Markov regime switching vector autoregressive models and nonlinear causality analysis: An application for the relationship between renewable energy cunsumption, CO2 emission and economic growth in OECD countries

The literature frequently mentions the non-linear time series modeling methods, which have been developed as an alternative to classical linear time series modeling methods in an attempt to model the financial variables more accurately since the 1980s. It is observed that non-linear time series models are more competent in explaining certain features of variables, compared to classical linear time series models. Although there are a lot of non-linear time series models in the literature the most preferred model is MS-VAR modeling method which is also used in this study. The MS-VAR modeling approach is a method that allows to use different circumstances and regimes or to be different dynamics in different regimes to model structural and behavioral changes of variables. The aim of this study is to contribute shaping the energy which is the major resource of progress of countries increasing production and life standards, in accordance with clean, renewable and sustainable targets. In this context, using the MS-VAR model approach that is mentioned in the theoretical part of the study, it is investigated the relationship between the CO2 emission, renewable energy consumption and economic growth covering the 1961-2011 period in 8 OECD countries that involve Austria, Canada, Portugal, Sweden, United States, Finland, Australia and Turkey. As a result of the analyses, it was determined that the CO2 emissions, renewable energy consumption and economic growth series have showed a changing structure according to nonlinear regimes and that MS-VAR models have reflected the economics of countries better than linear VAR model. Keywords: Markov Regime Switching, MS-VAR Model, MS-VECM, MS-Granger Causality