Energy sources and gross domestic product: international evidence.

Ahmad, Waseem ; Ahmed, Tanvir

ABSTRACT

Different energy sources are a necessary requirement for economic and social development of a country and no country has progressed from subsistence economy without the use of energy. The literature has found a significant relationship between energy consumption and economic growth. This work investigates the relationship between the gross domestic product (GDP) and various factors i.e. labour, capital and various forms of energy, region, income level and climatic zone. Present study uses cross country panel data of 40 countries over the period 1990 to 2011. The study applies Translog functional form and uses feasible generalised least squares approach to estimate the parameters. The results of the study show positive and significant elasticity of output with respect to labour, capital gas, oil and electricity. However, output elasticity of coal is negative and significant. The study further indicates significant impact of region, income level and climatic zone on GDP. The results suggest that sustainable supply of energy especially electricity will have a significant impact on the economic growth, while energy conservation policies have detrimental impact on the growth of a country.

JEL Classification: O47, Q42, Q43

Keywords: Gas, Oil, Electricity, Coal, Gross Domestic Product, Panel Data

INTRODUCTION

The relationship between energy consumption and economic growth received a significant amount of attention in energy economics literature [Al-Iraiani (2006)]. Rufael (2006) stated that different energy sources are a necessary requirement for economic and social development and no country in the world has progressed from subsistence economy without the use of energy. In this regard, four views have emerged over time about the relationship between energy consumption and output growth. One point of view is that energy is the prime source of value and other factors like labor and capital cannot do without energy. Many studies argue that the impact of energy use on growth depends on the structure of the economy and the stage of economic growth of the country concerned [Ghali and Sakka (2004)]. The bulk of the literature reports a uni-directional causality from energy consumption to economic growth. When the causality runs from energy consumption to economic growth, it is also called 'growth hypothesis'. Table 1 provides a list of the studies, which show such results. It implies that an increase in energy consumption has a significant impact on economic growth and if it is positive, then energy conservation policies have a detrimental impact on economic growth. Alternatively, if an increase in energy consumption has significant negative impact on GDP, it implies that growing economy needs a less amount of energy consumption, may be due to shift towards less energy intensive sectors [Payne (2010)]. Second point of view is that economic growth has a positive influence on energy consumption. There may be uni-directional causality from economic growth to energy consumption. Table 1 displays a list of studies showing such results. When the causality runs from economic growth to energy consumption, it is often referred to as 'conservation hypothesis'. It implies that energy conservation policies formulated to reduce energy consumption may not adversely affect economic growth. Third point of view is that the cost of energy use is very small compared to GDP and consequently its impact on economic growth is non-significant. There may be no causality between energy consumption and GDP; it is often referred to as 'neutrality hypothesis". A list of studies showing such results is given in Table 1. It implies that energy consumption has not a significant influence on economic growth, which means that neither conservation nor expansive policies pertaining to energy consumption have any effect on economic growth [Ozturk (2010)]. Fourth point of view is that when output and energy consumption are moving together towards a long-run equilibrium, and energy consumption and GDP are interdependent, and affect each other at the same time, there may be bi-directional causality [Payne (2010); Ozturk (2010)]. It implies that an increase (decrease) in GDP causes an increase (decrease) in energy consumption and similarly an increase (decrease) in GDP results in an increase (decrease) in energy consumption. It is also called 'feedback' hypothesis. A list of studies reporting such results is given in Table 1.

Different forms of causality between energy consumption and the economic growth have been reported by many studies in different countries (Table 2). Further multi-country studies also show similar results (Table 3). Thus empirical studies conducted on the energy consumption and economic growth yielded mixed results in terms of the above hypotheses; that is, some studies show causality running from energy consumption to economic growth, others report causality running from economic growth to energy consumption, while some studies find no causality or bi-directional causality. There is absence of consensus on the relationship between energy consumption and growth.

Karanfil (2009) has suggested that any future research using the same methods, variables and changing study period have no more potential to make a contribution to the existing energy consumption--economic growth literature. In order to avoid conflicting and unreliable results, Ozturk (2010) has suggested the use of new approaches, including panel data approach. Further, a majority of studies (Tables 1 to 3) estimate the causal relationship between aggregate energy consumption and economic growth. Use of aggregate energy consumption may mask the differential impact associated with various forms of energy consumption like gas, oil, electricity and coal [Payne (2010)]. The aim of this paper is to empirically investigate the relationship between output and energy use of various forms. We use a framework of neoclassical production economics where labour, capital and various forms of energy (i.e. gas, oil, electricity and coal) are treated as separate inputs. Within this framework, we use cross country panel data over the period 1990-2011. The results of the translog production function show that labor, capital, gas, oil and electricity have positive and significant impact on the GDP, while the coal has negative significant impact. This paper contributes in the following ways, first we use panel data approach to estimate the impact of energy consumption on economic growth. Second, we use the cross country data of countries having different levels of income to estimate the relation, which has not been done so far. Third, we estimate the relationship between energy consumption of various forms along with labor and capital on output.

The remainder of the paper is organised as follows. Section 2 is concerned with the data and variables, and reports methodology along with the description of the model. Section 3 presents the empirical results and Section 4 deals with the conclusion and policy implications.

DATA AND VARIABLES

In this study, cross country data have been used to estimate the production function by using real GDP as dependent variable and factors like total labour force, gross capital formation, and consumption of gas, oil, electricity and coal as independent variables. Besides these variables, dummy variables have been included in the model to capture the region specific, income level and climate effects. Data for real GDP are measured in constant 2005 US dollar and are obtained from the World Development Indicators [WDI, The World Bank (2011)]. Labour is a conventional input and is measured in millions, capital is measured in terms of gross capital formation in million US$ and is considered as a reliable proxy for capital stock [Jin and Yu (1996) and Shan and Sun (1998)]. The data for total labor force and gross capital formation are obtained from World Development Indicators [WDI, The World Bank (2011)]. Natural gas consumption is measured in billion cubic meters. Oil products include all liquid hydrocarbons obtained by refining of crude oil and NGL and by the treatment of natural gas, in particular LPG (liquid petroleum gas) production; it is measured in million tons. Electricity is measured in terawatt hours; it includes electricity consumption of private, public and industrial sectors. Coal is measured in million tons. The data for gas, oil, electricity and coal are obtained from Global Statistical Yearbook (http:// yearbook.enerdata.net/). Regional dummy variables are included to capture the regional specific effects. These regions are Europe, Commonwealth of Independent States, North America, Latin America, Asia, Pacific, Africa and Middle East. All World Bank countries have been divided into three groups on the basis of gross national income per capita i.e. low income ($1035 or less), middle income ($1036 to $12615) and high income ($12616 or more) (http://data.worldbank.org/about/country-classifications/ country-and-lending-groups). Dataset are available from 1990 to 2011 about the above variables only for 40 countries. These countries are either in the middle income or in high income group. Therefore only one dummy variable is used in the analysis. A list of countries included in this study is given in Appendix. The Koppen climate classification system divides the world's climate into 5 types on the basis of annual and monthly averages of temperature and precipitation. For the purpose of this study, last two types of climate i.e. Moist Continental Mid-latitude climate--E category (where the winter is cold and average temperature of the, coldest month is less than -3 [C.sup.0]) and Polar Climates--D category (where the soil is permanently frozen to depths of hundreds of meters or where the soil surface is permanently covered with snow and ice) have been grouped into one category. A dummy variable assumes a value of one if the country is mainly located in either of the above two climate zones, otherwise zero.

The descriptive statistics show that the average GDP of countries included in the sample is 833139 million US $. It may be noted that the countries included in the sample belong to the high income or middle income categories. The average value of dummy variable for middle income group shows about 48 percent countries included in the sample belong to the middle income category and 52 percent countries in the sample belong to the high income category. Due to non-availability of data about the low income countries, we could not include them in the analysis. The average value of electricity is 243.83 terawatt hour and the mean value of gas and oil are 45.30 million cubic meter and 58.60 million tons respectively (as shown in Table 4). The regional dummies show that about 35 percent countries included in the analysis are from European region and 15 percent countries included in the analysis are each from Latin America and Asian region. Dummy for climatic region shows that about 20 percent countries included in the analysis belong to D or E region.

The Model

The present study examines the relationship between gross domestic product (GDP) and various factors in production function framework such as total labor, gross capital formation and energy; energy is further divided into different forms such as oil, gas, electricity and coal. Mathematically it can be written as:

GDP = f(L, K, G, O, E, C) ... (1)

Where GDP represents the gross domestic product (GDP), L denotes total labor, K shows the capital, G represents gas, O denotes oil, E shows electricity and C indicates the coal consumption.

In this study a Translog function has been used; this function can be approximated by second order Taylor series. The Translog functional form imposes fewer restrictions on the production technology. It does not impose any a priori restriction on returns to scale and elasticity of substitution. Because of above mentioned reasons, it is widely used in the production economics literature [Kim (1992)]. We also used dummies for different regions, income levels and climatic zone. The detailed functional form can be written as follows:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Where [GDP.sub.jt] is gross domestic product of jth country in year t, [L.sub.jt], is total labor of jth country at t year, [K.sub.jt], is gross capital formation of jth country at t year, [G.sub.jt] is total gas consumption of jth country at t year, [O.sub.jt] is total consumption of oil products in jth country at t year, [E.sub.jt], is total domestic consumption of electricity in jth country at t year, [C.sub.jt] is total consumption of coal in jth country at t year, Deu, Deis, Dnamerica, Dlamerica, Dasia, Dpacific are different regional dummies, Dmiddle shows the dummy for middle income countries, Dcold denotes the dummy for cold climatic zone and [[mu].sub.jt] is the random error term.

The elasticity of GDP with respect to each input i.e. labor, capital, gas, oil, electricity and coal would be calculated by using:

[[epsilon].sub.i] [partial derivative]LnGDP/[partial derivative][LnX.sub.i] where [X.sub.i], represents labour, capital, gas, oil, electricity and coal. So the elasticity of each input can be written as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

Where [bar.LnL], [bar.LnK], [bar.LnG], [bar.LnO], [bar.LnE] and [bar.LnC] represent the average values.

The definition of variables and their expected signs are presented in Table 5.

Results

For estimation purpose, Translog model has been used on panel data of 40 countries from 1990 to 2011. In this regard likelihood ratio, heteroscedasticity and auto correlation tests were used for diagnostic purposes.

Likelihood ratio test is used to test the nested hypothesis of the model, in this regard; we compare the restricted (Cobb Douglas) and unrestricted (Translog) model. LR test helps us to identify whether the imposition of restriction holds or not. The LR test statistic is 535.09 and this value is significant at 1 percent level of significance. It indicates that the unrestricted model (Translog) performs better than Cobb Douglas.

In the presence of heteroscedasticity the estimates are unbiased but inefficient [Gujarati (2007)]. We use likelihood ratio test for testing existence of heteroscedasticity in the panel data [Ahmad and Anders (2012)]. The [chi square] value is 1063.15, which is significant at 1 percent level of significance. It shows that there is a problem of heteroscedasticity in the data.

Serial correlation in panel data model biases the standard error and makes the results inefficient. In the present study, we use Wooldridge test to test for serial correlation in the model. This test is easy to implement and requires relatively less assumptions [Drukker (2003)]. The result of the Wooldridge test statistic is 260.51, which is significant at 1 percent level of significance. The result of the test shows that there is a problem of autocorrelation in the data.

To fix the problem of heteroskedasticity and autocorrelation, we applied feasible generalized least square approach. It gives us unbiased and consistent results.

In the present study, we also applied Wald test to see the joint significance of different regions. The value of Wald test is 891.36 and it is significant at 1 percent level of significance. Thus on the basis of results of the model, the null hypothesis that there are no regional differences is strongly rejected as a composite hypothesis. Thus different regions have jointly significant impact on the GDP of the country.

The results of the estimated model are presented in Table 6. It presents the estimated coefficients and their standard errors. Overall results of model show that most of the coefficients are statistically significant. Based on the Translog production function estimates shown in Table 6, we derive the returns to scale and output elasticities with respect to the inputs. By taking sum of six output elasticities, we can get the value of return to scale. This value comes out to be 1.084 showing almost constant returns to scale.

The estimated elasticities of different inputs are given in Table 7. The elasticity estimates show that the coefficients of conventional inputs labor and capital are 0.04 and 0.43 respectively. These coefficients show that if there is 1 percent increase in labor, it will increase the GDP by 0.04 percent while 1 percent increase in the capital will result in an increase of 0.43 percent. A number of studies show that economic growth is influenced by the amount of energy as well as primary inputs i.e. labor and capital [Beaureau (2005)]. Lie and Liu (2011) reported the mean GDP elasticity with respect to labor and capital over the 10 years study period to be 0.302 and 0.614 respectively. Thus the study results show that capital intensive technology will be more beneficial for countries. The GDP elasticity estimates of gas, oil and electricity are positive; these results show that these energy inputs have positive impact on the GDP.

The GDP elasticity of gas, oil and electricity are 0.001, 0.19 and 0.45 respectively. These results show that among the various forms of energy, electricity is the most important factor in influencing the GDP. It is important to ensure its supply for sustainable economic development. These results indicate that electricity increase has the largest effect on the GDP while gas increase has the lowest positive impact on the GDP. The GDP elasticity of electricity shows that an increase of electricity by 1 percent will increase the GDP by 0.45 percent.

The GDP elasticity with respect to coal is negative. It shows that an increase in consumption of coal by 1 percent will decrease GDP by 0.03 percent. This results due to the fact that the average domestic consumption of coal showed either decreasing or stagnant behavior during the first thirteen years of study period. However, there was an increasing trend in the use of coal during the last nine year. Many countries showed substantial reduction in the domestic consumption of coal. For example, coal domestic consumption decrease from 448.81 million tons (MT) to 238.0 MT in Germany, 106.68 MT to 51.22 MT in United Kingdom, and 149.85 MT to 72.85 MT in Ukraine over the period 1990 to 2011. There was also reduction in coal consumption in Belgium, France, Romania, Spain, Kazakhstan, Uzbekistan and Columbia. However, there was an increase in the domestic consumption of coal in India from 220.86 MT to 703.28 MT, Indonesia from 8.27 MT to 71.25 MT and Turkey from 54.42 MT to 102.06 MT. Japan, Chile, Mexico, Malaysia, Thailand, South Africa also experienced an increase in coal consumption. Other countries like Italy, Netherlands, Portugal, Sweden, Egypt, Argentina, Nigeria, Algeria, Pakistan, Kuwait, Norway etc. either experience stagnant behavior or negligible use of coal.

CONCLUSION AND POLICY IMPLICATIONS

The paper determines the relationship between energy consumption in different forms and conventional inputs i.e. labour and capital with real gross domestic product in a production function framework. A Translog production function model is used on panel data of forty countries from 1990 to 2011. Feasible generalised least squares approach is applied in order to fix the problem of heteroskedasticity and autocorrelation. The results of the study show that all the independent variables included in the analysis have positive and significant impact on GDP except the coal variable. The study reveals that different regions, income level and climatic zones have significant impact on the GDP. Energy consumption in the form of electricity has the strongest impact on GDP than any other variable. The GDP elasticity estimate of electricity is 0.45, which shows that 1 percent increase in the electricity increases GDP by 0.45 percent. The GDP elasticity of electricity is substantially higher than any other form of energy. This suggests that policy maker should ensure sustainable electricity supply and place more emphasis on this form of energy. Any shocks to electricity supply will adversely affect the real GDP growth. In order to avoid the adverse effects of electricity supply, it is necessary for countries, especially developing countries facing its shortage, to plan and develop generation capacity to meet the electricity demand of their countries.

Waseem Ahmad <[email protected]> is Assistant Professor, Institute of Business Management Sciences, University of Agriculture, Faisalabad. Tanvir Ahmed is Associate Professor, Department of Economics, Forman Christian College (A Charted University), Lahore.

APPENDIX

Belgium, Finland, France, Germany, Italy, Netherlands, Poland, Portugal, Romania, Spain, Sweden, United Kingdom, Norway, Turkey, Kazakhstan, Ukraine, Uzbekistan, Canada, United States, Argentina, Brazil, Chile, Columbia, Mexico, Venezuela, India, Pakistan, Indonesia, Japan, Malaysia, Thailand, Australia, New Zealand, Algeria, Egypt, Nigeria, South Africa, Kuwait, Saudi Arabia, United Arab Emirates.

REFERENCES

Abosedra, S. and H. Baghestani (1989) New Evidence on the Causal Relationship between United States Energy Consumption and Gross National Product. Journal of Energy Development 14, 285-292.

Ahmad, W. and S. Anders (2012) The Value of Brand and Convenience Attributes In Highly Processed Food Products. Canadian Journal of Agricultural Economics 60, 113-133

Akarca, A. T. and T. V. Long (1980) On the Relationship between Energy and GNP: A Reexamination. Journal of Energy Development 5, 326-331.

Al-Iriani, M. A. (2006) Energy--GDP Relationship Revisited: An Example from GCC Countries Using Panel Causality. Energy Policy 34, 3342-3350.

Altinay, G. and E. Karagol (2004) Structural Break, Unit Root, and the Causality between Energy Consumption and GDP in Turkey. Energy Economics 26:6, 985-994.

Ang, J. B. (2008) Economic Development, Pollutant Emissions and Energy Consumption in Malaysia. Journal of Policy Modeling 30, 271-278.

Aqeel, A. and M. S. Butt (2001) The Relationship between Energy Consumption and Economic Growth in Pakistan. Asia Pacific Development Journal 8:2, 101-110.

Beaudreau, B. C. (2005) Engineering and Economic Growth. Structural Change and Economic Dynamics 16, 211-220.

Belloumi, M. (2009) Energy Consumption and GDP in Tunisia: Co-integration and Causality Analysis. Energy Policy 37:7, 2745-2753.

Bowden, N. and J. E. Payne (2009) The Causa! Relationship between US Energy Consumption and Real Output: A Disaggregated Analysis. Journal of Policy Modeling 31:2, 180-188.

Chang, T., H. P. Chu, and W. Y. Chen (2013) Energy Consumption and Economic Growth in 12 Asian Countries: Panel Data Analysis. Applied Economic Letters 20, 282-287.

Cheng, B. (1995) An Investigation of Co-integration and Causality between Energy Consumption and Economic Growth. Journal of Energy Development 21, 73-84.

Cheng, B. S. (1998) Energy Consumption, Employment and Causality in Japan: A Multivariate Approach. Indian Economic Review 33:1, 19-29.

Cheng, B. S. (1999) Causality between Energy Consumption and Economic Growth in India: An Application of Co-integration and Error-Correction Modeling. Indian Economic Review 34, 39-49

Chiou-Wei, S. Z., C. F. Chen, and Z. Zhu (2008) Economic Growth and Energy Consumption Revisited--Evidence from Linear and Non-linear Granger Causality. Energy Economics 30:6, 3063-3076.

Drukker, D. M. (2003) Testing for Serial Correlation in Linear Panel Data Models. The St at a Journal 3:2, 168-177.

Erdal, G., H. Erdal, and K. Esengun (2008) The Causality between Energy Consumption and Economic Growth in Turkey. Energy Policy 36:10, 3838-3842.

Erol, U. and E. S. H. Yu (1987) On the Causal Relationship between Energy and Income for Industrialised Countries. Journal of Energy Development 13, 113-122.

Fatai, K., L. Oxley, and F. Scrimgeour (2002) Energy Consumption and Employment in New Zealand: Searching for Causality. Paper presented at NZAEC on France, Wellington, 26-28 June.

Ghali, K. H. and M. I. T. El-Sakka (2004) Energy Use and Output Growth in Canada: A Multivariate Co-integration Analysis. Energy Economics 26, 225-238.

Glasure, Y. U. (2002) Energy and National Income in Korea: Further Evidence on the Role of Omitted Variables. Energy Economics 24, 355-365.

Gujarati, D. N. and Sangeetha (2007) Basic Econometrics. Fourth Edition. Tata McGraw Hill Publishing Company Limited New Delhi.

Halicioglu, F. (2009) An Econometric Study of C[O.sub.2] Emissions, Energy Consumption, Income and Foreign Trade in Turkey. Energy Policy 37, 1156-1164.

Ho, C-Y. and K. W. Siu (2007) A Dynamic Equilibrium of Electricity Consumption and GDP in Hong Kong: An Empirical Investigation. Energy Policy 35:4, 2507-2513.

Jin, J. C. and E. S. H. Yu (1996) Export-led Growth and the US Economy: Another Look. Applied Economics Letters 3, 341-344.

Karanfil, F. (2008) Energy Consumption and Economic Growth Revisited: Does the Size of Unrecorded Economy Matter? Energy Policy 36:8, 3029-3035.

Kraft, J. and A. Kraft (1978) On the Relationship between Energy and GNP. Journal of Energy and Development 3, 401-403.

Lee, C. C. (2005) Energy Consumption and GDP in Developing Countries: A Cointegrated Panel Analysis. Energy Economics 27, 415-427.

Lee, C. C. (2006) The Causality Relationship between Energy Consumption and GDP in G-11 Countries Revisited. Energy Policy 34, 1086-1093.

Lee, C. C. and C. P. Chang (2007a) The Impact of Energy Consumption on Economic Growth: Evidence from Linear and Nonlinear Models in Taiwan. Energy 32:1, 2282-2294.

Lee, C. C. and C. P. Chang (2007b) Energy Consumption and GDP Revisited: A Panel Analysis of Developed and Developing Countries. Energy Economics 29, 1206-1223.

Lise, W. and K. Van Montfort (2007) Energy Consumption and GDP in Turkey: Is There a Co-integration Relationship? Energy Economics 29, 1166-1178.

Masih, A. and R. Masih (1996) Energy Consumption and Real Income Temporal Causality, Results for a Multi-country Study Based on Co-integration and Error-Correction Techniques. Energy Economics 18, 165-183.

Murray, D. A. and G. D. Nan (1996) A Definition of the Gross Domestic Product-Electrification Inter Relationship. Journal of Energy and Development 19, 275-283.

Oh, W. and K. Lee (2004) Causal Relationship between Energy Consumption and GDP: The Case of Korea 1970-1999. Energy Economics 26:1, 51-59.

Ozturk, I. (2010) A Literature Survey on Energy-Nexus. Energy Policy 38, 340-349.

Paul, S. and R. N. Bhattacharya (2004) Causality between Energy Consumption and Economic Growth in India: A Note on Conflicting Results. Energy Economics 26:6, 977-983.

Payne, J. E. (2009) On the Dynamics of Energy Consumption and Output in the US. Applied Energy 86:4, 575-577.

Payne, J. E. (2010) Survey of the International Evidence on the Causal Relationship between Energy Consumption and Growth. Journal of Economic Studies 37, 53-95.

Rufael, Y. W. (2006) Electricity Consumption and Economic Growth: A Time Series Experience for 17 African Countries. Energy Policy 34, 1106-1114

Shan, J. and F. Sun (1998) Export-Led Growth Hypothesis for Australia: An Empirical Re-investigation. Applied Economics Letters 5, 423-428.

Soytas, U., R. Sari and O. Ozdemir (2001) Energy Consumption and GDP Relation in Turkey: A Co-integration and Vector Error Correction Analysis. In Economies and Business in Transition: Facilitating Competitiveness and Change in the Global Environment Proceedings. Global Business and Technology Association, pp. 838-844.

Soytas, U. and R. Sari (2003) Energy Consumption and GDP: Causality Relationship in G-7 Countries and Emerging Markets. Energy Economics 25, 33-37.

Soytas, U. and R. Sari (2006) Energy Consumption and Income in G7 Countries. Journal of Policy Modeling 28, 739-750.

Soytas, U. and R. Sari (2009) Energy Consumption, Economic Growth, and Carbon Emissions: Challenges Faced by an EU Candidate Member. Ecological Economics 68:6, 1667-1675.

Stern, D. I. (2000) Multivariate Co-integration Analysis of the Role of Energy in the US Macro-economy. Energy Economics 22, 267-283.

Yu, E. S. H. and B. K. Hwang (1984) The Relationship between Energy and GNP: Further Results. Energy Economics 6, 186-190.

Yu, E. S. H. and J. Y. Choi (1985) The Causal Relationship between Energy and GNP: An International Comparison. Journal of Energy and Development 10, 249-272.

Yu, E. S. H. and J. C. Jin (1992) Co-integration Tests of Energy Consumption, Income, and Employment. Resources and Energy 14, 259-266.

Zamani, M. (2007) Energy Consumption and Economic Activities in Iran. Energy Economics 29:6, 1135-1140.

Zhang, X. P. and X. M. Cheng (2009) Energy Consumption, Carbon Emissions, and Economic Growth in China. Ecological Economics 68:10, 2706-2712.

Table 1
Studies Showing Various Types of Causality from Energy
Consumption to Economic Growth

Country            Authors            Period         Methodology

Causality from Energy Consumption to Economic Growth

USA         Stem (2000)              1948-1994   Co-integration,
                                                 Granger causality

Turkey      Soytas, et al. (2001)    1960-1995   Co-integration,
                                                 Granger causality

Taiwan      Lee and Chang (2007a)    1955-2003   Granger causality,
                                                 co-integration, VECM

Hong Kong   Ho and Siu (2007)        1966-2002   Co-integration, VEC
                                                 model

Causality from Economic Growth to Energy Consumption

USA         Kraft and Kraft (1978)   1947-1974   Granger causality

India       Cheng (1999)             1992-1995   Co-integration, ECM,
                                                 Granger causality

Pakistan    Aqeel and Butt (2001)    1955-1996   Hsiao's version of
                                                 Granger causality

                                                 Method,
                                                 Co-integration

Iran        Zamani (2007)            1967-2003   Granger causality,
                                                 Co-integration, VECM

Turkey      Karanfil (2008)          1970-2005   Granger Causality
                                                 Test, Co-
                                                 integration test

China       Zhang and Cheng (2009)   1960-2007   Granger causality

No Causality between Economic Growth and Energy Consumption

New         Fatai, et al. (2002)     1960-1999   Granger causality,
  Zealand                                        ARDL, Toda and
                                                 Tamamoto test

Turkey      Halicioglu (2009)        1960-2005   Granger causality,
                                                 ARDL, co-
                                                 integration

USA         Payne (2009)             1949-2006   Toda-Yamamoto
                                                 causality test

Turkey      Belloumi (2009)          1960-2000   Toda-Yamamoto
                                                 causality test

Bi-directional Causality between Economic Growth and Energy Consumption

Korea       Glasure (2002)           1961-1990   Co-integration,
                                                 error correction,
                                                 variance
                                                 decomposition

Canada      Ghali and El-Sakka       1961-1997   Co-integration,,
            (2004)                               VEC, Granger
                                                 causality

India       Paul and Bhattacharya    1950-1996   Co-integration and
            (2004)                               Granger causality

Turkey      Erdal, et al. (2008)     1970-2006   Pair-wise Granger
                                                 causality, Johansen
                                                 co-integration

Table 2
Studies Showing Different Energy Consumption and Economic
Growth Causality for the Selected Countries

Countries          Causality Relationship

            GDP [right            EC [right
            arrow] EC             arrow] GDP

India       Cheng (1999)          Masih (1996)

Japan       Cheng (1998),         Soytas and Sari
            Lee (2006)            (2003)

Korea       Yu and Choi (1985),   Oh and Lee (2004)
            Soytas and Sari
            (2003)

Malaysia    Ang (2008)            Chiou-Wei, et al.
                                  (2008)

Turkey      Lise and Van          Murray and Nan
            Montfort (2007),      (1996), Soytas, et
            Karanfil (2008)       al. (2001), Soytas
                                  and Sari (2003)

USA         Kraft (1978),         Stern (2000),
            Abosedra and          Soytas and Sari
            Baghestani (1989)     (2006), Bowden
                                  and Payne (2009)

Countries           Causality Relationship

            EC [left arrow]    GDP--EC
            [right arrow]GDP

India       Paul and           Soytas and Sari
            Bhattacharya       (2003)
            (2004)

Japan       Erol and Yu        --
            (1987)

Korea       Glasure (2002)     --

Malaysia    --                 Masih (1996)

Turkey      Erdal, et          Altinay and Karagol
            al. (2008)         (2004), Altinay and
                               Karagol (2007),
                               Karanfil (2008),
                               Soytas and Sari
                               (2009), Halicioglu
                               (2009)

USA         Lee(2006)          Akarca and Long
                               (1980), Yu and
                               Hwang (1984), Yu
                               and Choi (1985), Yu
                               and Jin (1992),
                               Cheng (1995),
                               Soytas and Sari
                               (2003), Chiou-Wei,
                               et al. (2008), Payne
                               (2009)

Table 3
Causal Relationships between Energy Consumption and
Economic Growth for Multi-Country Studies

Authors         Period       Countries

Soytas and      1950-1992   G-7 Countries
Sari (2003)

Lee (2005)      1975-2001   18 Developing
                            Countries

Lee (2006)      1960-2001   11 Developed
                            Countries

Soytas and      1960-2004   G-7 Countries
Sari (2006)

Lee and         1965-2002   22 Developed
Chang (2007b)   1971-2002   Countries, 18
                            Developing
                            Countries

Chiou-Wei,      1954-2006   Asian
el al. (2008)               Countries and
                            USA

Chang, el       1970-2010   12 Asian
at (2013)                   Countries

Authors          Methodology           Causality Relationship

Soytas and      Co-              EC [left arrow][left arrow] GDP
Sari (2003)     integration      (Argentina)
                and Granger      GDP [right arrow] EC (Italy, Korea)
                causality        EC [right arrow] GDP (Turkey,
                                 France, Japan, Germany)

Lee (2005)      Panel VECM       EC [right arrow] GDP

Lee (2006)      Granger          GDP--EC(Germany,
                causality test   UK)
                                 EC [left arrow][right arrow] GDP
                                 (Sweeden, USA)
                                 EC [right arrow] GDP (Belgium,
                                 Netherlands, Canada,
                                 Switzerland)

Soytas and      Multivariate     GDP [right arrow] EC (Germany)
Sari (2006)     co-              EC [right arrow] GDP (France, USA)
                integration,     EC [left arrow][right arrow] GDP
                ECM,             (Canada, Italy, Japan, UK)
                generalised
                variance
                decompositions

Lee and         Panel VARs and   GDP [right arrow] EC (developing
Chang (2007b)   GMM              countries)
                                 EC [left arrow][right arrow] GDP
                                 (developed countries)

Chiou-Wei,      Granger          GDP--EC(USA,
el al.(2008)    causality        Thailand, South Korea)
                                 GDP [right arrow] EC (Philippines,
                                 Singapore)
                                 EC [right arrow] GDP (Taiwan, Hong
                                 Kong, Malaysia,
                                 Indonesia)

Chang, el       Panel            EC--GDP (China,
at (2013)       causality        Indonesia, Japan,
                analysis         Malaysia, Pakistan,
                                 Philippines, Singapore,
                                 South Korea, Taiwan)
                                 EC [right arrow] GDP (Philippines)
                                 GDP [right arrow] EC (India)
                                 EC [left arrow][right arrow] GDP
                                 (Thailand and Vietnam)

Table 4
Descriptive Statistics of Variables Used in the Analysis

Variable (a)     Mean      Standard Deviation

GDP             833139         1844001.00
G                45.30           96.74
O                58.62           126.67
E               243.83           548.80
C                72.58           164.07
L                34.14           67.79
K              164277.70       348603.40
Deu              0.35             0.48
Deis             0.08             0.26
Dnamerica        0.05             0.22
Dlamerica        0.15             0.36
Dasia            0.15             0.36
Dpacific         0.05             0.22
Dmiddle          0.48             0.50
Dcold            0.20             0.40

(a) Definitions of variables are given in Table 5.

Table 5
Variable Definitions and Expected Signs

Variables   Variable Description                               Expected
                                                                 Sign

GDP         Gross domestic product (million US $)
L           Total labor force (millions)                         +ve
K           Gross capital formation (million US $)               +ve
G           Gas domestic consumption (million cubic meters)      +ve
O           Oil products domestic consumption (million tons)     +ve
E           Electricity domestic consumption (terawatt hour)     +ve
C           Coal and lignite domestic consumption                +ve
            (million tons)
Deu         Deu=1 if the observation belongs to European
            region,  otherwise 0
Deis        Dcis=1 if the observation belongs to
            Commonwealth of Independent States region,
            otherwise 0
Dnamerica   Dnamerica=1 if the observation belongs to
            North American region, otherwise 0
Dlamerica   Dlamerica=1 if the observation belongs to
            Latin American region, otherwise 0
Dasia       Dasia=1 if the observation belongs to
            Asian region,  otherwise 0
Dpacific    Dpacific=1 if the observation belongs to
            Pacific region, otherwise 0
Dmiddle     Dmiddle=1 if the observation belongs to
            middle income country, otherwise 0
Dcold       Dcold=1 if the observation belongs to a
            country which is located in D and/or E
            Koppen climate classification system,
            otherwise 0

Table 6
Estimates of the Inter Country Translog Production Function

Variable             Coefficient

                       9.522 *
Constant               (0.581)
                      -1.503 *
Lngas                  (0.125)
                       1.596 *
Lnoil                  (0.342)
                      0.564 **
Lnelectric             (0.305)
                       0.198 *
Lncoal                 (0.070)
                       0.901 *
Lntlabor               (0.204)
                      -0.762 *
Lngcapital             (0.088)
                        0.003
[Lngas.sup.2]          (0.006)
                      0.106 **
[Lnoil.sup.2]          (0.057)
                       0.142 *
[Lnelec.sup.2]         (0.044)
                      -0.006 *
[Lncoal.sup.2]         (0.002)
                      -0.051 *
[Lntlabor.sup.2]       (0.013)
                       0.089 *
[lngcapital.sup.2]     (0.006)
                      -0.267 *
Lngasoil               (0.029)
                       -0.009
Lngaselec              (0.020)
                      -0.031 *
Lngascoal              (0.008)
                       0.057 *
Lngastlabor            (0.019)
Lngasgcapita           0.211 *
1                      (0.016)
                        0.003
Lnoilelec              (0.076)
                       0.091 *
Lnoilcoal              (0.016)
                        0.075
Lnoiltlabor            (0.048)
                      -0.159 *
Lnoilgcapital          (0.041)
                       0.047 *
Lneleccoal             (0.021)
                      -0.071 *
Lnelectlabor           (0.034)
                      -0.123 *
Lnelectgcapital        (0.038)
                        0.008
Lncoaltlabor           (0.006)
                      -0.058 *
Lncoalgcapital         (0.009)
                       -0.063
Lntlaborgcapital       (0.026)
                        0.026
Dcold                  (0.032)
                       0.638 *
Deu                    (0.041)
                      -0.312 *
Dcis                   (0.065)
                       -0.030
Dnamerica              (0.077)
                       0.468 *
Dlamerica              (0.037)
                       0.165 *
Dasia                  (0.039)
                       0.681 *
Dpacific               (0.062)
                      -0.130 *
Dmiddle                (0.032)

Estimates obtained by using FGLS procedure.
Standard error of the Coefficient is given in the parenthesis.
* and ** represent statistical significance at 5 percent and 10
percent level of significance respectively.

Table 7
Elasticities Estimates of Different Inputs

Input   Elasticity Estimate

G            0.0018 *
O            0.1914 *
E            0.4521 *
C           -0.0328 *
L            0.0438 *
K            0.4280 *

* Represents statistical significance at 5 percent
level of significance.