By now it is clear that man-made global warming cannot be avoided, but only limited. Therefore, the question of how best to combine mitigation and adaptation strategies becomes crucial for climate policy. By first principles of economics, the key task is to link climate damages to the incentives driving greenhouse gas emissions. A major problem is that climate damages usually happen at great temporal and spatial distance from the actions causing them. Moreover, the causal mechanisms involved are stochastic in nature and are shaped by a whole array of factors. Finally, global climate agreements will be soft on targets and even softer on implementation for a long time to come.

Under these circumstances, the European Emissions Trading System offers interesting opportunities for social learning. I propose to use small fractions of the revenues generated by this system to build a fund for public-private partnerships insuring climate risks. The existing European solidarity fund formed after the great floods of 2002 may be used for this purpose. In the talk, I show how the problem of attributing fractions of actual damages to climate change can be addressed with methods of Bayesian decision theory.

Similar schemes could be developed in U.S., China, and other regions. The UN Framework Convention for Climate Change could then be used to gradually strengthen the linkages between such schemes, until global emissions have been reduced to the point of avoiding dangerous interference with the climate system.