Explaining the structure of CEO incentive pay with decreasing relative risk aversion

It is established that the standard principal-agent model cannot explain the structure of commonly used CEO compensation contracts if preferences with constant relative risk aversion are postulated. However, we demonstrate that this model has potentially a high explanatory power with preferences with decreasing relative risk aversion, in the sense that a typical CEO contract is approximately optimal for plausible preference parameters.

► The explanatory power of the principal-agent model for the structure of CEO incentive pay is low with CRRA preferences.
► The explanatory power is potentially high with preferences with decreasing relative risk aversion.
► The optimal contract looks similar to typically observed contracts, and it does not generate substantial cost savings.
► The predictions of the model are consistent with the shift away from stock-options in the 2000s.