Bootstrap panel causality: An Empirical application on transition economies

This study investigates the energy-led growth hypothesis for Transition Economies via analysing Granger Causality between GDP Growth per capita and Energy Use Per Capita for. For this purpose 17 countries located at Central and Eastern Europe and Caucasian region (Albania, Armenia, Azerbaijan, Blearus, Bulgaria, Czech Republic, Georgia, Hungary, Latvia, Lithuania, Macedonia, Moldov, Poland, Romania, Slovak Republic, Slovenia and Ukraine) are chosen and a panel dataset consisting of these countries for the period of 1990-2011 is studied. In the study Granger Causality is investigated using Bootstrapped Panel Causality approach proposed by Kónya (2006), which gives consistent results in case of cross-sectional dependecy and heterogeneity between countries and SUR is used to estimate the causality models. Causality is examined for two scenerios; one with a trend and one without a trend. According to results, there is bidirectional causality between Energy Use per capita and GDP growth per capita for Albania, Armenia, Belarus, Georgia, Hungary and Latvia; unidirectional causality running from Energy Use per capita to GDP growth per capita for Czech Republic, Lithuania, Romania and Slovak Republic and unidirectional causality running from GDP growth per capita to Energy Use per capita for the case of Moldov. Obtained results show that energy led growth hypothesis is supported by 10 countries. These countries should improve their energy policies considering this information. On the other hand, as energy led growth hypothesis is not supported by Azerbaijan, Bulgaria, Macedonia, Moldov, Poland, Slovenia and Ukranie these countries can adept energy saving policies.