Promising somewhat more than anyone expects to be delivered remains ingrained in international agreements, clouding their expected aggregate outcomes and how to assess the Parties’ performance. This persistence is explained by systematic overpromising being predictable and, up to a point, productive. I characterize this regime and its consequences and provide an empirical application to the Kyoto Protocol. Overpromising by governments or Parties can be part of a sustainable strategy for electoral success, and varies with socio-economic determinants that characterize the group regime. Thus targets need to be adjusted for regression-predicted overpromising to reveal the rationally-expected outcomes and individual countries’ performance is best identified by deviations of outcomes from their adjusted rather than the agreed targets.