Conclusion
The world faces major environmental problems, such as air pollution, global warming and climate change today since (i) the world depends on fossil energy sources and (ii) GHGs, such as CO2, CH4, N2O and fluorinated gases, arise from the usage of fossil sources. Policy makers consider renewable energy sources regarded as clean energy sources to reduce environmental problems. Hence, many global meetings have been organized, and renewable energy production and consumption have been encouraged in many countries in recent years. These developments have inevitably affected the energy literature, and many papers have been conducted in order to examine the effects of renewable energy consumption and/ or production on GHGs. When one analyses these papers, he/ she will observe that they employ fixed parameter estimation methods and ignore time-varying parameters of the determinants of GHGs. Therefore, there seems to be a research gap about the time-varying effects of non-renewable and renewable energy consumption/production on GHGs.
To fulfil this gap to a certain extent, this paper investigates the effects of electricity production from non-renewable and renewable sources on CO2 emissions in Turkey by employing both fixed parameter and time-varying parameter estimation methods over the period 1970–2013. To this end, the paper, first, performs unit root and cointegration tests and explores that the variables are integrated of order one and that there is a cointegration relationship among variables. Then, the paper employs the FMOLS and DOLS estimators and the Kalman filter. The findings obtained from the methods yield that CO2 emissions are positively related to GDP, electricity production from non-renewable sources and electricity production from renewable sources. Another important finding that the methods present is that the coefficient of electricity production from renewable sources is lower than that of electricity production from nonrenewable sources. Put differently, electricity production from renewable sources leads to fewer CO2 emissions than electricity production from non-renewable sources does. On the other hand, one can observe from the figure which presents timevarying parameters that the break dates indicated by Narayan and Popp (2010) unit root and Maki (2012) cointegration tests do not have considerable effects on time-varying parameters. In other words, time-varying parameters did not exhibit radical and/or permanent changes after the breaks. That is to say, the breaks did not have considerable effects on the relationships among variables.