The relationship between energy consumption, CO2 emissions, and economic growth in Turkey: evidence from Fourier approximation

Given that Turkey has recently committed itself for the first time to reducing its CO₂ emissions in the interest of sustainable growth in not only Turkey but also the world as a whole, this paper examines the relationship between energy consumption, CO₂ emissions, and economic growth in Turkey for the period 1960–2014. In view of the different findings concerning causality and the character of the relationships between these variables revealed in our review of past studies (in most cases using quite different methods), this paper utilizes several different but related methodological approaches for identifying causal relationships. These include both the Toda and Yamamoto (1995) approach, the Fourier Toda-Yamamoto for Cumulative Frequency approach developed by Nazlioglu et al. (2016), vector error correction model (VECM) methodology, and the asymmetric Granger causality test proposed by Hatemi-J (Empir Econ 43:447–456, Hatemi-j 2012). Our results show that, when we apply the popular Toda-Yamamoto model, causality in these relationships is not confirmed even among any of the relevant variables in Turkey. Yet, when the Fourier Toda-Yamamoto tests for cumulative frequency are employed, we find unidirectional causality running from GDP per capita to emissions of CO₂ per capita. Moreover, when we utilize the VECM methodology, the results show that long-run causality exists from GDP per capita and energy to CO₂ emissions. When we apply the asymmetric causality tests, the results provide even stronger evidence for a unidirectional causal relationship from GDP per capita to CO₂ emissions. As a result, the latter sets of results, based on more realistic conditions, suggest very strongly that, if Turkey is to meet the objectives of its ambitious Climate Change Action Plan commitment to the United Nations to reduce its CO₂ per capita emissions relative to its past trends by up to 21% over the coming 2021–2030 decade, it is going to get very serious about the best way to do this as soon as possible.