Energy consumption and Economic Growth Nexus in the Baltic Countries: Causality Approach

Transcription

1 Volume VII Number 4 December 2014 Energ consumtion and Economic Growth Nexus in the Baltic Countries: Causalit Aroach Gitana Dudzevičiūtė 1, Rima Tamošiūnienė 2 Abstract. The relationshi between energ consumtion and economic growth has been a hot issue in academic research. The results of the investigations have been varied due to quite different national economic develoment and diversit of research methods as well. This aer investigates the causal relationshi between energ consumtion and economic growth in the Baltic countries, such as Lithuania, Latvia and Estonia in the eriod of ear between 1995 and To that end, Granger causalit test is alied in order to suggest which variable in the model has significant imact on the future value of other variable in the sstem. Based on the causal relationshi, regression model has been emloed for the estimation of interrelationshi between energ consumtion and economic growth. The emirical results could be summarized as following: 1) in terms of correlation analsis between energ consumtion and economic growth, there is a weak and insignificant relationshi for Lithuania, strong and significant for Latvia and moderate one for Estonia; 2) in terms of causalit running from GDP to energ consumtion, the causal relationshi exists for Estonia. The main imlication of this finding is that the energ conservation olicies will not have imact on economic growth; moreover, the results of the research highlight the absence of Granger causalit between variables observed in the case of Lithuania and Latvia. In this light, national energ olicies will not have imact on economic growth as well as economic growth will not has influence on energ consumtion. Index Terms: energ consumtion, economic growth, Granger causalit, regression analsis. JEL: C10, C12, O47, Q43 I. INTRODUCTION The growth of energ consumtion in the Euroean Union countries oses serious develoment constraints in recent ears. According to rognosis the global energ consumtion will continue to increase in the next twent ears. It is estimated that global oil consumtion will increase one ercent and global natural gas consumtion will increase 1.5 ercent on average er ear, so that total increase in energ demand will amount to 45 ercent b 2030 (Ozkan 2011; Dudzevičiūtė 2012). 1 Gitana Dudzevičiūtė is with the Facult of Business Management, Vilnius Gediminas Technical Universit, Sauletekio ave. 11, LT Vilnius, and with the Deartment of Management, The General Jonas Žemaitis Militar Academ of Lithuania, Šilo str. 5A, LT Vilnius Lithuania. 2 Rima Tamošiūnienė is with the Facult of Economics and Finance Management, Mkolas Romeris Universit, Ateities str. 20, LT Vilnius, Lithuania. In academic research, the relationshi between energ consumtion and economic growth is analzed from different theoretical asects. Scientists discuss the causalit between these two variables the question whether this relationshi exists and what is its nature has been raised. Some of them (Ho et al. 2007; Akinlo 2009; Ozturk 2010; Munim et al. 2010; Tsani 2010; Lee et al. 2011; Aergis 2012; Yildirim et al. 2014) assume unidirectional causalit from energ consumtion to economic growth and emhasize ver significant role of energ consumtion on economic growth. This relationshi denotes an energ deendent econom and in the case of limited access to energ sul can result in oor economic develoment. Under this scenario, economic develoment rogrammes should aim at imroving access to energ for all oulations and sectors (Squalli 2007). While others signif that economic growth is the central factor which causes energ consumtion and indicates an econom which is less energ deendent. In this case energ management olicies, such as investments in energ efficienc will have no adverse imact on GDP growth (Ouedraogo 2013). Also in recent academic research findings are resented that iml a mutual and comlementar relationshi between energ consumtion and economic growth (Bartleet and Gounder 2010; Esso 2010; Eggoh et al. 2011; Belke et al. 2011; Fuinhas and Marques 2012) or indicate that the energ consumtion has no imact on economic growth (Huang et al. 2008; Pane 2009; Acaravci and Ozturk 2010; Yalta and Cakar 2012). In the case of the absence of causalit between energ consumtion and economic growth, olicies to romote higher levels of energ consumtion will not have an imact on economic growth (Ouedraogo 2013). This research attemts to rovide more reliable estimates of the causal relationshi between energ consumtion and economic growth in the Baltic countries and uses annual data of The aer is organized as follows. Section 2 gives a short summar of the relevant emirical literature review on the relationshi of energ consumtion and economic growth and research methodolog. The investigations are summarized and the main insights are rovided. Section 3 analses the main trends and causal relationshi between energ consumtion and economic growth in three countries, such as Lithuania, Latvia and Estonia. Granger causalit test and regression model are alied for the 12

2 KSI Transactions on KNOWLEDGE SOCIETY estimations. Section 4 concludes summarizing the main trends observed. II. EMPIRICAL RESEARCH REVIEW AND METHODOLOGY A. Emirical research overview As discussed above, in the context of economic growth, energ resources la a crucial role. Over the ast decade, the causal relationshi between energ consumtion and economic growth has been analzed extensivel b energ secialists. Recentl ublished researches on the energ growth nexus were done in two was (Dudzevičiūtė 2013) some researchers analze the causal relationshi between searate energ resources (electricit, gas, coal, renewable resources and others) and economic growth (Bobinaitė et al. 2011; Li 2011; Georgantooulas 2012; Tang and Tan 2012), others (Chontanawat et al. 2008; Odularu and Okonkwo 2009; Bojnec and Paler 2011; Zhixin and Xin 2011; Chen et al. 2012; Camo and Sarmiento 2013; Yildirim et al. 2014) examine energ growth nexus focusing on total energ consumtion. In general, the findings of the investigations show a strong relationshi between energ and economic growth. However, the two focus variables have been a oint of disagreement in the academic research. An imortant research of the imact of energ consumtion on economic growth was done b Chontanawat el al. (2008). The stud examined the relationshi for over 100 countries, divided into two categories such as develoed OECD countries and develoing non-oecd countries. The emirical estimates showed that there is the causalit from energ to economic growth and it is stronger in the develoed countries than in develoing ones. Another stud of energ growth nexus was conducted b Yildirim et al. (2014) in the next 11 countries. The findings imlied that energ conservation-oriented olicies should be imlemented in Bangladesh, Egt, Iran, Korea, Mexico, Indonesia, Pakistan and Philiines. In the case of Turke, a unidirectional causal relationshi was found between energ consumtion and economic growth. It indicates that energ conservation olic oses an obstacle for economic growth in Turke. Camo and Sarmiento (2013) examined the long run relationshi between energ consumtion and economic growth for 10 Latin American countries. The results showed interconnection between two variables. According to the research results, a 1% increase in total energ consumtion increases economic growth b 0.59%. At the same time, a 1% growth in GDP causes an increase in energ consumtion b 0.59%. This finding suorts the olicies that romote energ conservation and efficienc (Camo and Sarmiento 2013). The stud of Lin and Wesseh (2014) revealed unidirectional causalit running from energ consumtion to economic growth in South Africa. The findings of research iml that energ conservation olicies will negativel imact on economic growth in South Africa. Given the wide range of emirical evidences in the scientific literature on the causal relationshi between energ consumtion and economic growth, no general conclusion can be made regarding the differences in research results between develoed and develoing countries. A similar icture aears in the scientific researches of the searate resources of energ (electricit, renewable and non-renewable resources and others) and growth nexus. More secificall, electricit consumtion has become the dominant form and it is considered to be ver imortant for economic growth. Man studies have been conducted on this toic (Chen et al. 2007; Squalli 2007; Ho et al. 2007; Gurgul and Lach 2011; Georgantooulas 2012). The findings of these studies iml different forms of relationshi between energ resources and economic growth such as unidirectional causalit, the absence of causalit, a mutual and comlementar relationshi. The direction of causalit between energ consumtion and economic growth has significant olic imlication. On the one hand, in the case of unidirectional causalit running from energ consumtion to economic growth, conserving or reducing energ could reduce economic growth. On the other hand, in the case of unidirectional causalit from economic growth to energ consumtion, conservation olicies ma be imlemented with little or no effects on economic growth. When causalit does not exist, it means that neither conservative nor exansive olicies regarding energ consumtion have an imact on economic growth. All researchers have concluded that there are interdeendenc of energ consumtion and economic growth, but the ractices of different countries lead to different results regarding the resence of causalit. B. Research methodolog This research was guided b the estimation of energ consumtion growth nexus in the Baltic countries. The author refer to methodolog on energ growth relationshi considered in studies of different countries (Chontanawat et al. 2008; Odularu and Okonkwo 2009; Bojnec and Paler 2011; Zhixin and Xin 2011; Eddrief- Cherfi and Kourbali 2012; Chen et al. 2012; Yildirim et al. 2014). The ratios of energ consumtion er caita and GDP er caita (Eddrief-Cherfi and Kourbali 2012; Yildirim et al. 2014) are used in order to comare three Baltic countries Lithuania, Latvia and Estonia as well. Granger (1969) causalit test is alied for the estimation of causalit and the direction of relationshi between energ consumtion and economic growth. Granger causalit test requires estimating the following two regression equations (Kozhan 2010; Stern 2011): t = β 1,0 + β 1, i t i + β 1, + j x t j + ε 1t i = 1 j = 1 (1) x t = β 2, 0 + β 2, i t i + β 2, + j x t j + ε 1t i = 1 j = 1 (2) 13

3 Volume VII Number 4 December 2014 where: is the number of lags, β- arameter, ε- error. If the arameters β 1,+j are jointl significant then the null hothesis that x does not Granger cause can be rejected. Similarl, if the arameters β 2,i are jointl significant then the null hothesis that does not Granger cause x can be rejected. Granger causalit test is based on the concet of causal ordering and assumtion as follows: a variable x is said to Granger cause another variable if ast values of x hel redict the current level of given all other aroriate information (Stern 2011). Finall, a regression model would be more aroriate for more detailed analsis of energ consumtion economic growth nexus. When the relationshi of the variables has been established as linear, a regression equation can be alied as follows (Kozhan 2010): i = a + bx i (3) where: i - deendent variable (energ consumtion or economic growth taking into consideration the results of Granger causalit), x i indeendent variable (energ consumtion or economic growth taking into consideration the results of Granger causalit), a, b- arameters. The arameter b indicates how much ma change on average, if x is changed b unit. All calculations are made using Windows-based econometric software Eviews v Section 3 analses the relationshi between energ consumtion and economic growth. III. INVESTIGATION OF THE INTERRELATIONSHIP BETWEEN ENERGY CONSUMPTION AND ECONOMIC GROWTH A. Energ consumtion and economic growth trends in the Baltic countries Two aroaches are emloed in this research. First, annual data analsis of the three Baltic countries Lithuania, Latvia and Estonia is carried out in the eriod of The aim of this exercise is to comare three Baltic countries according to energ consumtion and economic growth ratios and analze the relationshi between these two variables. Second, the causal relationshi and its direction is examined using Granger causalit test; and regression model is made to describe reliance of one factor from another in quantitative wa. Figure 1 and figure 2 reresent the tendencies of energ consumtion er caita and GDP er caita in the Baltic countries. The analsis of these indicators has rovided emirical insights into the long-run. Indeed, the general findings for all three countries have been as follows: 1) slight variation of energ consumtion er caita, which has made over 9 ercent; and 2) ver big variation of GDP er caita, which has made over 50 ercent. Over a eriod of , ever Estonian has used energ twice as much than Lithuanian and Latvian. In 2012, the ratio of energ consumtion er caita was amounted to 4.6 TOE in Estonia and 2.4 and 2.2 TOE in Lithuania and Latvia resectivel. The Estonian energ consumtion er caita increased from 3.9 TOE in 1995 to 4.6 TOE in In Latvia, the growth of the same ratio was from 1.9 TOE in 1995 to 2.2 TOE in In , the Lithuanian energ consumtion er caita remained at the same level and made 2.4 TOE in ,0 4,5 4,0 3,5 3,0 2,5 2,0 1,5 1,0 0,5 0, Estonia Latvia Lithuania Source: Eurostat data Figure 1. Energ consumtion er caita, TOE Over a eriod of , three Baltic countries have recorded the growth of GDP er caita. The greatest increase has been recorded in Estonia. It has increased from EUR 2 thou in 1995 to EUR 13 thou in In Lithuania, the growth was from EUR 1.4 thou to EUR 11 thou, and in Latvia from EUR 1.5 thou to EUR 10.9 thou. Estonia has declared relativel higher ratio of GDP er caita than other Baltic countries in Average value of this indicator has made EUR 7.4 thou in Estonia and EUR 5.8 thou in Lithuania and Latvia Estonia Latvia Lithuania Source: Eurostat data Figure 2. GDP er caita, Euro In order to evaluate the relationshi and its direction between energ consumtion and economic growth correlation analsis has been erformed. The results have been summarized in Table 1. The stud of correlation analsis has shown different results of Baltic countries regarding energ consumtion economic growth nexus. It is noticeable that Lithuania has revealed weak and ositive relationshi between these factors. The t-test alied for this relationshi has suggested statistical insignificance. The case of Latvia and Estonia has revealed strong and moderate relationshis of 14

4 KSI Transactions on KNOWLEDGE SOCIETY the same direction between energ consumtion er caita and GDP er caita. The t-test for these countries has shown that these relationshis are reliable; and energ consumtion and economic growth have been directl related i.e. when one ratio increases the other grows as well and vice versa. TABLE 1. RESULTS OF CORRELATION ANALYSIS BETWEEN ENERGY CONSUMPTION AND GDP RATIOS Countr Relationshi t- t- Statistical strength critical observed significance Lithuania 0.27 Weak Insignificant Latvia 0.87 Strong Significant Estonia 0.69 Moderate Significant Correlations between the trending variables sa nothing about causation and siml reflect that both factors have weak, strong or moderate trends relative to the fluctuations around the trend. Correlation does not show causation. In this light, Granger causalit testing to modeling the relationshi between energ consumtion and economic growth has been alied to this research. It has revealed if energ consumtion has caused economic growth or economic growth has driven increasing energ consumtion; or the absence of causalit. The answers to these questions are imortant for analzing and imlementing energ olicies. Next section is devoted for this issue. IV. GRANGER CAUSALITY TESTING Granger causalit test is to stud the forerunner-lag relationshi between the two variables such as energ consumtion er caita and GDP er caita have been investigated in this section of the aer. A variable energ consumtion is said to Granger cause another variable economic growth if ast values of energ consumtion hel redict the current level of economic growth given all other aroriate information. This is based on the concet of causal ordering. Similarl, if economic growth in fact causes energ consumtion, then given the ast histor of economic growth it is likel that information will hel redict energ consumtion. Table 2 resents the results of Granger causalit test. TABLE II. GRANGER CAUSALITY TEST RESULTS Countr/hothesis Lags Observations F-statistic Probabilit Test results Lithuania Null hothesis: GDP does not Granger cause of EC Acceted Acceted GDP does not Granger cause of EC Rejected Acceted GDP does not Granger cause of EC Acceted Acceted GDP does not Granger cause of EC Acceted Acceted Latvia Null hothesis: GDP does not Granger cause of EC Acceted Acceted GDP does not Granger cause of EC Acceted Acceted GDP does not Granger cause of EC Acceted Acceted GDP does not Granger cause of EC Acceted Acceted Estonia Null hothesis: GDP does not Granger cause of EC Rejected Acceted GDP does not Granger cause of EC Rejected Acceted GDP does not Granger cause of EC Rejected Acceted GDP does not Granger cause of EC Rejected Acceted 15

5 Volume VII Number 4 December 2014 The results of Granger causalit test rovide new emirical insights into the long run relationshi between energ consumtion er caita and GDP er caita. The rejection rule is alied, when the robabilit value is no more than the level of significance at 5% and 10%. The research suggests different results of the Baltic countries. In general, the case of Lithuania reveals that no causalit exists between variables, excet two lags situation, when GDP er caita leads the energ consumtion er caita at 7% level of significance. In the case of Latvia, it is shown that the null hothesis that GDP does not Granger-cause of energ consumtion and energ consumtion does not Granger-cause of GDP cannot be rejected. In this light, we can conclude that there is absence of Granger causalit between these two variables, no matter that the strongl correlate. According to Ouedraogo (2013), under this scenario, olicies to romote energ access and higher levels of consumtion will not have imact on economic growth. The case of Estonia shows the running causalit from GDP er caita growth to energ consumtion er caita. This unidirectional causalit signifies that econom is less energ deendent and economic growth causes the develoment of energ sector. Under this scenario, energ conservation olicies, such as investments in energ efficienc and demand management olicies will have no adverse imact on GDP growth (Ouedraogo 2013). Aling regression model, this causal relationshi could be described as follows: = (4) i x i where: i - deendent variable energ consumtion er caita, x i indeendent variable GDP er caita. According to the regression model, an increase b EUR 1thou in GDP er caita increases total energ consumtion er caita b 0.07 TOE. The regression model describes arox. 48% of the variation in energ consumtion er caita. IV. CONCLUSION The aer investigates the causal relationshi between energ consumtion and economic growth in the Baltic countries in a eriod of Granger causalit test is alied in order to suggest which variable in the model has significant imact on the future value of other variable in the sstem. Based on the causal relationshi, regression model has been emloed for the estimation of interrelationshi between variables observed. The emirical results highlight a weak and insignificant correlation between energ consumtion er caita and GDP er caita for Lithuania, strong and significant relationshi for Latvia and moderate one for Estonia. In terms of causalit, there is causal relationshi running from GDP to energ consumtion in Estonia. This unidirectional causalit signifies that econom is less energ deendent and economic growth causes the develoment of energ sector. The assumtion could be made that the energ conservation olicies will not have imact on economic growth. Economic growth, however, imacts on energ consumtion. According to the regression model based on Granger causalit test s results and alied for the case of Estonia, an increase b EUR 1thou in GDP er caita increases total energ consumtion er caita b 0.07 TOE. This regression model describes arox. 48% of the variation in energ consumtion er caita. What is more, the research highlights the absence of Granger causalit between variables observed in the case of Lithuania and Latvia. In this light, national energ olicies will not have imact on economic growth as well as economic growth will not has influence on energ consumtion. REFERENCES Akinlo, A. E. (2009). Electricit consumtion and economic growth in Nigeria: Evidence from co-integration and cofeature analsis, Journal of Polic Modeling 31: Aergis, N.; Pane, J. E. (2012). Renewable and non-renewable energ consumtion-growth nexus: Evidence from a anel error correction model, Energ Economics, Elsevier 34(3): Bartleet, M.; Gounder, R. (2010). Energ consumtion and economic growth in New Zealand: results of trivariate and multivariate models, Energ Polic 38: Belke, A.; Dobnik, F.; Dreger, C. (2011). Energ consumtion and economic growth: new insights into the cointegration relationshi, Energ Economics 33(5): Bobinaitė, V.; Juozaavičienė, A.; Konstantinavičiūtė, I. (2011). Assessment of causalit relationshi between renewable energ consumtion and economic growth in Lithuania. Engineering Economics 22(5): htt://dx.doi.org/ /j01.ee Bojnec, S.; Paler, D. (2011). Economic efficienc, energ consumtion and sustainable develoment, Journal of Business Economics and Management 12: doi: / Camo, J.; Sarmiento, V. (2013). The relationshi between energ consumtion and GDP: evidence from a anel of 10 Latin American countries, Latin American Journal of Economics 50 (2): Chen, P.Y.; Chen, S.T.; Chen, C.C. (2012). Energ consumtion and economic growth new evidence from meta analsis, Energ Polic 44: Chen, S.T.; Kuo, H.I.; Chen, C.C. (2007). The relationshi between GDP and electricit consumtion in 10 Asian countries, Energ Polic 35: Chontanawat, J.; Hunt, L. C.; Pierse, R. (2008). Does energ consumtion cause economic growth?: evidence from a sstematic stud of over 100 countries, Journal of Polic Modeling, Elsevier 30(2): Dudzevičiūtė, G. (2012). Lithuanian securit оf energ sul and cоnsumtiоn in the cоntext оf the Eurоean Uniоn cоuntries. Internatiоnal Scientific Cоnference Whither Оur Ecоnоmies Оctоber Prоceedings: ISSN (оnline) Dudzevičiūtė, G. (2013). The research оf the ecоnоmic structural changes: energ cоnsumtiоn asect, Jоurnal оf Securit and Sustainabilit Issues 2(4): htt://dx.dоi.оrg/ /jssi (2). Eddrief-Cherfi, S.; Kourbali, B. (2012). Energ consumtion and economic growth in Algeria: cointegration and causalit 16

6 KSI Transactions on KNOWLEDGE SOCIETY analsis. International Journal of Energ Economics and Polic 2 (4): Eggoh, J.C.; Bangake, C.; Rault, C. (2011). Energ consumtion and economic growth revisited in African countries, Energ Polic 39 (11): Esso, L.J. (2010). Threshold cointegration and causalit relationshi between energ use and growth in seven African countries, Energ Economics 30: Eurostat database on energ indicators. Available on the Internet: htt://e.eurostat.ec.euroa.eu/ortal/age/ortal/energ/dat a/database. Fang, Y. (2011). Economic welfare imacts from renewable energ consumtion: the China exerience, Renewable and Sustainable Energ Reviews 15 (9): Fuinhas, J.; Marques, A. C. (2012). An ARDL aroach to the oil and growth nexus: Portuguese evidence, Energ Sources 7(3): Georgantooulas, A. (2012). Electricit consumtion and economic growth: analsis and forecasts using VAR/VEC aroach for Greece with caital formation, International Journal of Energ Economics and Polic 2 (4): Granger, C.W.J. (1969). Investigating causal relations b econometric models and crosssectral methods, Econometrica 37 (3): Gurgul, H.; Lach, L. (2011). The electricit consumtion versus economic growth of the Polish econom, Energ Economics 34 (2): Ho, Chun-Yu.; Siu, Kam Wing, (2007). A dnamic equilibrium of electricit consumtion and GDP in Hong Kong: An emirical investigation, Energ Polic, Elsevier 35(4): Kozhan, R. (2010). Financial Econometrics. Roman Kozhan & Ventus Publishing AS Lee, Chien-Chiang; Chiu, Yi-Bin, Oil rices, nuclear energ consumtion, and economic growth: new evidence using a heterogeneous anel analsis, Energ Polic, Elsevier 39(4): Li, J.; Li, Z. (2011). A causalit analsis of coal consumtion and economic growth for China and India, Natural Resources 2: Lin, B.; Wesseh, P.K. (2014). Energ consumtion and economic growth in South Africa reexamined: a nonarametric testing aorach, Renewable and sustainable energ reviews 40: Munim, J. M. A.; Hakim, M. M.; Abdullah-Al-Mamun, M Analsis of energ consumtion and indicators of energ use in Bangladesh, Economic Change Restructuring 43: Odhiambo, N. (2010). Energ consumtion, rices and economic growth in three SSA countries: A comarative stud, Energ Polic 38 (5): Odularu, G. O.; Okonkwo, Ch. (2009). Does energ consumtion contribute to economic erformance? Emirical evidence from Nigeria, Journal of Economics and Business 12 (2): Ouedraogo N. S. (2013). Energ consumtion and economic growth: evidence from the economic communit of West African States, Energ Economics 36: Ozkan, G. (2011). The Nabucco roject within the context of energ sul securit and international olitics, China-USA Business Review, 10 (8): Ozturk, I. (2010). A literature surve on energ-growth nexus, Energ Polic 38: Pane, J. E. (2010). Surve of the international evidence on the causal relationshi between energ consumtion and growth, Journal of Economic Studies 37(1): Squalli, J. (2007). Electricit consumtion and economic growth: bounds and causalit analses of OPEC members, Energ Economics 29 (6): Stern, D. I. (2011). From correlation to Granger causalit. Crawford School Research Paer No Tang, C.F.; Tan, E. CH. (2012). Electricit consumtion and economic growth in Portugal: evidence from a multivariate framework analsis, Energ Journal 33 (4), 27. htt://dx.doi.org/ / Tsani, S. (2010). Energ consumtion and economic growth: a causalit analsis for Greece. Energ Economics 32: Yildirim, E.; Sukruoglu, D.; Aslan, A. (2014). Energ consumtion and economic growth in the next 11 countries: the bootstraed autoregressive metric causalit aroach, Energ Economics 44: Zhixin, Z.; Xin, R. (2011). Causal relationshis between energ consumtion and economic growth, Energ Procedia 5: