Abstract
Given a sample of bids from independent auctions, this paper examines thequestion of inference on auction fundamentals (e.g. valuation distributions,welfare measures) under weak assumptions on information structure. The questionis important as it allows us to learn about the valuation distribution in arobust way, i.e., without assuming that a particular information structureholds across observations. We leverage the recent contributions of\cite{Bergemann2013} in the robust mechanism design literature that exploit thelink between Bayesian Correlated Equilibria and Bayesian Nash Equilibria inincomplete information games to construct an econometrics framework forlearning about auction fundamentals using observed data on bids. We showcaseour construction of identified sets in private value and common value auctions.Our approach for constructing these sets inherits the computational simplicityof solving for correlated equilibria: checking whether a particular valuationdistribution belongs to the identified set is as simple as determining whethera {\it linear} program is feasible. A similar linear program can be used toconstruct the identified set on various welfare measures and counterfactualobjects. For inference and to summarize statistical uncertainty, we proposenovel finite sample methods using tail inequalities that are used to constructconfidence regions on sets. We also highlight methods based on Bayesianbootstrap and subsampling. A set of Monte Carlo experiments show adequatefinite sample properties of our inference procedures. We illustrate our methodsusing data from OCS auctions.