The climate change debate has drawn attention to the problem of greenhouse gases emissions into the atmosphere. One of the most important issues in the policy debate is the role that should be played by developing countries in joining the commitment of developed countries to reduce GHG emissions, and particularly CO2 emissions. This debate calls into play the relationship between energy consumption, CO2 emissions and economic development. In this paper we use a panel data model for 110 world countries to estimate the relationship between CO2 emissions and GDP and to produce emission forecast. The paper contains three major results: (i) the empirical relationship between carbon dioxide and income is well described by non-linear Gamma and Weibull specifications as opposed to more usual linear and log-linear functional forms; (ii) our single equation reduced form model is comparable in terms of forecasted emissions with other more complex, less data driven models; (iii) despite the decreasing marginal propensity to pollute, our forecasts show that future global emissions will rise. The average world growth of CO2 emissions between 2000 and 2020 is about 2.2%/yr, while that of non-Annex 1 countries is posted at 3.3%/yr.