مصرف انرژی و رشد اقتصادی: تحلیل علیتی برای یونان

This paper investigates the causal relationship between aggregated and disaggregated levels of energy consumption and economic growth for Greece for the period 1960–2006 through the application of a later development in the methodology of time series proposed by Toda and Yamamoto (1995). At aggregated levels of energy consumption empirical findings suggest the presence of a uni-directional causal relationship running from total energy consumption to real GDP. At disaggregated levels empirical evidence suggests that there is a bi-directional causal relationship between industrial and residential energy consumption to real GDP but this is not the case for the transport energy consumption with causal relationship being identified in neither direction. The importance of these findings lies on their policy implications and their adoption on structural policies affecting energy consumption in Greece suggesting that in order to address energy import dependence and environmental concerns without hindering economic growth emphasis should be put on the demand side and energy efficiency improvements.

Recent experiences with unprecedented high levels of energy prices, especially of oil prices and the commitment to international initiatives on reduction of greenhouse gas emissions such as the Kyoto protocol have revitalized the debate on the implementation of energy conservation policies. In this context, policies aiming at the gradual curtailing of energy needs have to consider the potential causal linkages between economic growth and energy consumption. Into this direction empirical research dating back to the 1970s and the first energy crisis offers an extensive and significant volume of investigation. The empirical focus has concentrated on whether economic growth boosts energy consumption or whether energy consumption takes precedence over economic growth. The investigation and the understanding of the direction of the causality between energy consumption and economic growth are of great value due to their policy implications. The existence of any causal relationships running from energy consumption to economic growth would indicate the dependence of the economy on energy with the latter being a stimulus to economic growth. In the presence of such causal relationships any structural policies aiming at the reduction of energy consumption like the policies following the Kyoto protocol or the recent surge in the fuel prices might possibly slow economic growth. This paper takes a fresh look at the causal relationship between energy consumption and economic growth for Greece for the period 1960–2006. The empirical investigation is conducted with the employment of a later development in the methodology of time series proposed by Toda and Yamamoto (1995). The aim is to identify the causal linkages between energy consumption and economic growth and to further assess the degree of energy consumption dependence for Greece overcoming limitations possibly present in earlier assessments. In addition to this we chose Greece, a medium sized economy, as an example to our empirical investigation with the aim to draw conclusions that could be further useful for the analysis of other small and medium sized economies like Portugal, Ireland, Spain, Malta, Luxemburg, Belgium and Cyprus all of the latter being countries faced with similar energy conservation policies deriving from commitments to international initiatives such as the Kyoto protocol and high levels of dependence on energy imports. Aiming at a rich insight on the causal relationships between energy consumption and economic growth, consumption patterns are assessed at aggregated as well as at disaggregated levels accounting for the energy consumption of the industrial, transport and residential sectors. At aggregated levels the findings remain important as heavy dependence of the country on energy consumption would render it weak to adverse effects on economic growth that environmental policies and energy conservation initiatives could cause. At disaggregated levels the findings help us understand better the relative importance that each sector bears as a causal factor of economic growth providing us at the same time with a clearer picture of the developments on energy consumption patterns across time and sectors and of their subsequent interactions with economic growth. The assessment of the relationship between energy consumption and economic growth in the case of a medium size country like Greece remains interesting in the context of serving as a representative empirical investigation case where practical conclusions could serve as the basis for further comparative analysis. In addition the employment of a later methodology in time series analysis could complement previous studies looking on the causal relationships between energy consumption and economic growth for Greece overcoming their methodological limitations. The theoretical approaches to date regarding the relationship between energy consumption and economic growth offer different models. In the traditional neo-classical growth model energy is introduced as an intermediate input next to the basic factors of land, labor and capital contributing to economic growth both directly and/or indirectly (Akarka and Long, 1979, Stern, 1993, Stern, 2000, Payne, 2008 and Yuan et al., 2008). On the biophysical approach energy becomes an important determinant of income with economies depending significantly on energy being affected heavily by changes in energy consumption (Cleveland et al., 1984). Following different theoretical approaches empirical investigations in the literature to date on the nature and the direction of the causal relationships between energy consumption and economic growth yield various and often contradictive results. The plethora of the latter appears to be the outcome of a diverse set of samples studied with different characteristics in each case, a variety of variables included, as well as of the implementation of different econometric approaches developed. The pioneer work of Kraft and Kraft (1978) with the application of a standard Granger (1969) test finds a uni-directional long-run relationship running from GDP to energy consumption in the USA for the period 1947–1974. Yu and Hwang (1984) employing the same methodology find no causality in the case of the USA for the period 1947–1979. Ever since, empirical investigations extended to developed and developing countries yielding mixed and often contradictive results. Yu and Choi (1985) employ the standard Granger (1969) test for the period 1954–1976 for a set of countries finding that causality runs from GDP to energy consumption for Korea, causality runs in the opposite direction for the Philippines while no causality is identified in the case of the USA, Poland and UK. Erol and Yu (1987) find the uni-directional relationship running from energy consumption to growth for Japan for the period 1950–1982. Employing the same test for Taiwan, Hwang and Gum (1992) for the period 1955–1993 and Yang (2000) for the period 1954–1997 find a bi-directional causal relationship between energy consumption and economic growth. Similarly Cheng (1997) and Wolde-Rufael (2004) find long-run causal relationship running from energy consumption to GDP in the case of Brazil for the period 1963–1993 and the case of Shanghai for the period 1952–1999 accordingly. Lee (2006) investigates the existence of causal relationships between energy consumption and economic growth for 11 developed countries for the period 1960–2001 deriving mixed results on the nature and the direction of causality. Cheng and Lai (1997) employ Hsiao's (1981) Granger causality test for Taiwan for the period 1995–1993. Their results in contrast to the work of Hwang and Gum (1992) and Yang (2000) suggest the causal relationship being uni-directional running form GDP to energy consumption. Chiou-Wei et al. (2008) apply both linear and nonlinear Granger causality tests to examine the causal relationship between energy consumption and economic growth for a sample of Asian newly industrialized countries as well as the USA for the period 1954–2006. Their study finds evidence supporting a neutrality hypothesis for the USA, Thailand, and South Korea. Moreover they find the existence of a uni-directional causality running from economic growth to energy consumption for Philippines and Singapore while energy consumption may have affected economic growth for Taiwan, Hong Kong, Malaysia and Indonesia. Chontanawat et al. (2008) test for causality between energy and GDP using a consistent data set and Granger test methodology for 30 OECD countries and 78 non-OECD countries. They find that causality from energy to GDP is found to be more prevalent in the developed OECD countries compared to the developing non-OECD countries. In the light of these findings the existing literature fails to come to a consensus on the nature of causal relationships between energy consumption and economic growth. The conflicting and inconsistent results may be attributed to different frameworks of institutions, structures and policies adopted by the countries as well as to the different methodological applications employed.1 The empirical investigation to date of the relationship between energy consumption and economic growth in the case of Greece offers a range of techniques employed looking mainly on energy demand associated to output and prices. Mitropoulos et al. (1982) employing Kalman filter techniques suggest that price elasticities behave as a cluster against energy demand. Samouilidis and Mitropoulos (1984) assess the interdependence of economic growth and energy consumption over the 1960s and the 1970s suggesting decreasing income and price elasticity of energy demand. Donatos and Mergos (1989) conclude that energy demand remains rather inelastic with respect to prices when looking at the developments over the period 1963–1984. In a more recent time framework Zonzilos and Lolos, 1996 and Christodoulakis and Kalyvitis, 1998 yield similar results on energy demand being rather inelastic with respect to prices. Hondroyiannis et al. (2002) re-examine the relationship between energy consumption, output and prices for the period 1960–1996 looking at their interdependency through the employment of vector error-correlation model estimation. Their findings suggest the existence of a long-run relationship between energy consumption, output and prices. The present attempt aims to complement the empirical investigations of the causal relationship between energy consumption and economic growth for Greece assessing recent developments of the latter and to overcome limitations present in the previous empirical approaches through the employment of a later technique proposed by Toda and Yamamoto (1995). Toda and Yamamoto, 1995 and Dolado and Lutkepohl, 1996 develop an alternative procedure for testing the Granger causality in possibly integrated and cointegrated systems of any integration order through the employment of an augmented level Vector autoregression (VAR) model. In this way Granger causality tests can be performed allowing for the long-run information often ignored in systems that require first differencing and pre-whitening (Masih and Masih, 1999 and Awokuse and Yang, 2003). Moreover the methodology may be preferred as it bypasses the need of unit root and cointegration pre-tests, tests that are often implemented in models like the vector error correlation model therefore overcoming the problems of biased pre-tests. The rest of the paper is organized as follows. Section 2 reviews the developments of energy consumption and economic growth in Greece. Section 3 presents the data and the econometric specification employed. Section 4 summarizes the empirical investigation and the estimation results at both aggregated and disaggregated levels of energy consumption. The last section summarizes the conclusions of the analysis and discusses the policy implications.

This paper has investigated the causal relationship between energy consumption and economic growth for Greece for the period 1960–2006. In order to assess the existence of any causal relationships energy consumption has been assessed at aggregated and at disaggregated levels. The empirical findings suggest the existence of a uni-directional causal relationship running from total energy consumption to real GDP. At disaggregated levels of energy consumption empirical findings suggest that there is present a bi-directional relationship between real GDP and industrial energy consumption as well as between real GDP and residential energy consumption. The tests of non-causality between transport energy consumption and real GDP for Greece suggest that there is no causal relationship running in either direction. The results suggest that output is affected by changes in total energy consumption in a way that policies aiming at energy conservation would not retard economic growth in Greece. These findings seem to confirm the growth hypothesis in the case of Greece implying that implementation of energy conservation policies would not adversely affect output but in contrast it would give impetus to economic growth. Empirical results are in tandem with the developments in Greece and have considerable policy implications with regards to the adoption and the implementation of structural reforms and the promotion of energy conservation policies. A closer look at the developments in the country suggests that this has been the case in Greece where surging growth rates of energy consumption over 1960–1973, an outcome of the industrialization process, were followed by considerably lower growth rates of output. Sluggish economic development following in the 1980s was the outcome of weak stop–go policies that failed to implement the necessary structural adjustments, industrial reconstruction and efficiency improvements. Failure to implement energy efficiency and energy conservation policies impeded economic growth leading to relatively lower growth rates. In contrast after the 1990s the implementation of structural policies that resulted in the diminishing, albeit high, rates of total energy consumption and the adoption of energy conservation policies has given acceleration and impetus to economic growth. At disaggregated levels of energy consumption empirical findings suggest that there is present a bi-directional relationship between real GDP and industrial energy consumption. The existence of a feedback relationship indicates the interdependence of the developments regarding industrial energy consumption and real GDP. In this respect energy consumption in the industrial sector and real GDP appear to be endogenous thus rendering misleading any single equation forecasts. The findings seem to confirm the experience of inefficient energy use in the industrial sector in Greece where increasing energy costs and retarded policies of promoting efficient energy use in the industrial sector have led to the increasing costs of production and the transformation of the sector in a non-competitive industrial sector constantly contracting over time. In this respect, strong forward and backward linkages of the industrial sector with the rest of the sectors along with the transformation of the Greek economy and in particular the increased importance of the tertiary sector has to be considered when understanding later developments in energy consumption and economic growth in Greece. With regards to the causal relationship between residential energy consumption and economic growth the results suggest the existence of a feedback relationship indicating the presence of interdependences between the developments in residential energy consumption and real GDP. Income growth has been followed by higher demand for residential energy amenities, improving living standards and increased numbers of consumers in the residential sector. Similarly increasing access to residential energy consumption seems to have had a detrimental effect upon output through the demand for qualitative services and the improved household comfort. For the transport sector the results suggest the neutrality hypothesis between transport energy consumption and real GDP. The tests of non-causality between transport energy consumption and real GDP for Greece suggest no causal relationship running in either direction indicating that developments on the earlier would not have any significant impact on the later and vice versa. In this context output developments do not seem to boost energy demand in the transport sector. Indeed changes in energy consumption patterns in the transport sector in Greece have been much related to preferences and necessities hence being exogenously determined and not by changes in income. In practice changing energy consumption patterns in the transport sector reflect developments as an outcome of preferences, trends and fashion, consumption tastes and habits that may not be directly related to changes in economic activity. The estimations have important implications for Greece. In the light of the existence of a causal relationship from total energy consumption to economic growth energy conservation and efficiency improvement policies should be taken into consideration and promoted accordingly. The experience with the implementation of structural policies in the 1990s that gave an impetus to lower energy requirements and to higher economic growth is supportive of the implementation of robust policies of energy efficiency. Policy makers should push forward the implementation of energy conservation policies in order to address not only obligations deriving from commitments to international initiatives such as the Kyoto protocol but also to address high energy imports dependence. Into this direction Greece is left with significant room for action. Energy sector in Greece has largely benefited from the EU support and the implementation of EU energy policy directives. Nevertheless Greece has not been able to avoid derogations in the implementation timetable. In the light of the empirical evidence energy policy should aim for energy conservation and emphasize more on the demand rather than on the supply side making first priority the implementation of energy efficiency projects rather than the penetration of natural gas into the domestic market. At disaggregated levels the support on specific sectors of the industry provided to date mainly in order to achieve social objectives should be offered not by maintaining the same inefficient production pattern but by facilitating energy efficiency improvements. With regards to residential energy consumption in the presence of the identified interdependences inevitable increase in the tariffs for residential energy will have a detrimental impact upon growth unless specific energy efficiency improvements are implemented. The impact of opening up of the electricity market and the subsequent challenge of the caped consumer prices upon residential energy consumption and growth should be addressed and it can be limited through the implementation of energy efficiency and energy saving improvements. With regards to the transport sector the findings offer significant implications especially when considering environmental concerns and commitments of Greece under the Kyoto protocol. In the presence of no interdependence between the transport sector energy consumption and output and considering the fact that the transport sector is one of the most energy intensive and polluting sectors in Greece policies aiming at energy efficiency and at the adoption of alternative environmentally friendly sources of energy use would facilitate compliance to Kyoto targets and decreased energy dependence without hindering economic growth. Into this direction the government has to further sufficiently design and communicate its energy policy and policy goals to the broader public. Policy makers should ensure that communities and local authorities have an understanding of the national energy situation. Given the increased need for quality and sustainable energy consumption agents may alter their energy behavior especially when considering energy efficiency, inevitably increasing energy tariffs and environmental friendly alternatives. In support of this financial incentives proposed under EU directives for energy saving improvements in buildings and vehicles as well as the promotion of less polluting technology should be specified and put into practice. In this way Greece may effectively address energy conservation policies, high energy dependence and environmental considerations without hindering economic growth. The results provide an indication of the causal relationships between energy consumption and economic growth for Greece and future research could further enrich our understanding of the later by identifying and incorporating a larger set of relevant variables in the analysis overcoming potential omitted variables limitations of the present attempt.