Recent papers by Wagner and Vollebergh et al. point out some fundamental econometric problems with traditional methods of estimating the environmental Kuznets curve (EKC) and propose alternative approaches that avoid these issues. Wagner notes that traditional methods do not take into account the presence of powers of unit root variables and cross-sectional dependence in the data while Vollebergh et al. point out that the time effects are not uniquely identified in the EKC model. The between estimator is a simple estimator that also addresses the concerns of these authors. It makes no a priori assumption about the nature of the time effects and is likely to provide consistent estimates of long-run relationships in real world data situations. I apply several common panel data estimators including the between estimator to OECD and global carbon and sulfur emissions datasets. The between estimates of the sulfur-income elasticity are 0.732 in the OECD and 1.067 in the global data set and the estimated carbon-income elasticity is 1.612 in the OECD and 1.509 globally.
|Publication status||Published - 2010|