Over the last decades, democratic countries have signed hundreds of international environmental treaties, but most of them fail to specify sanctions on non-compliers. This appears as a puzzle in standard economic models. To rationalize it, we develop a positive theory of international treaties where the political incumbent negotiates treaties that will influence the next policymaker as well as the probability of winning future elections. The incumbent does not get any electoral advantages by signing a strong treaty, or by not signing any treaty at all, since the incumbent and the challenger would then behave in a similar way once elected. If a treaty is weak, however, it will be complied with by the relatively "green" party but not by the relatively "brown" party. Since the median voter's preference vary with the level of the sanction, the reelection probability will always be maximized by signing a weak treaty, regardless of the identity of the incumbent. Our theory also predicts that countries will rely too much on technology (rather than sanctions) as a way to make compliance credible, and the equilibrium scope of a treaty may be too broad.