Energy Sources, Part B, 6:111–117, 2011 Copyright © Taylor & Francis Group, LLC ISSN: 1556-7249 print/1556-7257 online DOI: 10.1080/15567240802533971

The Causal Relationship Between Nuclear Energy Consumption and Economic Growth in India J.-Y. HEO,1 S.-H. YOO,2 and S.-J. KWAK3 1

Department of Energy Policy, Graduate School of Energy & Environment, Seoul National University of Science & Technology, Seoul, Korea 2 Department of International Area Studies, Hoseo University, Chungnam, Korea 3 Department of Economics, Korea University, Seoul, Korea Abstract This article attempts to investigate the long-run and short-run causality issues between nuclear energy consumption and economic growth in India by using the co-integration and error-correction models. It employs the annual data covering the period 1969–2006. The results indicate that there is a uni-directional causality running from nuclear energy consumption to economic growth without any feedback effect. Thus, considering the fact that nuclear energy consumption fosters economic growth, policies for increasing nuclear energy supply investment is, therefore, likely to enhance economic growth in India. Keywords causality, economic growth, India, nuclear energy consumption

Introduction Historical records of the early 1970s show that almost 25% of global electricity was generated from oil while nuclear’s share constituted only 3%. By 2002, however, the global electricity supply structure changed. Oil’s share declined to 7.2% while nuclear constituted 16.6%. Nuclear absorbed approximately 75% of the decrease of oil’s share (17.8%). There is no doubt that nuclear energy has made a major intrusion in the electricity market (International Energy Agency, 2004). Since the 1990s the emphasis of nuclear power plant construction has shifted towards Asia and developing countries. New nuclear power plants are most attractive where energy demand growth is rapid, alternative resources are scarce, and energy supply security is a priority—China, India, Japan, and Korea. Moreover, using nuclear energy as a source of energy reduces air pollution and greenhouse gas emissions (Toth and Rogner, 2006). India is on an economic overdrive in the post liberalization era and has developed a great craving for energy. Although coal is the dominant energy source providing 72% of the electricity in India, it is an exhaustible source of energy likely to last for only another 5 or 6 decades. Amongst all the options available to India today, nuclear energy (2.9% of the fuel mix for generating power) emerges as a ray of hope (Nuclear Power Corporation of India Limited, 2007). Public policy makers in India have shown a great deal of interest in the role that nuclear energy consumption plays in economic growth. The nuclear energy infrastructure Address correspondence to Seung-Hoon Yoo, Department of Energy Policy, Graduate School of Energy & Environment, Seoul National University of Science & Technology, 172 Gongreung2-Dong, Nowon-Ku, Seoul, 139–743, Republic of Korea. E-mail: [email protected]

111

112

J.-Y. Heo et al.

of India is becoming an increasingly important component of the economy. Grover and Chandra (2006) present a scenario for the growth of electricity in India. To meet the projected demand, the article presents that the nuclear contribution towards electricity generation has to increase from the present 3% to about a quarter of the total. To proactively cope with increasing nuclear energy demand accompanying rapid economic growth, India should endeavor to uncover the causal relationship between nuclear energy consumption and economic growthto make appropriate nuclear energy policy. This task has become one of the most important agendas for India in the present and will be so in the near future. The purpose of this article is, therefore, to investigate causality between nuclear energy consumption and economic growth, and to obtain policy implications from the results. Ghosh (2006) examined the long-run equilibrium relationship between total petroleum products consumption and economic growth in India. However, only a few studies have been done on the relationship between nuclear energy consumption and economic growth. Yoo and Jung (2005) detected a uni-directional causality which runs from nuclear energy consumption to economic growth in Korea. To this end, the authors attempt to provide more careful consideration of the causality issues by applying the modern rigorous techniques of Granger causality to the Indian data. The remainder of the article is organized as follows. The next section presents an overview of the proposed methodology. The penultimate section explains the data employed and reports the empirical findings. Some concluding remarks are made in the final section.

Methodology Granger-causality and Stationarity The first attempt at testing for the direction of causality was proposed by Granger (1969). The Granger-causality test is a convenient and very general approach for detecting any presence of a causal relationship between two variables. The test is quite simple and straightforward. A time series .X/ is said to Granger-cause another time series .Y / if the prediction error of current Y declines by using past values of X in addition to past values of Y . The Granger-causality test method is selected to be used in this study over other alternative techniques because of the favorable Monte Carlo evidence reported by Guilkey and Salemi (1982) and Geweke et al. (1983), particularly for small samples in empirical works. In order to conduct to the Granger-causality test, a series of variables is required to be stationary. It has been shown that using non-stationary data in causality tests can yield spurious causality results (Granger and Newbold, 1974). Therefore, following Engle and Granger (1987), the authors first test the unit roots of X and Y to confirm the stationarity of each variable. This is done by using the Phillips-Perron (PP) test (Phillips and Perron, 1988) over alternative tests, in that the PP test is known to be robust for a variety of serial correlations and time-dependent heteroscedasticities. If any variable is found to be non-stationary, we must take the first difference and then apply the causality test with differenced data. Co-integration The concept of co-integration can be defined as a systematic co-movement among two or more economic variables over the long run. If X and Y each are non-stationary and

Nuclear Energy and Economic Growth in India

113

co-integrated, then any standard Granger-causal inferences will be invalid and a more comprehensive test of causality based on an error-correction model (ECM), should be adopted (Engle and Granger, 1987). However, if X and Y are both non-stationary and the linear combination of the series of two variables is non-stationary, then a standard Granger-causality test should be adopted (Toda and Phillips, 1993; Yoo and Kwak, 2004). Therefore, it is necessary to test for the co-integration property of the series of nuclear energy consumption and economic growth before performing the Granger-causality test. When both series are integrated of the same order, we can proceed to test for the presence of co-integration. The Johansen co-integration test procedure (Johansen and Juselius, 1990) is used for this purpose. Error-correction Model In the error-correction modeling procedure, X Granger-causes Y , if either the estimated coefficients on lagged values of X or the estimated coefficient on lagged value of error term from co-integrated regression is statistically significant. Similarly, Y Granger-causes X, if either the estimated coefficients on lagged values of Y or the estimated coefficient on lagged value of error term from co-integrated regression is statistically significant. This procedure specifically allows for a causal linkage between two or more variables stemming from an equilibrium relationship, thus characterizing the long-run equilibrium alignment that persists beyond the short-run adjustment. If two variables are non-stationary, but they become stationary after the first differencing, and co-integrated, the ECMs for the Granger causality test can be specified accordingly as follows: Yt D ˇ10 C

L11 X

ˇ11i Yt

i

C

i D1

Xt D ˇ20 C

L21 X i D1

L12 X

ˇ12j Xt

j

C ˇ13 "t

1

C u1t ;

(1)

j

C ˇ23 "t

1

C u2t ;

(2)

j D1

ˇ21i Xt

i

C

L22 X

ˇ22j Yt

j D1

where Xt and Yt represent natural logarithms of nuclear energy consumption and real gross domestic product (GDP), respectively, is the difference operator, Ls are the numbers of lags, ˇs are parameters to be estimated, ut s are the serially uncorrelated error terms, and "t 1 is the error correction term (ECT), which is derived from the long run co-integration relationship, Yt D 0 C 1 Xt C "t where s are parameters to be estimated and "t is error term. In each equation, the change in the dependent variable is caused not only by their lags, but also by the previous period’s disequilibrium in level, "t 1 . Given such a specification, the presence of short-run and long-run causality can be tested. Let us consider Eq. (1). If the estimated coefficients on lagged values of nuclear energy consumption (ˇ12 s) are statistically significant, then the implication is that nuclear energy consumption Grangercauses real GDP in the short-run. This test can be conducted by a joint F-test. On the other hand, long-run causality can be found by testing the significance of the estimated coefficient of ECT (ˇ13 ) by a t-test. Finally, the strong Granger-causality can be exposed through a joint test of the statistical significance of ˇ12 s and ˇ13 by a joint F-test. Similar reasoning is possible for examining whether real GDP Granger-causes nuclear energy consumption in Eq. (2).

114

J.-Y. Heo et al.

Empirical Results Data In order to investigate whether there is a causal relationship between nuclear energy consumption and economic growth, data covering the period 1969-2006 are used. The choice of the starting period was constrained by the availability of data on nuclear energy consumption. Nuclear energy consumption is expressed in terms of million tonnes oil equivalent. The GDP series of India are real GDP in constant 2000 prices in US dollars. The variables used in the models are: LNC, natural logarithm of nuclear energy consumption; and LGDP, natural logarithm of real GDP. The data on the two variables were obtained from BP (2007) and World Bank (2007), respectively. Results of Unit Roots and Co-integration Tests When testing for unit roots and co-integrations the authors have chosen to use a probability value of 0.10 in this study, which is an appropriate level of significance to be used with small sample sizes such as that used here. The results of the unit root tests for the series of LNC and LGDP variables are shown in Table 1. The PP test provides the formal test for unit roots in this study. The p-values of PP values calculated for the two series are larger than 0.10. This indicates that the series of all the variables are non-stationary at 10% level of significance, and thus any causal inferences from the two series in levels are invalid. However, non-stationarity can be rejected for first differences of these series at a 1% level of significance. Hence, the Granger-causality models are estimated with first-differenced data. As mentioned previously, the basic idea behind co-integration is to test whether a linear combination of two individually non-stationary time series is itself stationary. Given that integration of two series is of the same order, it is necessary to test whether the two series are co-integrated over the sample period. The results of the Johansen co-integration test for the series LNC and LGDP are reported in Table 2. The likelihood ratio tests show that the null hypothesis of absence of co-integrating relation .R D 0/ can be rejected at a 10% level of significance, but that the null hypothesis of existence of at most one co-integrating relation .R 1/ cannot be rejected at a 10% level of significance. This implies that there is only one co-integrating equation at a 10% level of significance. Evidence in this study indicates that the integrated

Table 1 Results of Phillips-Perron (PP) unit root tests Levels Variables LNC LGDP

PP-values 20.81[7] 2.77[2]

First differences p-values 0.060 0.946

PP-values 47.30[2]** 39.64[2]**

p-values 0.000 0.001

Note: The numbers inside the brackets are the optimum lag lengths determined using Akaike’s information criterion described in Pantula et al. (1994). The pvalues are calculated under the null hypothesis of nonstationarity. ** represents the rejection of the null hypothesis at a 1% level of significance.

Nuclear Energy and Economic Growth in India

115

Table 2 Results of Johansen co-integration tests Null hypothesis The number of co-integrating equation is 0 .R D 0/. The number of co-integrating equation is at most 1 .R 1/.

Likelihood ratio test statistic

p-values

17.422*

0.064

0.445

0.501

Note: The optimal lag length is chosen as 8 by using Akaike’s information criterion described in Pantula et al. (1994). The p-values are calculated under the corresponding null hypothesis. R denotes the number of co-integrating equation. * indicates the rejection of the null hypothesis at a 10% level of significance.

variables have inherent co-movement tendency over the long run. Thus, the authors conclude that nuclear energy consumption and real GDP are co-integrated. That is, there is a long-run relationship between nuclear energy consumption and real GDP for India. Results of the Error-correction Model and Granger-causality Tests In the ECM, the first difference of each endogenous variable (nuclear energy consumption and real GDP) was regressed on a period lag of the co-integrating equation and lagged first differences of all the endogenous variables in the system, as shown in Eqs. (1) and (2). The lag lengths, L11 , L12 , L21 and L22 , in Eqs. (1) and (2) were chosen by using Akaike’s information criterion described in Pantula et al. (1994). The results of the tests on causality are presented in Table 3. A significance level of 10% is also used for causality tests. Short-run causality is found to run from nuclear energy consumption to real GDP. However, the reverse short-run causality does not exit. That is, there is uni-directional short-run Granger-causality from nuclear energy

Table 3 Results of causality tests based on the error correction model Source of causation Short-run

Long-run

Joint (short-/long-run)

F-statistics

t-statistics

F-statistics

Null hypothesis

LNC

LGDP

Nuclear energy consumption does not cause economic growth. Economic growth does not cause nuclear energy consumption.

6.04** (0.004)

—

—

0.55 (0.736)

"t

1

3.60** (0.004) 0.37 (0.713)

LNC, "t

1

LGDP, "t

1

8.41** (0.001)

—

—

0.45 (0.832)

Note: The lag lengths are chosen by using Akaike’s information criterion described in Pantula et al. (1994). Finally, L11 D 6, L12 D 9, L21 D 2, and L22 D 5 were derived. More detailed estimation results are omitted here for brevity. However, they are available from the senior author upon request. The numbers in parentheses below the statistics are p-values calculated under the null hypothesis of no causation. ** denotes the rejection of the null hypothesis at 1% levels of significance.

116

J.-Y. Heo et al.

consumption to economic growth. The coefficient of the ECT is found to be significant in Eq. (1) and not in Eq. (2), which indicates that long-run Granger-causality from nuclear energy consumption to real GDP exists, but the reverse does not. Results of the significance of the estimated coefficients on lagged values of change in nuclear energy consumption (LNC), along with the ECT in Eq. (1) are consistent with the presence of strong Granger-causality running from nuclear energy consumption to economic growth. These indicate that whenever a shock occurs in the system, nuclear energy consumption would make short-run adjustments to restore long-run equilibrium. However, the statistical insignificance of the estimated coefficients on lagged values of change in real GDP (LGDP), along with the ECT in Eq. (2) means that there is no strong Granger-causality running from real GDP to nuclear energy consumption. Thus, according to the overall results, we can conclude that there is uni-directional causality from nuclear energy consumption to economic growth. These overall results are almost similar to those of Yoo and Jung (2005), but contrary to those of Ghosh (2006). Economic growth in India tends to cause energy consumption in accordance with results of Ghosh (2006). The Indian government, however, started nuclear energy researche to utilize the nuclear energy for military purposes, not to stimulate economic growth. Unlike the other energy resources, therefore, a reverse directional causality running from nuclear energy consumption to economic growth exists.

Conclusions The results of this study showed that there are uni-directional short-run and long-run causality from nuclear energy consumption to economic growth and uni-directional strong causality from nuclear energy consumption to economic growth. Overall, the authors found that causality runs from nuclear energy consumption to economic growth without any feedback effect. It has a number of implications for policy analysts and forecasters in India. A high level of nuclear energy consumption leads to high level of real GDP. The nuclear energy consumption is the initial receptor of an exogenous impact and the equilibrium is restored through adjustment in the real income variable. This implies that nuclear energy consumption infrastructure shortage may restrain the economic growth in India. Increasing real GDP requires enormous nuclear energy consumption, though there are many other factors contributing to economic growth, and nuclear energy is only one of such factors. In order not to adversely affect economic growth, efforts must be made to encourage government and industry to increase nuclear energy supply investment. The fast growth of India’s GDP is expected to persist for a few decades to come. Historical primary energy and electricity growth rates during 1981–2000 were 6% per year and 7.8% per year, respectively. Based on the growth rates proposed by Grover and Chandra (2006), per capita electricity generation would reach about 5,300 kWh per year in the year 2052 and total about 8,000 billion kWh. A policy for increasing nuclear energy supply investment is, therefore, likely to enhance economic growth for India. In pursuit of continuing economic growth, India will need to put more efforts into increasing nuclear energy supply investment when implementing national nuclear energy supply infrastructure as a strategy toward advanced development in the long haul. Thus, this study generates confidence in decisions to invest in the nuclear energy supply infrastructure.

Nuclear Energy and Economic Growth in India

117

References British Petroleum (BP). 2007. Statistical review of world energy 2007. Available at: http://www. bp.com Engle, R. F., and Granger, C. W. J. 1987. Cointegration and error correction: representation, estimation and testing. Econometrica 55:251–276. Geweke, J., Meese, R., and Dent, W. 1983. Comparing alternative tests for causality in temporal systems: analytic results and experimental evidence. J. Econometrics 21:161–194. Ghosh, S. 2006. Future demand of petroleum products in India. Energ. Policy 34:2032–2037. Granger, C. W. J. 1969. Investigating causal relation by econometric and cross-sectional method. Econometrica 37:424–438. Granger, C. W. J., and Newbold, P. 1974. Spurious regressions in econometrics. J. Econometrics 2:111–120. Grover, R. B., and Chandra, S. 2006. Scenario for growth of electricity in India. Energ. Policy 34:2834–2847. Guilkey, D. K., and Salemi, M. K. 1982. Small sample properties of the three tests of causality for Granger causal ordering in a bivariate stochastic system. Rev. Econ. Stat. 64:668–680. International Energy Agency. 2004. World energy outlook 2004. Paris: International Energy Agency/ OECD. Johansen, S., and Juselius, K. 1990. Maximum likelihood estimation and inference on cointegration with applications to the demand for money. Oxford B. Econ. Stats. 52:169–210. Nuclear Power Corporation of India Limited. 2007. Annual report 2006–2007. Mumbai, India: Nuclear Power Corporation of India Ltd. Pantula, S. G., Gonzalez-Farias, G., and Fuller, W. A. 1994. A comparison of unit-root test criteria. J. Busd. Econ. Stat. 12:449–459. Phillips, P. C. B., and Perron, P. 1988. Testing for a unit root in time series regression. Biometrika 75:335–346. Toda, H. Y., and Phillips, P. C. B. 1993. Vector autoregressions and causality. Econometrica 61:1367–1393. Toth, F. L., and Rogner, H. H. 2006. Oil and nuclear power: Past, present, and future. Energ. Eco. 28:1–25. World Bank. 2007. World development indicators. Washington, DC: World Bank. Yoo, S.-H., and Jung, K.-O. 2005. Nuclear energy consumption and economic growth in Korea. Prog. Nucl. Energ. 46:101–109. Yoo, S.-H., and Kwak, S.-J. 2004. Information technology and economic development in Korea: A causality study. Int. J. Tech. Manage. 27:57–67.

The Causal Relationship Between Nuclear Energy Consumption and Economic Growth in India J.-Y. HEO,1 S.-H. YOO,2 and S.-J. KWAK3 1

Department of Energy Policy, Graduate School of Energy & Environment, Seoul National University of Science & Technology, Seoul, Korea 2 Department of International Area Studies, Hoseo University, Chungnam, Korea 3 Department of Economics, Korea University, Seoul, Korea Abstract This article attempts to investigate the long-run and short-run causality issues between nuclear energy consumption and economic growth in India by using the co-integration and error-correction models. It employs the annual data covering the period 1969–2006. The results indicate that there is a uni-directional causality running from nuclear energy consumption to economic growth without any feedback effect. Thus, considering the fact that nuclear energy consumption fosters economic growth, policies for increasing nuclear energy supply investment is, therefore, likely to enhance economic growth in India. Keywords causality, economic growth, India, nuclear energy consumption

Introduction Historical records of the early 1970s show that almost 25% of global electricity was generated from oil while nuclear’s share constituted only 3%. By 2002, however, the global electricity supply structure changed. Oil’s share declined to 7.2% while nuclear constituted 16.6%. Nuclear absorbed approximately 75% of the decrease of oil’s share (17.8%). There is no doubt that nuclear energy has made a major intrusion in the electricity market (International Energy Agency, 2004). Since the 1990s the emphasis of nuclear power plant construction has shifted towards Asia and developing countries. New nuclear power plants are most attractive where energy demand growth is rapid, alternative resources are scarce, and energy supply security is a priority—China, India, Japan, and Korea. Moreover, using nuclear energy as a source of energy reduces air pollution and greenhouse gas emissions (Toth and Rogner, 2006). India is on an economic overdrive in the post liberalization era and has developed a great craving for energy. Although coal is the dominant energy source providing 72% of the electricity in India, it is an exhaustible source of energy likely to last for only another 5 or 6 decades. Amongst all the options available to India today, nuclear energy (2.9% of the fuel mix for generating power) emerges as a ray of hope (Nuclear Power Corporation of India Limited, 2007). Public policy makers in India have shown a great deal of interest in the role that nuclear energy consumption plays in economic growth. The nuclear energy infrastructure Address correspondence to Seung-Hoon Yoo, Department of Energy Policy, Graduate School of Energy & Environment, Seoul National University of Science & Technology, 172 Gongreung2-Dong, Nowon-Ku, Seoul, 139–743, Republic of Korea. E-mail: [email protected]

111

112

J.-Y. Heo et al.

of India is becoming an increasingly important component of the economy. Grover and Chandra (2006) present a scenario for the growth of electricity in India. To meet the projected demand, the article presents that the nuclear contribution towards electricity generation has to increase from the present 3% to about a quarter of the total. To proactively cope with increasing nuclear energy demand accompanying rapid economic growth, India should endeavor to uncover the causal relationship between nuclear energy consumption and economic growthto make appropriate nuclear energy policy. This task has become one of the most important agendas for India in the present and will be so in the near future. The purpose of this article is, therefore, to investigate causality between nuclear energy consumption and economic growth, and to obtain policy implications from the results. Ghosh (2006) examined the long-run equilibrium relationship between total petroleum products consumption and economic growth in India. However, only a few studies have been done on the relationship between nuclear energy consumption and economic growth. Yoo and Jung (2005) detected a uni-directional causality which runs from nuclear energy consumption to economic growth in Korea. To this end, the authors attempt to provide more careful consideration of the causality issues by applying the modern rigorous techniques of Granger causality to the Indian data. The remainder of the article is organized as follows. The next section presents an overview of the proposed methodology. The penultimate section explains the data employed and reports the empirical findings. Some concluding remarks are made in the final section.

Methodology Granger-causality and Stationarity The first attempt at testing for the direction of causality was proposed by Granger (1969). The Granger-causality test is a convenient and very general approach for detecting any presence of a causal relationship between two variables. The test is quite simple and straightforward. A time series .X/ is said to Granger-cause another time series .Y / if the prediction error of current Y declines by using past values of X in addition to past values of Y . The Granger-causality test method is selected to be used in this study over other alternative techniques because of the favorable Monte Carlo evidence reported by Guilkey and Salemi (1982) and Geweke et al. (1983), particularly for small samples in empirical works. In order to conduct to the Granger-causality test, a series of variables is required to be stationary. It has been shown that using non-stationary data in causality tests can yield spurious causality results (Granger and Newbold, 1974). Therefore, following Engle and Granger (1987), the authors first test the unit roots of X and Y to confirm the stationarity of each variable. This is done by using the Phillips-Perron (PP) test (Phillips and Perron, 1988) over alternative tests, in that the PP test is known to be robust for a variety of serial correlations and time-dependent heteroscedasticities. If any variable is found to be non-stationary, we must take the first difference and then apply the causality test with differenced data. Co-integration The concept of co-integration can be defined as a systematic co-movement among two or more economic variables over the long run. If X and Y each are non-stationary and

Nuclear Energy and Economic Growth in India

113

co-integrated, then any standard Granger-causal inferences will be invalid and a more comprehensive test of causality based on an error-correction model (ECM), should be adopted (Engle and Granger, 1987). However, if X and Y are both non-stationary and the linear combination of the series of two variables is non-stationary, then a standard Granger-causality test should be adopted (Toda and Phillips, 1993; Yoo and Kwak, 2004). Therefore, it is necessary to test for the co-integration property of the series of nuclear energy consumption and economic growth before performing the Granger-causality test. When both series are integrated of the same order, we can proceed to test for the presence of co-integration. The Johansen co-integration test procedure (Johansen and Juselius, 1990) is used for this purpose. Error-correction Model In the error-correction modeling procedure, X Granger-causes Y , if either the estimated coefficients on lagged values of X or the estimated coefficient on lagged value of error term from co-integrated regression is statistically significant. Similarly, Y Granger-causes X, if either the estimated coefficients on lagged values of Y or the estimated coefficient on lagged value of error term from co-integrated regression is statistically significant. This procedure specifically allows for a causal linkage between two or more variables stemming from an equilibrium relationship, thus characterizing the long-run equilibrium alignment that persists beyond the short-run adjustment. If two variables are non-stationary, but they become stationary after the first differencing, and co-integrated, the ECMs for the Granger causality test can be specified accordingly as follows: Yt D ˇ10 C

L11 X

ˇ11i Yt

i

C

i D1

Xt D ˇ20 C

L21 X i D1

L12 X

ˇ12j Xt

j

C ˇ13 "t

1

C u1t ;

(1)

j

C ˇ23 "t

1

C u2t ;

(2)

j D1

ˇ21i Xt

i

C

L22 X

ˇ22j Yt

j D1

where Xt and Yt represent natural logarithms of nuclear energy consumption and real gross domestic product (GDP), respectively, is the difference operator, Ls are the numbers of lags, ˇs are parameters to be estimated, ut s are the serially uncorrelated error terms, and "t 1 is the error correction term (ECT), which is derived from the long run co-integration relationship, Yt D 0 C 1 Xt C "t where s are parameters to be estimated and "t is error term. In each equation, the change in the dependent variable is caused not only by their lags, but also by the previous period’s disequilibrium in level, "t 1 . Given such a specification, the presence of short-run and long-run causality can be tested. Let us consider Eq. (1). If the estimated coefficients on lagged values of nuclear energy consumption (ˇ12 s) are statistically significant, then the implication is that nuclear energy consumption Grangercauses real GDP in the short-run. This test can be conducted by a joint F-test. On the other hand, long-run causality can be found by testing the significance of the estimated coefficient of ECT (ˇ13 ) by a t-test. Finally, the strong Granger-causality can be exposed through a joint test of the statistical significance of ˇ12 s and ˇ13 by a joint F-test. Similar reasoning is possible for examining whether real GDP Granger-causes nuclear energy consumption in Eq. (2).

114

J.-Y. Heo et al.

Empirical Results Data In order to investigate whether there is a causal relationship between nuclear energy consumption and economic growth, data covering the period 1969-2006 are used. The choice of the starting period was constrained by the availability of data on nuclear energy consumption. Nuclear energy consumption is expressed in terms of million tonnes oil equivalent. The GDP series of India are real GDP in constant 2000 prices in US dollars. The variables used in the models are: LNC, natural logarithm of nuclear energy consumption; and LGDP, natural logarithm of real GDP. The data on the two variables were obtained from BP (2007) and World Bank (2007), respectively. Results of Unit Roots and Co-integration Tests When testing for unit roots and co-integrations the authors have chosen to use a probability value of 0.10 in this study, which is an appropriate level of significance to be used with small sample sizes such as that used here. The results of the unit root tests for the series of LNC and LGDP variables are shown in Table 1. The PP test provides the formal test for unit roots in this study. The p-values of PP values calculated for the two series are larger than 0.10. This indicates that the series of all the variables are non-stationary at 10% level of significance, and thus any causal inferences from the two series in levels are invalid. However, non-stationarity can be rejected for first differences of these series at a 1% level of significance. Hence, the Granger-causality models are estimated with first-differenced data. As mentioned previously, the basic idea behind co-integration is to test whether a linear combination of two individually non-stationary time series is itself stationary. Given that integration of two series is of the same order, it is necessary to test whether the two series are co-integrated over the sample period. The results of the Johansen co-integration test for the series LNC and LGDP are reported in Table 2. The likelihood ratio tests show that the null hypothesis of absence of co-integrating relation .R D 0/ can be rejected at a 10% level of significance, but that the null hypothesis of existence of at most one co-integrating relation .R 1/ cannot be rejected at a 10% level of significance. This implies that there is only one co-integrating equation at a 10% level of significance. Evidence in this study indicates that the integrated

Table 1 Results of Phillips-Perron (PP) unit root tests Levels Variables LNC LGDP

PP-values 20.81[7] 2.77[2]

First differences p-values 0.060 0.946

PP-values 47.30[2]** 39.64[2]**

p-values 0.000 0.001

Note: The numbers inside the brackets are the optimum lag lengths determined using Akaike’s information criterion described in Pantula et al. (1994). The pvalues are calculated under the null hypothesis of nonstationarity. ** represents the rejection of the null hypothesis at a 1% level of significance.

Nuclear Energy and Economic Growth in India

115

Table 2 Results of Johansen co-integration tests Null hypothesis The number of co-integrating equation is 0 .R D 0/. The number of co-integrating equation is at most 1 .R 1/.

Likelihood ratio test statistic

p-values

17.422*

0.064

0.445

0.501

Note: The optimal lag length is chosen as 8 by using Akaike’s information criterion described in Pantula et al. (1994). The p-values are calculated under the corresponding null hypothesis. R denotes the number of co-integrating equation. * indicates the rejection of the null hypothesis at a 10% level of significance.

variables have inherent co-movement tendency over the long run. Thus, the authors conclude that nuclear energy consumption and real GDP are co-integrated. That is, there is a long-run relationship between nuclear energy consumption and real GDP for India. Results of the Error-correction Model and Granger-causality Tests In the ECM, the first difference of each endogenous variable (nuclear energy consumption and real GDP) was regressed on a period lag of the co-integrating equation and lagged first differences of all the endogenous variables in the system, as shown in Eqs. (1) and (2). The lag lengths, L11 , L12 , L21 and L22 , in Eqs. (1) and (2) were chosen by using Akaike’s information criterion described in Pantula et al. (1994). The results of the tests on causality are presented in Table 3. A significance level of 10% is also used for causality tests. Short-run causality is found to run from nuclear energy consumption to real GDP. However, the reverse short-run causality does not exit. That is, there is uni-directional short-run Granger-causality from nuclear energy

Table 3 Results of causality tests based on the error correction model Source of causation Short-run

Long-run

Joint (short-/long-run)

F-statistics

t-statistics

F-statistics

Null hypothesis

LNC

LGDP

Nuclear energy consumption does not cause economic growth. Economic growth does not cause nuclear energy consumption.

6.04** (0.004)

—

—

0.55 (0.736)

"t

1

3.60** (0.004) 0.37 (0.713)

LNC, "t

1

LGDP, "t

1

8.41** (0.001)

—

—

0.45 (0.832)

Note: The lag lengths are chosen by using Akaike’s information criterion described in Pantula et al. (1994). Finally, L11 D 6, L12 D 9, L21 D 2, and L22 D 5 were derived. More detailed estimation results are omitted here for brevity. However, they are available from the senior author upon request. The numbers in parentheses below the statistics are p-values calculated under the null hypothesis of no causation. ** denotes the rejection of the null hypothesis at 1% levels of significance.

116

J.-Y. Heo et al.

consumption to economic growth. The coefficient of the ECT is found to be significant in Eq. (1) and not in Eq. (2), which indicates that long-run Granger-causality from nuclear energy consumption to real GDP exists, but the reverse does not. Results of the significance of the estimated coefficients on lagged values of change in nuclear energy consumption (LNC), along with the ECT in Eq. (1) are consistent with the presence of strong Granger-causality running from nuclear energy consumption to economic growth. These indicate that whenever a shock occurs in the system, nuclear energy consumption would make short-run adjustments to restore long-run equilibrium. However, the statistical insignificance of the estimated coefficients on lagged values of change in real GDP (LGDP), along with the ECT in Eq. (2) means that there is no strong Granger-causality running from real GDP to nuclear energy consumption. Thus, according to the overall results, we can conclude that there is uni-directional causality from nuclear energy consumption to economic growth. These overall results are almost similar to those of Yoo and Jung (2005), but contrary to those of Ghosh (2006). Economic growth in India tends to cause energy consumption in accordance with results of Ghosh (2006). The Indian government, however, started nuclear energy researche to utilize the nuclear energy for military purposes, not to stimulate economic growth. Unlike the other energy resources, therefore, a reverse directional causality running from nuclear energy consumption to economic growth exists.

Conclusions The results of this study showed that there are uni-directional short-run and long-run causality from nuclear energy consumption to economic growth and uni-directional strong causality from nuclear energy consumption to economic growth. Overall, the authors found that causality runs from nuclear energy consumption to economic growth without any feedback effect. It has a number of implications for policy analysts and forecasters in India. A high level of nuclear energy consumption leads to high level of real GDP. The nuclear energy consumption is the initial receptor of an exogenous impact and the equilibrium is restored through adjustment in the real income variable. This implies that nuclear energy consumption infrastructure shortage may restrain the economic growth in India. Increasing real GDP requires enormous nuclear energy consumption, though there are many other factors contributing to economic growth, and nuclear energy is only one of such factors. In order not to adversely affect economic growth, efforts must be made to encourage government and industry to increase nuclear energy supply investment. The fast growth of India’s GDP is expected to persist for a few decades to come. Historical primary energy and electricity growth rates during 1981–2000 were 6% per year and 7.8% per year, respectively. Based on the growth rates proposed by Grover and Chandra (2006), per capita electricity generation would reach about 5,300 kWh per year in the year 2052 and total about 8,000 billion kWh. A policy for increasing nuclear energy supply investment is, therefore, likely to enhance economic growth for India. In pursuit of continuing economic growth, India will need to put more efforts into increasing nuclear energy supply investment when implementing national nuclear energy supply infrastructure as a strategy toward advanced development in the long haul. Thus, this study generates confidence in decisions to invest in the nuclear energy supply infrastructure.

Nuclear Energy and Economic Growth in India

117

References British Petroleum (BP). 2007. Statistical review of world energy 2007. Available at: http://www. bp.com Engle, R. F., and Granger, C. W. J. 1987. Cointegration and error correction: representation, estimation and testing. Econometrica 55:251–276. Geweke, J., Meese, R., and Dent, W. 1983. Comparing alternative tests for causality in temporal systems: analytic results and experimental evidence. J. Econometrics 21:161–194. Ghosh, S. 2006. Future demand of petroleum products in India. Energ. Policy 34:2032–2037. Granger, C. W. J. 1969. Investigating causal relation by econometric and cross-sectional method. Econometrica 37:424–438. Granger, C. W. J., and Newbold, P. 1974. Spurious regressions in econometrics. J. Econometrics 2:111–120. Grover, R. B., and Chandra, S. 2006. Scenario for growth of electricity in India. Energ. Policy 34:2834–2847. Guilkey, D. K., and Salemi, M. K. 1982. Small sample properties of the three tests of causality for Granger causal ordering in a bivariate stochastic system. Rev. Econ. Stat. 64:668–680. International Energy Agency. 2004. World energy outlook 2004. Paris: International Energy Agency/ OECD. Johansen, S., and Juselius, K. 1990. Maximum likelihood estimation and inference on cointegration with applications to the demand for money. Oxford B. Econ. Stats. 52:169–210. Nuclear Power Corporation of India Limited. 2007. Annual report 2006–2007. Mumbai, India: Nuclear Power Corporation of India Ltd. Pantula, S. G., Gonzalez-Farias, G., and Fuller, W. A. 1994. A comparison of unit-root test criteria. J. Busd. Econ. Stat. 12:449–459. Phillips, P. C. B., and Perron, P. 1988. Testing for a unit root in time series regression. Biometrika 75:335–346. Toda, H. Y., and Phillips, P. C. B. 1993. Vector autoregressions and causality. Econometrica 61:1367–1393. Toth, F. L., and Rogner, H. H. 2006. Oil and nuclear power: Past, present, and future. Energ. Eco. 28:1–25. World Bank. 2007. World development indicators. Washington, DC: World Bank. Yoo, S.-H., and Jung, K.-O. 2005. Nuclear energy consumption and economic growth in Korea. Prog. Nucl. Energ. 46:101–109. Yoo, S.-H., and Kwak, S.-J. 2004. Information technology and economic development in Korea: A causality study. Int. J. Tech. Manage. 27:57–67.