Research

We study revenue-maximizing auction design when bidders can collude ex ante—i.e., before learning their values or deciding whether to participate in the auction. Bidders’ ex-post private information does not preclude efficient collusion, and the seller therefore cannot guarantee a higher expected revenue than from posting an optimal price and rationing the good. However, posted-price mechanisms leave the seller vulnerable to inefficient collusion. We construct mechanisms that guarantee the non-collusive posted-price expected revenue against a broad range of collusion—including by any ring that may comprise a subset of bidders, can commit to punish bidders who refuse to join, and of which some bidders may be unaware. Achieving such a guarantee requires protecting bidders’ ex-ante expected utilities against all such forms of collusion—a property that we term strong collusion-robustness. With symmetric bidders, our mechanism combines a posted price with a first-price knockout auction conducted on behalf of the bidders.

A principal sets bonuses to ensure that the induced game will have a unique outcome where all agents work rather than shirk. I primarily explore three notions of outcomes: (1) Nash equilibrium, (2) correlated equilibrium, (3) rationalizability. It is weakly more expensive for the principal to uniquely implement working under (3) than under (2), which in turn, is more expensive than implementing under (1). I show that when the production technology is anonymous, these weak inequalities hold at equality. When the assumption of anonymity is weakened to a condition I call aligned marginal contributions, I show that there can be a strict gap between uniquely implementation under Nash compared to correlated equilibrium, but there remains no gap between implementation under correlated equilibrium compared to rationalizability. Finally, I provide a sufficient condition under which there is a strict gap between achieving a unique correlated equilibrium and unique rationalizable strategies and characterize the cost of uniquely implementation in solution concepts that refine Nash equilibrium.

I study how an upstream market research firm provides information about demand conditions to a downstream product market. Demand can be high or low. The market research firm provides private signals to atomistic firms that compete with each other by producing quantities of an undifferentiated product. The market research firm chooses its information to maximize the fee it collects. I solve for equilibria of this game and demonstrate that the market research firm chooses an equilibrium to maximize the joint producer surplus. In environments where uncertainty about demand is small and strategic externalities between producers are weak, full information maximizes consumer, producer, and total surplus. In other environments, consumer, producer, and total surplus are not necessarily aligned.