Empirical investigation of global warming models in economics have started basically with simple calibrations of control models where the whole world was represented by a single agent. Theoretical advances in disaggregated game theoretic models have also motivated the analysis of disaggregated empirical models which have been extended to computable dynamic general equilibrium models or explicit solution of games related to global warming. The present paper is an empirical investigation of co-operative and noncooperative solutions in global warming using a dynamic disaggregated model of five groups of countries. This paper differs from previous empirical analysis by econometrically estimating benefit functions for each group of countries using cointegration techniques and by explicitly solving an optimal control model for the co-operative solution and a differential game model with linear Markov strategies for the noncooperative solution. The Markov strategies assumption is more realistic since it allows for the emissions of the groups to be affected by the actions of the other groups through the accumulation of CO2.