Testing stationarity of per capita energy consumption across the globe: a better picture using the wavelet framework.

The issue of examining the stationarity of per capita energy consumption (PCEC) series is central in energy literature. Most of the existing studies examine this issue by employing unit root tests based on the time domain only providing a partial picture of the problem at hand due to not including the information related to the frequency domain. In contrast to existing studies, the present study uses a recently projected wavelet-based unit root testing scheme-Fourier wavelet KSS (FWKSS) introduced by Aydin and Pata. Energy 207:118245 (2020) which has several advantages over the existing wavelet unit root tests such as the ones proposed by Fan and Gençay (2010). Specifically, the FWKSS takes care of any possible smooth structural break in the data series under examination. This study re-examines the stationarity of PCEC across the globe covering for the first time a total of 133 countries and including all income groups constituting 49 high-, 76 middle- and 8 low-income countries. The selection of countries is made based on the availability of at least 25 years of annual time series data. The empirical results based on latest available data reveal that among 133 chosen countries, 67 countries are found to have non-stationary PCEC including 20 high-, 43 middle- and 4 low-income countries while the remaining 66 countries are found to have stationary PCEC including 29 high-, 33 middle- and 4 low-income countries. PCEC is found to be stationary in 79 of them including 35 high-, 40 middle- and 4 low-income countries. Some relevant policy implications are discussed as well.