The environmental Kuznets curve (EKC) hypothesis which was developed by Grossman and Krueger implies that clean environment demand increases over a certain level of income. Convergence of carbon emission can be considered an extension of EKC hypothesis, which emphasizes that the carbon emissions of countries will converge to a certain level. Convergence is important for the identification of the trend of carbon emission and to design emission abatement policies. This study investigates the stochastic convergence of per capita carbon emission in Organization for Economic Co-operation and Development (OECD) countries for 1960-2014 under cross-sectional dependence by using recently developed panel data methods. Empirical results are as follows: (i) per capita carbon emission is generally non-stationary according to the CADF test and the CIPS test. (ii) Panel KPSS test indicates that carbon emission per capita is overwhelmingly stationary in both country-specific and panel levels; (iii) Fourier panel KPSS test indicates that per capita carbon emission is overwhelmingly stationary at the country-specific level and non-stationary at the panel level. These findings are strongly in favor of convergence of per capita carbon emission among OECD countries. Consequently, it can be said that emission abatement policies aren't essential among these countries.
Keywords: Carbon dioxide emission; Environmental pollution; Fourier panel KPSS; Panel data analysis; Stochastic convergence.