A Dynamic Nonlinear Panel Decomposition Model for the Global Environmental Kuznets Curve

We propose a decomposition of the Environmental Kuznets Curve (EKC) based on a differenceof-sigmoid functions with one for the economic-growth effect and one for the green-economy effect. This nonlinear function replaces the traditional quadratic specification in panel data models. We further extend the framework with unobserved stochastic trends to capture time-varying features such as evolving energy mixes and technological advances. The panel model can be represented in state-space form. We show that parameter estimation can be carried out by nonlinear regression and Kalman filter methods. In our empirical study, we consider a panel of pollution (CO2 emissions) and economic output variables for 15 major economies (G7, Australia and 7 emerging countries), representing approximately 75% of the global economy. We extract the common EKC where its parameters are pooled or partially pooled across all countries. The empirical results show statistically significant evidence of a global relevant EKC.