The authors complement the existing knowledge in the renegotiation literature on infrastructure concessions by analyzing government-led renegotiations. They first propose a multiple-period theoretical framework in which both Pareto-improving and rent-shifting renegotiations at the initiative of the government can occur. They then perform an empirical analysis based on a sample of 307 water and transport projects in five Latin American countries between 1989 and 2000. While some of the main insights from the previous literature are unchanged, for example concerning the importance of having a regulator in place when awarding concessions and the fragility of price cap regulatory schemes, there are also significant differences as predicted by the model, in particular with respect to the effect of investment and financing, as well as the corruption variables. The authors provide additional evidence showing that a good regulatory framework is especially important in contexts with weak governance and political opportunism.