A Bayesian perspective on the Environmental Kuznets Curve

Existing studies on the Environmental Kuznets Curve (EKC) offer mixed results and ignore the issue of parameter uncertainty in a frequentist inference framework. In this paper, we adopt a Bayesian approach to test the EKC hypothesis using a static panel model for 170 countries over the period 1990–2019. Our contribution is both theoretical and empirical. On the theoretical side, we propose two new encompassing prior Bayesian tests to evaluate the presence and characteristics of the EKC in a panel setting. On the empirical side, this novel approach allows us to confirm the EKC hypothesis for CO2 emissions and to examine and quantify the high uncertainty reflected in the wide posterior distribution of turning points. We identify the source of this uncertainty and show that isolating and excluding oil countries from the panel substantially reduces uncertainty and lowers tuning points. This finding may suggest a more optimistic outlook for achieving the environmental benefits of growth. However, this result should be interpreted with caution, as static panel data models capture cross-country associations rather than long-run equilibrium relationships.