Exploring the Carbon Kuznets Hypothesis

The Carbon Kuznets hypothesis conjectures an inverse U–shape relation between GDP and carbon dioxide emissions. We investigate a number of empirical problems with this hypothesis by way of both econometric analysis and CGE modelling. The econometric analysis takes into account the possibility of unit root non–stationary regressors. On a panel of 107 countries covering the years from 1986 to 1998 we find evidence for unit root non–stationarity in log GDP and log emissions. Our discussion therefore focusses of potential pitfalls in estimating the Carbon Kuznets curve in the context of non–stationary panels context. We conclude that current practice in the literature fails to take these potential problems adequately into account. The second conceptual problem considered in the paper is the question of how to interpret an observed inverse U–shaped relationship. With the help of a small GCE model, we illustrate the danger of using observed GDP–emission patterns directly as a policy guide. Our model economy, where decarbonization is exogenous, demonstrates in particular that a carbon policy relating to income levels may not be appropriate even in the face of an observed inverse U–pattern between income and emissions. JEL Classification: Q20, C12, C13