Available online at www.sciencedirect.

com

Journal of Policy Modeling 34 (2012) 879889

Energy consumption, economic growth and environmental pollutants in Indonesia

Yaghoob Jafari , Jamal Othman, Abu Hassan Shaari Mohd Nor
Faculty of Economics and Management, Universiti Kebangsaan Malaysia (UKM), 43600 Bangi, Selangor, Malaysia Received 12 March 2012; received in revised form 22 April 2012; accepted 30 May 2012 Available online 12 June 2012

Abstract This paper examines the long run Granger causality relationship between economic growth, carbon dioxide emissions and energy consumption from 1971 to 2007 in Indonesia, controlling for capital stock and urban population. Using TodaYamamoto (TY) procedure, it has been found that there was no relationship between these variables except the causality effect that runs from urban population to energy consumption. The absence of these linkages suggests that energy conservation strategies in Indonesia may not produce desirable effect on emission reductions, and Indonesia does not have to relinquish economic growth. However, the results may not be sufcient to warrant Indonesias choice of specic policies and strategies to limit carbon emissions in the context of combating global climate change. This is because the study does not take into account net carbon release from deforestation and forest degradation. There seems to be substantial scope for Indonesia to limit her CO2 emission via reducing deforestation and improved forest management through the REDD Plus framework. 2012 Society for Policy Modeling. Published by Elsevier Inc. All rights reserved.
JEL classication: Q43; Q53; Q56 Keywords: Carbon dioxide emissions; Energy consumption; Economic growth in Indonesia

1. Introduction One of the most important events over the last decade until today is global warming. Many studies have acknowledged that carbon dioxide (CO2 ) emissions are among the most common

Corresponding author. Tel.: +60 3 89213762; fax: +60 3 89215789. E-mail address: Yaghoob.jafari@gmail.com (Y. Jafari).

0161-8938/$ see front matter 2012 Society for Policy Modeling. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpolmod.2012.05.020

880

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

and greatest factors contributing to global warming (IPCC, 1996). Consumptions of conventional fuel, including crude oil, natural gas and coal stand for the largest portion (56.6%) of Green House Gases (GHGs) emitted to the atmosphere (IPCC, 1996). International organizations, especially the United Nation-based institutions have strived to reduce the amount of global warming substances. This is especially done through the binding and intergovernmental agreements known as the Kyoto Protocol (see Halicioglu, 2009). Indonesia has been the rst large developing nation and the third top producer of green house gases to ratify the protocol, thus committing herself to reduce GHG emissions. The country announced her target of GHG emissions reduction about 2641% below the Business as Usual (BAU) scenario with a targeted economic growth rate of 7% by 2020 (Copenhagen Accord, 2009). An important aim of Indonesias national energy policies has been to reduce the domestic fossil-based energy consumption (National Council on Climate Change, 2009). The Indonesian government believes that reducing the energy consumption may promote the sustainable development efforts and energy security programme without causing a decline in economic growth rate. However, it is not clear whether reducing energy consumption by itself can be a solution to the environmental degradation problem. Furthermore, if energy consumption and economic growth were truly intertwined, energy consumption reduction may in fact hinders economic growth. Additionally, considering the targeted economic growth rate of 7% through 2020, the vital question is that is it possible to have sustained growth in the long run without accumulation of pollution? In particular, if economic growth causes CO2 emission then growth policies could lead to a degradation in environmental quality. However, depending on the nature of the long run relationship between CO2 emissions, economic growth, and energy consumption in different economies, countries may resort to different policy options in contributing to the ght against global warming (Soytas & Sari, 2006a, 2006b). This paper attempts to shed more light on the energy consumption, carbon emissions and economic growth relations in Indonesia that may provide more information to the policy makers in the country in choosing from a variety of policy options in line with the Kyoto Protocol and its potential successor, the REDD Plus framework. Indonesia appears to be an interesting case study given that it is the third top producer of green house gases as of 2007 (WBI, 2010). In recent years, Indonesias carbon emissions have grown exponentially. The amount of CO2 emission over the 19712007 period has increased by 919%, while the amount of fossil fuel consumption has increased at a much slower pace of some 430% (calculated based on WBI, 2010). The choice of Indonesia is also motivated by the fact that there have been no specic studies on Indonesia that have employed advanced techniques to link energy consumption, CO2 emissions and economic growth in a coherent framework. Well acknowledging Indonesias commitment to limit the rise in global temperature via CO2 reductions while establishing relevant national policies that safeguards their interest, the purpose of this paper is to investigate the relationship between Indonesias economic growth, energy consumption and pollutant emissions in a multivariate framework controlling for other crucial variables such as urban population and capital stock. 2. Literature review Generally, there are three research strands in investigating the relation between economic growth, carbon emission and energy use. The rst strand focuses on the environmental pollutants and economic growth nexus. It is closely related to testing the validity of the so-called Environmental Kuznets Curve (EKC) hypothesis, which argues an inverted U-shaped relationship between the

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

881

level of environmental degradation and income growth. So, in the development process, countries will recognize their levels of environmental degradation would increase until some income threshold is met and then decrease afterwards. Following the seminal work of Grossman and Krueger (1992), a large and growing body of literature has investigated the relationship between economic growth and environmental pollution. The outcomes from many empirical studies, however, were rather controversial and seriously being criticized for the lack of feedback from environmental pollutants to economic output as income is assumed to be an exogenous variable. The second strand focuses on the link between economic output and energy consumption since the carbon emissions are generally attributed to the burning of fossil fuels. Ever since the original empirical research of Kraft and Kraft (1978), an increasing body of literature exists on the association between energy consumption and economic growth. Nevertheless, these researches also produce conicting results. Huang, Hwang, and Yang (2008), provided a good review on the empirical results from causality tests. The results of Granger causality between energy consumption and economic growth as gathered from the literature can be classied under the following three categories; (i) there exists no causal relationship between energy consumption and economic growth; (ii) economic growth leads energy consumption; and (iii) energy consumption leads economic growth. Lastly, the third strand combines the last aforementioned nexus in order to capture the inter-temporal linkages in output, energy and pollution. This approach examines the dynamic relationships between economic growth, environmental pollutants and energy consumption simultaneously. In all the research strands, there are few studies that attempted to examine the relationship between economic growth, environmental pollutants and energy consumption in Indonesia. Applying the error correction method on the Indonesian economy, Masih and Masih (1996) and Asafu-Adjaye (2000) found unidirectional Granger causality running from GDP to energy consumption. Fatai, Oxley, and Scrimgeour (2004) and Chiou-Wei, Chen, and Zhu (2008) on the other hand found unidirectional link from energy consumption to GDP. However, Soytas and Sari (2009) and Ozturk, Aslan, and Kalyoncu (2010) found no cointegration between economic growth and GDP in Indonesia. As far as the cause and effect relationship between the economic growth, energy consumption and carbon emissions is concerned, existing literature reveals that ndings of empirical studies differ substantially and are not conclusive. It can also be asserted that the nature and direction of causality may vary from one country to the other. 3. Methodology To study the causality among the variables, the Error Correction Model (ECM) or Vector Autoregressive (VAR) could be employed. One could conduct an ECM model if all the variables present in the model are integrated of the same order and are cointegrated. The VAR in the rst differences of the variables can be utilized if all the variables are integrated of order one. Hence, pre-testing on the variables is required whether the variables are stationary, integrated, or cointegrated. Toda (1995) pointed out that the causality inference in ECM might suffer from the severe bias as the pre-test for cointegration in ECM (proposed by Johansen, 1988) are extremely sensitive to the value of the nuisance parameters in the known sample. Toda and Phillips (1993) and Sims, Stock, and Watson (1990) indicate that in a system that includes unit roots, standard Wald statistics based on Ordinary Least-Squares (OLS) estimation of level VAR model for testing coefcient restrictions have nonstandard asymptotic distributions and cannot be applied to mixed integration orders. Difculties related with estimation of such VAR models have been illustrated by Park and Phillips (1989) and Toda and Phillips (1993, 1994), among others.

882
140 120 100 80 60 40 20 0
Index

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

Capital

1200 1000 800 600 400 200 0

Index

CO2

600 500 400 300 200 100 0 1971 1995 2001 2007

Index Energy Consumpon

1000 800 600 400 200 0

Toda and Yamamoto (1995) (hereafter TY) proposed an augmented VAR approach which can be employed whether the variables are stationary, integrated of any arbitrary level or any arbitrary order of cointegration. As pointed out by Zapata and Rambaldi (1997), the TY approach has much practical appeal when there is uncertainty about the order of integration. Henceforth, in this study, we employ the TY procedure to examine the linkages among the economic growth, energy use and the environmental pollutants. 4. Empirical analysis This paper uses annual time series data for the period 19712007 which include real GDP (at constant 2000 prices), energy consumption (EC) (kt of oil equivalent), CO2 (kt of CO2 equivalent, excluding deforestation and forest degradation), urban population (UPOP) and capital stock (K) (at constant 2000 prices) for Indonesia. The selection of the starting period was constrained by the availability of data on energy consumption. Following Soytas, Sari, and Ewing (2007), we employed total data rather than per capita data as dividing the variables by number of population will only scale down the variables. In addition, as pointed out by Friedl and Getzner (2003), based on the spirit of the Kyoto Protocol, it would be appropriate to use total emissions rather than per capita as it calls for the reduction in the percentage of emission. The data, except for capital stock, were obtained from World Development Indicators (2010) published by the World Bank. The time series capital stocks are obtained from the empirical study by Pierre (2008). The data were converted into natural logarithms so that they can be interpreted in growth terms. Following Zhang and Cheng (2009), in order to illustrate the change in the trend of each series in the same scale, we construct an index for each series using 1971 as the base year. The diagrams in Fig. 1 show the trend of the respective series. As illustrated in the gure all the variables (except for capital stock) were increasing over the period. The effect of ASEAN recession (19971998) is also clear from the sudden decrease in all variables except for urban

2006 2001 1996 1991 1986 1981 1976 1971

Index GDP

1978

1999

2006

1985

1992

1989

600 500 400 300 200 100 0

Index

1971

1977

Urban Populaon

1983

2006 2001 1996 1991 1986 1981 1976 1971

1981

1986

1991

1996

2001

Fig. 1. Trends of variables (before taking logarithm, 1971 = 100).

1971

1976

2006

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

1200 Index 1000 800 600 400 200 0 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007

883

Energy Consumpon

CO2

GDP

Urban Populaon

Capital

Fig. 2. Trends of all variables (before taking logarithm, 1971 = 100).

population. All the trends as depicted in Fig. 2 suggest that the series of GDP, EC, CO2 , UPOP and K tend to move not closely together over time. 4.1. Unit root test In order to employ the TodaYamamoto (TY) procedure, we need to identify the maximum order of integration for the series under consideration. In order to test for the stationarity of each variable of interest and to nd the maximum order of cointegration, three unit root tests namely augmented Dickey and Fuller (1979) (ADF), Phillips and Perron (1988) (PP), Kwiatkowski, Phillips, Schmidt, and Shin (1992) (KPSS) were conducted. In reference to previous studies, ADF and PP tests which were designed on the basis of the null hypothesis that a series is I(1) have a low power of rejecting the null. Therefore, in order to obtain more robust results we relied on the KPSS unit root test. The unit root test results are shown in Table 1. The results are rather contradictory where the maximum order of integration for all variables do not exceed two. Perron (1989) pointed that the failure to allow for an existing break leads to a bias that reduces the capability to reject a false unit root null hypothesis. In other words, for the series that are found to be I(1), there may be a possibility that they are, in fact, stationary around the structural break(s), I(0), but are erroneously classied as I(1). To overcome this, we need to take into account the structural break for possible changes in intercept, slope or both. Zivot and Andrews (1992) (hereafter ZA) proposed a unit root testing procedure which allows for an estimated break in the trend function under the alternative hypothesis. ZA procedure has been employed to test the order of integration of series. The results from ZA test are summarized in Table 2. Similar to the above result for the order of integration, ZA tests show that the maximum order of integration does not go beyond two. Hence, we identify the maximum order of integration to be 2(d = 2). 4.2. Granger causality Since the variables of interest are not integrated of the same order, the TY procedure is possibly the most appropriate approach to test for the granger causality. In order to apply the TY procedure, the optimal lag length of VAR is required. To estimate the optimal lag length several criteria are used. Table 3 shows the results from the respective lag length criteria. The LR and SC suggested the lag length of one, while FPE, AIC, and HQ pointed out the optimal lag length of 4. Nevertheless, the diagnostic test on the augmented VAR model (4 + 2)

884 Table 1 Unit root test results.

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

ADF Level Intercept GDP CO2 EC CAPU UPOP Intercept and trend GDP CO2 EC CAPU UPOP First difference Intercept GDP CO2 EC CAPU UPOP Intercept and trend GDP CO2 EC CAPU UPOP Second difference Intercept UPOP Intercept and trend UPOP

PP

KPSS

2.257283 (0) 1.524329 (0) 0.988760 (0) 1.510192 (0) 2.263132 (1) 1.674478 (1) 2.690423 (0) 1.362511 2.668277 (0) 2.894598 (9)

2.257283 2.041306 1.024669 1.461284 0.288513 1.393169 2.582414 1.362511 2.700209 2.62066

0.723253** 0.726798** 0.722432** 0.519435** 0.729097** 0.180775** 0.129514*** 0.130996*** 0.095967 0.185079**

4.142123* (0) 5.414039* (0) 6.171673* (0) 6.371992 (0) 0.526617 (0) 4.420723* (0) 5.447265* (0) 6.250257* (0) 6.289404* 1.978435 (0) 6.364775* (0) 6.675155* (0)

4.142123* 5.407708* 6.170652* 6.371992* 3.299517** 4.387632* 5.641227* 6.251114* 6.292105* 1.669746 6.559076* 15.77034*

0.345203 0.195001 0.157520 0.059049 0.581824** 0.071023 0.120347 0.088146 0.047845** 0.173238**

0.339468 0.050000

Notes: All variables in natural logs. The null hypothesis of all tests except KPSS are unit roots. The null of KPSS, on the other hand, states that the variable is stationary. * Signicance at 1%. ** Signicance at 5%. *** Signicance at 10%.

signies that the VAR (6) is not stable. Therefore, the augmented VAR (K + d = 3) is estimated by employing the Seemingly Unrelated Regression (SUR) framework. The results of robustness test on the augmented VAR (3) are summarized in Table 4. The table shows that the adjusted R-square values are pretty high and all the equations are robustly explained. The JarqueBera (JB) test results show that there are no serious violations of normality, only the residual from energy consumption (EC) equation is found to reject the null assumption of normality. The results of BreuschGodfrey (BG) test reveal that there is no serial correlation for all equations at 5% signicance level. The White test indicates that there is no heteroscedasticity for all equations. Lagrange multiplier test points out that there is no autoregressive conditional hetroscedasticity (ARCH) while the Ramsey Reset test signies that we cannot reject the null of no model misspecication and the parameters seem stable for all equations. Furthermore the

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889 Table 2 ZA unit root test. Variable GDP Level 9.001(A)** 3.775(B) 6.562(C)** 3.922(A) 3.585(B) 3.851(C) 2.953(A) 3.059(B) 6.104(C)** 6.560(A)** 2.90(B) 6.522(C)** 0.015(A) 1.768(B) 1.755(C) Break 1998 1998 1998 1998 1978 1977 1990 1996 1990 1998 1994 1999 1981 1997 1997 First deference 4.587(B)** 5.338(A)** 4.986(B)** 5.713(C)** 6.779(A)** 6.608(B)** 6.547(B)** 4.396(A) 3.83(B) 4.880(C) Break 2000 1980 1984 1979 1996 1991 Second deference Break

885

CO2

EC

1999 1981 1983 1981 5.818(A)** 5.061(B)** 6.001(C)**

UPOP

1988 1996

Notes: The letters in parentheses indicate the models AC . Model A allows for a change in the level of the series; model B allows for a change in the slope of the trend of a series, while model C combines both the changes. Break denotes the time of the structural change. ** Signicance at the 5% level.

Table 3 VAR lag order selection criterion. Lag 0 1 2 3 4 LogL 136.1598 373.0533 396.8922 430.3168 467.5144 LR NA 387.6439* 31.78520 34.43746 27.05277 FPE 2.43e10 6.56e16 7.87e16 6.46e16 6.28e16* AIC 7.949080 20.79111 20.72074 21.23132 21.97057* SC 7.722336 19.43065* 18.22656 17.60343 17.20895 HQ 7.872787 20.33336 19.88153 20.01065 20.36843*

Notes: LogL, log likelihood; LR, sequential modied likelihood ratio test statistic; FPE, nal prediction error; AIC, Akaike information criterion; SC, Schwarz information criterion. HQ, Hannan-Quinn information criterion (each test at 5% level of signicance). * Lag order selected by the criterion.

Table 4 Diagnostic test results. Dependent variable GDP CO2 EC K UPOP Adj R-Square 0.9966 0.9874 0.9965 0.8261 0.9999 JB test 1.329 0.131 8.728* 0.952 0.059 BG test 4.373838 2.474972 0.105155 5.917367 3.975522 White test 1.4207 0.5333 0.8357 0.851 2.7619 ARCH-LM 1.1747 0.0574 0.1622 0.1943 3.4924 Ramsey test 0.0437 0.0866 0.0391 0.2406 0.0021

Notes: JB test null is normality. BG test null is no serial correlation. ARCH LM null is no ARCH effects. White test null is no heteroscedasticity. Ramsey Reset test null is no specication errors with one tted term using LR. * Signicance at the 5%.

886

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

Table 5 The results of Wald test statistics. GDP GDP CO2 EC K UPOP 0.141490 0.936529 1.203244 0.235194 CO2 0.062301 0.102520 0.006777 1.989838 EC 0.172312 0.107633 0.191544 0.620058 K 0.117904 0.470915 1.582707 1.193637 UPOP 2.234239 0.368408 3.146550** 0.967240

Wald statistics follow an asymptotic Chi-square distribution with p degrees of freedom. Signicance implies that the column variable Granger causes the row variable. ** Signicance at the 5%.

estimated VAR is stable (stationary) as all roots have modulus less than one and lie inside the unit circle. Hence, we can proceed with the Granger causality tests. The results of Wald test statistics for the Granger non-causality hypotheses are presented in Table 5. As can be seen from the table, there is only one signicant result at the 5% signicance level. Urban population Granger cause energy consumption in the long run but the reverse does not hold since the causality is not bi-directional, as is expected. Hence, urban population level improves the forecasts of energy consumption. Most industrial zones inevitably are located close to urban population areas. Transportation activities are also more extensive in the urban sector. Hence, accelerating industrialization or urbanization processes naturally would have positive effects on energy use. This reects the importance of urban population segment in terms of government policy targets in the pursuit of clean energy and technology implementation, consistent with sustainable development objectives and energy security in Indonesia. The results fail to reject the null hypothesis that there is no causality (in any direction) between carbon emissions and economic growth. This nding is striking as it indicates that much of Indonesias value adding activities has not been signicantly associated with CO2 emissions from conventional industrial processes. While the nding that energy consumption does not lead to CO2 emissions is quite counter-intuitive, other studies seem to lend support to the result. For instance, the study by Ozturk and Acaravci (2010) on the Turkeys economy. Soytas et al. (2007) also found that energy consumption does not granger cause CO2 emission in the United States. Such unexpected nding may be due to changes in the mix of energy use in which the ratio of fossil-based fuels may be increasing at a much faster pace relative to renewable energy or gas while total energy consumption may only increase steadily. The increasing divergence between the trend of energy consumption and CO2 emissions as depicted in Fig. 2 seems to lend support to this suggestion. Likewise, the results fail to reject the null hypothesis that there is no Granger causality (in any direction) between real GDP and energy consumption at the 5% signicant level in Indonesia. In the modelling sense, neither energy nor economic growth seems to improve the forecasts of each other. This result is consistent with the aforementioned nding that economic growth does not lead to CO2 emissions and vice versa. These results are in line with that of Hung and Shaw (2002), which pointed out that in low income level countries, no relationship exists between energy consumption and economic growth. In addition, Akarca and Long (1980), Erol and Yu (1987), Yu and Hwang (1984), Yu and Choi (1985), Yu, Jue, and Youmin (2008), Yu and Jin (1992), Soytas and Sari (2003), Altinay and Karagol (2004) also found no causal relationships between the two.

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

887

4.3. Policy implications One of the greatest challenges facing Indonesia, the third largest producer of environmental pollutants is climate change. Considering that the preamble to the UNFCCC acknowledges that the global nature of climate change calls for the widest possible collaboration of countries, Indonesia has pledged to trim her carbon emissions by 26% by 2020 and extend it up to 41% under the international assistance, using the Business as Usual (BAU) scenario as the base. However, the country is confronted with the issue of balancing its energy usage, economic growth and green house gases. To understand the relationships between national CO2 emissions and GDP and to assist current deliberations about emission projections, this paper investigated the causal relation between pollutant emissions, economic growth, energy consumption, capital stock and urban population in Indonesia for the 19712007 period by applying the Toda and Yamamoto approach. The estimation results show that there is no signicant relationship between GDP, CO2 emissions and energy consumption in Indonesia. This indicates that while the implementation of energy conservation policies in Indonesia may not affect Indonesias economic growth signicantly, such policies may also be ineffective to reduce CO2 emissions in the country. The latter implication may be due to the small distributive and total effects of CO2 emissions attributed to GDP and sectoral growth, respectively. However, our causality tests indicate that there was a unidirectional Granger causality running from urban population to energy consumption. This suggests increases in the level of energy consumption in Indonesia coincide more strongly with increases in urban population rather than overall economic growth. Inevitably, urban population growth, as in many developing countries, is linked with increased modernity, industrialization processes, transportation, and heat island effects which augment energy consumption. This reects the importance of energy conservation strategies targeting the urban sector. Policy options may focus on urban energy demand management encompassing urban land use regulations, infrastructure design and investment, industrial zoning, as well as market-based instruments such as energy pricing, emissions trading, taxation and road pricing for urban areas. Such strategies may induce more efcient energy use and/or a switch from conventional energy to cleaner or renewable energy. The inter-sectoral economic impacts that result from changes in energy mix may also be considerable. For instance, the move to increase the use of bio-diesel from palm oil potentially leads to adverse impacts in the food and other chemical sector. Energy conservation strategies may not pose signicant impact on the countrys GDP, but may have substantial inter-sectoral impacts in which some sectors may gain, while some others may lose. It will thus be an interesting future study to appraise the inter-sectoral impacts of energy conservation or CO2 emissions reductions in Indonesia. 5. Conclusion It has been shown in our empirical work that Indonesias economic growth, CO2 emissions and energy consumption have no causality effects on each other. The results seem to imply that energy conservation strategies in Indonesia may not lead to signicant emission reductions, and the country may be able to enjoy double dividends achieving both economic growth and environmental protection simultaneously. Nevertheless, the econometric results may not be sufcient to warrant Indonesias choice of specic policies and strategies to limit carbon emissions in the context of combating global climate change. This is especially because the study only accounted for CO2 emission that emanated from conventional industrial processes and energy use, while disregarding net carbon release from deforestation, forest degradation and peat swamp conversion. There

888

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

seem to be more scope for Indonesia to limit her CO2 emission via the REDD Plus frameworks which calls for reduction in emissions through reduced forest conversion and degradation as well as through better forest management strategies. Based on the causality framework using a linear logarithmic specication, this study has demonstrated that the EKC hypothesis may not apply to Indonesia, if one would consider CO2 emissions from conventional sources as an appropriate representation of environmental pollutants. It will be highly benecial for future work to look into the same policy issue but taking into account all sources of CO2 emissions in Indonesia, including deforestation, forest degradation and peat swamp conversion into agricultural areas. Acknowledgements We are very grateful to four anonymous referees whose constructive comments have helped to improve the quality of the paper. We are also grateful to the Editor of the Journal, A.M. Costaf for his encouragement. The usual caveats apply. References
Akarca, A. T., & Long, T. V. (1980). On the relationship between energy and GDP: A reexamination. Journal of Energy Development, 5, 326331. Altinay, G., & Karagol, E. (2004). Structural break, unit root, and the causality between energy consumption and GDP in Turkey. Energy Economics, 26, 985994. Asafu-Adjaye, J. (2000). The relation between energy consumption, energy prices and economic growth: Time series evidence from Asian Development Countries. Energy Economics, 22, 615625. Chiou-Wei, S. Z., Chen, C. F., & Zhu,. Z. (2008). Economic growth and energy consumption: Evidence from linear and nonlinear Granger Causality. Energy Economics, 30, 30603076. Copenhagen Accord. (2009). United Nations framework convention on climate change. Dickey, D. A., & Fuller, W. A. (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Society, 75, 427431. Erol, U., & Yu, E. S. H. (1987). On the causal relationship between energy and income for industrialized countries. Journal of Energy and Development, 13, 113122. Fatai, K., Oxley, L., & Scrimgeour, F. G. (2004). Modeling the causal relationship between energy consumption and GDP in New Zealand, Australia, India, Indonesia, the Philippines and Thailand. Mathematics and Computers in Simulation, 64, 431445. Friedl, B., & Getzner, M. (2003). Determinants of CO2 emissions in a small open economy. Ecological Economics, 45, 133148. Grossman, G., & Krueger, A. (1992). Environmental impacts of a North American free trade agreement. Centre for economic policy research discussion paper, no. 644. Halicioglu, F. (2009). An econometric study of CO2 emissions, energy consumption, income and foreign trade in Turkey. Energy Policy, 37, 11561164. Huang, B. N., Hwang, M. J., & Yang, C. W. (2008). Causal relationship between energy consumption and GDP growth revisited: A dynamic panel data approach. Ecological Economics, 67, 4154. Hung, M. F., & Shaw, D. (2002). Economic growth and the environmental Kuznets curve in Taiwan: A simultaneity model analysis. Mimeo: Department of Economics, National Cheng-Chi University. IPCC Second Assessment on Climate Change. (1996). International panel on climate change. Cambridge, UK: Cambridge University Press. Johansen, S. (1988). Statistical analysis of cointegration vectors. Journal of Economic Dynamics and Control, 12, 231254. Kraft, J., & Kraft, A. (1978). On the relationship between energy and GNP. Journal of Energy and Development, 3, 401403. Kwiatkowski, D., Phillips, P. C. B., Schmidt, P., & Shin, Y. (1992). Testing the null hypothesis of stationary against the alternative of a unit root. Journal of Econometrics, 54, 159178.

Y. Jafari et al. / Journal of Policy Modeling 34 (2012) 879889

889

Masih, A. M. M., & Masih, R. (1996). Energy consumption, real income and temporal causality: Results from a multi-country study based on cointegration and errorcorrection modeling techniques. Energy Economics, 18, 165183. National Council on Climate Change. (2009). National Economic, Environment and Development Study (NEEDS) for climate change: Indonesia country study. Jakarta: National Council on Climate Change. Ozturk, I., & Acaravci, A. (2010). CO2 emissions, energy consumption and economic growth in Turkey. Renewable and Sustainable Energy Reviews, 14, 32203225. Ozturk, I., Aslan, A., & Kalyoncu, H. (2010). Energy consumption and economic growth relationship: Evidence from panel data for low and middle income countries. Energy Policy, 38, 44224428. Park, J. P., & Phillips, P. C. B. (1989). Statistical inference in regression with integrated process: Part 2. Econometric Theory, 5, 95131. Perron, P. (1989). The great crash, the oil price shock and the unit root hypothesis. Econometrica, 57, 13611401. Phillips, P. C. B., & Perron, P. (1988). Testing for a unit root in time series regressions. Biometrica, 75, 335346. Pierre V. D. E. (2008). Capital formation and capital stock in Indonesia, 19502007. Working paper, Australian National University, 24. Sims, C., Stock, J., & Watson, M. (1990). Inference in linear time series models with unit roots. Econometrica, 58, 113144. Soytas, U., & Sari, R. (2003). Energy consumption and GDP: Causality relationship in G7 countries and emerging markets. Energy Economics, 25, 3337. Soytas, U., & Sari, R. (2006a). Energy consumption and income in G7 countries. Journal of Policy Modeling, 28, 739750. Soytas, U., & Sari, R. (2006b). Can China contribute more to the ght against global warming? Journal of Policy Modeling, 28, 837846. Soytas, U., & Sari, R. (2009). Are global warming and economic growth compatible? Evidence from ve OPEC countries. Applied Energy, 86, 18871893. Soytas, U., Sari, R., & Ewing, B. T. (2007). Energy consumption, income and carbon emissions in the United States. Ecological Economics, 62, 482489. Toda, H. Y., & Phillips, P. C. B. (1993). Vector autoregressions and causality. Econometrica, 61, 13671393. Toda, H. Y., & Phillips, P. C. B. (1994). Vector autoregression and causality: A theoretical overview and simulation study. Econometric Reviews, 13, 259285. Toda, H. Y., & Yamamoto, T. (1995). Statistical inference in vector autoregressions with possibly integrated processes. Journal of Econometrics, 66, 225250. Toda, H. Y. (1995). Finite sample performance of likelihood ratio tests for cointegrating ranks in vector autoregressions. Econometric Theory, 11, 10151032. World Development Indicators. (2010). World Bank. Yu, E. S. H., & Choi, J. Y. (1985). The causal relationship between energy and GNP: An international comparison. Journal of Energy and Development, 10, 249272. Yu, E. S. H., & Hwang, B. K. (1984). The relationship between energy and GNP: Further results. Energy Economics, 6, 186190. Yu, S. H., & Jin, J. C. (1992). Cointegration tests of energy consumption, income and employment. Resources and Energy Economics, 14, 259266. Yu, W., Jue, G., & Youmin, X. (2008). Study on the dynamic relationship between economic growth and china energy based on cointegration analysis and impulse response function. China Population, Resources and Environment, 18, 5661. Zapata, H. O., & Rambaldi, A. N. (1997). Monte Carlo evidence on cointegration and causation. Oxford Bulletin of Economics and Statistics, 59, 285298. Zhang, X. P., & Cheng, X. M. (2009). Energy consumption, carbon emissions, and economic growth in China. Ecological Economics, 68, 27062712. Zivot, E., & Andrews, D. (1992). Further evidence of the great crash, the oil-price shock and the unit-root hypothesis. Journal of Business and Economic Statistics, 10, 251270.