2010-04 Analyzing policy risk and accounting for strategy

We examine the potential for simple auction mechanisms to efficiently allocate arrival and departure slots during Ground Delay Programs (GDPs). The analysis is conducted using a new approach to predicting strategic behavior called Predictive Game Theory (PGT). The difference between PGT and the familiar Equilibrium Concept Approach (ECA) is that PGT models produce distribution-valued solution concepts rather than set-valued ones. The advantages of PGT over ECA in policy analysis and design are that PGT allows for decision-theoretic prediction and policy evaluation. Furthermore, PGT allows for a comprehensive account of risk, including two types of risk, systematic and modeling, that cannot be considered with the ECA. The results show that the second price auction dominates the first price auction in many decision-relevant categories, including higher expected efficiency, lower variance in efficiency, lower probability of significant efficiency loss and higher probability of significant efficiency gain. These findings are despite the fact that there is no a priori reason to expect the second price auction to be more efficient because none of the conventional reasons for preferring second price over first price auctions, i.e. dominant strategy implementability, apply to the GDP slot auction setting.