Ang (2005) is employed in this study. Furthermore, additive version of this method is utilized in which the change in one variable is decomposed as summation of changes in the components of that variable. Therefore, changes in CO2 emissions can be decomposed into changes in overall activity (activity effect), activity mix (structure effect), sectoral energy intensity (intensity effect), sectoral energy mix (energy-mix effect) and CO2 emission factor (emission-factor effect).

Considering the pace of increase in CO2 emissions, it is necessary to understand the share of different sectors in emissions, the sources and changes in the sources of emissions from 1970s to mid-2000s. Accordingly, effective policies for reducing CO2 emissions can be developed which can also be considered as a step towards sustainable use of energy sources.

There is a vast literature on decomposition analysis on energy consumption and CO2 emissions. However, applications of these techniques on the Turkish economy are limited. There are a couple of studies that decompose the energy consumption in Turkey. One of them which investigates sectoral energy consumption is Ediger and Huvaz (2006). Applying additive LMDI method by Ang et al. (1998) primary energy is mainly found to be determined by the production effect, structure effect and intensity effect for agriculture, industry and service sectors in the 1980–2000 period. The authors find that the production effect is more important than other effects and this is more apparent in industry and services sectors.

Türüt-Aşık et al. (2008) analyze the sources of the change in energy intensity at four digit sectoral level in the Turkish manufacturing industry, for the period of 1992–2001. Using energy decomposition method suggested by Choi and Ang (2003), it is found that the change in energy intensity of the manufacturing sector results from the change in sectoral energy intensities rather than the change in sectoral shares.

Another strand of literature applies similar techniques on CO2 emissions. An example is Karakaya and Özçağ (2003). For 1973–1999 period, the growth effect and population effect are identified as the main contributors to CO2 emissions in the Turkish economy. These effects are followed by fossil-fuel intensity effect (share of fossil fuels in total energy) and conversion-efficiency effect (primary energy required to deliver energy for final consumption). The authors claim that increasing energy demand due to economic growth in the Turkish economy has been satisfied increasingly by fossil fuels and decreasingly by renewable sources.

Lise (2006) reaches similar conclusions. CO2 emissions for the period 1980–2003 are analyzed for four aggregated sectors; agriculture, industry, transportation and services. In this study refined Laspeyres method is used. The author concludes that while the growth of the economy (scale effect) is responsible for most of the CO2 emissions, the carbon intensity and composition effects also contribute to CO2 emissions. Energy intensity effect on the other hand is only responsible from a modest reduction in CO2 emissions.

Different from the aggregate decomposition studies that are mentioned above Akbostancı et al. (2008) look into the CO2 emissions of Turkish manufacturing industry at four digit level by using four fuel groups for 1995–2001 period. Similar to the current study, authors utilize additive LMDI method developed by Ang (2005). It is found that main causes of the change in CO2 emissions during the study period are the changes in total industrial activity and energy intensity. It is also indicated that among the fuels used, coal is the main determining factor and among the sectors iron and steel basic industries is the single dirtiest sector dominating the industrial CO2 emissions. Following iron and steel sector, manufacture of cement, lime and plaster, petroleum refineries and manufacture of synthetic resins, plastic materials and man made fibers except glass sectors are the other higher CO2 emitters.

Current study improves the previous decomposition studies of CO2 emissions by updating the coverage of the time period as well as utilizing the additive LMDI which is not used in the previous studies. Advantage of the LMDI method is that it does not result in an unexplained residual term. The decomposition literature deals with the residual term by different methods, which are mainly arbitrary like equally distributing the residual among the components.1 This study utilizes the decomposition method suggested by Ang (2005) which deals with this problem. Additionally, another novelty is that the decomposition method used by the current study enables us to look into the contributions of energy sources to the changes in CO2 emissions. This way we are able to track the changes in different energy sources used throughout the study period and impact of this change in the energy consumption structure on the CO2 emissions. This allows the researchers to devise policy implications from this perspective as well.

Rest of the study is organized as follows. Turkish energy policy is discussed in Section 2, decomposition analysis methodology and the data set are discussed in 3 Methodology, 4 Data set respectively. Results of the decomposition analysis are presented in Section 5, and finally Section 6 concludes the paper.

Section snippets

Energy policy of Turkey

In terms of total and per capita energy consumption, developed countries are ahead of developing countries. Most of the energy is consumed by high income countries and USA is the number one among the highest energy consuming countries. Even if it can be argued that the developed countries have reached the saturation point in terms of their energy use, this is not true for developing countries. According to Churchill (1993), in the following decades the increases in energy demand will emanate

Methodology

In terms of decomposition analysis we use LMDI method developed by Ang (2005) to figure out the leading factors that contribute to the change in CO2 emissions in the Turkish economy. LMDI method does not bear a residual term and all zeros in the data set may be replaced by a small positive constant (Liu et al., 2007). Mainly because of these two properties of this method, many applications utilizing this method could be found in the literature.4

Data set

In this study, we utilize the energy consumption data for Turkey from World Energy Council Turkish National Committee (WECTNC) for 1970–2006 period.5 Data set is used in three sector

Results

Before presenting the decomposition analysis results we discuss the general characteristics of the Turkish economy and its energy use during the study period. General trend of the economic activity in the Turkish economy can be followed from Fig. 1. It is apparent that the volatility of the Turkish economy is more in the 1990s compared to the 1970s and 1980s. From the real GDP series the impact of economic and political crisis of 1980 is seen as a slowdown in the economy. However, when we look

Conclusion

In this study, CO2 emissions of the Turkish economy in 1970–2006 period are analyzed in three sectors: agriculture, industry and services. The analysis also considers the impact of different energy sources used by these sectors in terms of four groups: solid fuels, petroleum, natural gas and electricity. Different components of the change in CO2 emissions are calculated by LMDI method of decomposition suggested by Ang (2005).

Our analysis shows that, the main component that determines the

Acknowledgements

This study has been supported by METU Scientific Research Project Fund (BAP-2008-04-03-07). We would like to thank WECTNC for providing the energy data. Earlier versions of this study were presented in The Ninth Biennial Conference of The International Society for Ecological Economics, 2008 Nairobi Kenya, and European Association of Evolutionary Political Economy Conference 2008, Rome Italy.

References (33)

  • C. Sayın et al.

    Assessing of energy policies based on Turkish agriculture: current status and some implications

    Energy Policy

    (2005)
  • Akbostancı, E., Tunç, G. İ., Türüt-Aşık, S., 2008. CO2 Emissions of Turkish manufacturing industry: a decomposition...
  • T. Baysan et al.

    Turkey's trade liberalization in the 1980s and prospects for its sustainability

  • Churchill, A.A., 1993. Energy demand and supply in the developing world, 1990–2020. In: Proceedings of the World Bank...
  • Houghton, J.T., Meira Filho, L.G, Lim, B., Treanton, K, Mamaty, I, Bonduki, Y, Griggs, D.J., Callender, B.A., 1996....
  • IEA, 2001. Energy Policies of IEA Countries: Turkey 2001 Review, OECD/IEA,...
  • Cited by (0)

    View full text