We empirically test stability of climate change coalitions with the STAbility of Coalitions model (STACO). The model comprises twelve world regions and captures important dynamic aspects of the climate change problem. We apply the stability concept of internal and external stability to a cartel formation game. It is shown that only if benefits from global abatement are sufficiently high, stable coalitions emerge, though they only marginally improve upon the Nash equilibrium. We explain this phenomenon by analyzing the individual incentive structure of all regions and relate our results to the predictions of theory.