Private Information in Dynamic Games

National Bureau Of Economic Research Inc, Cambridge MA

Investigators

Abstract

This project aims to provide and empirically test new economic theories about auctions that take place over time. The project will also develop a new theory about pricing in collusive markets. Auctions are a type of market that is used to sell many different goods and services. Developments in information technology have made use of this type of market increasingly attractive to both buyers and sellers. However, much of what economists know about auctions is based on static models where an auction happens once. This is unrealistic; in many settings buyers and/or sellers interact repeatedly in a series of auctions. In particular, the project will analyze the case where cost shocks may be serially correlated. This type of correlation arises when firms periodically improve their production process or gain expertise, or when supply or labor contracts are periodically renegotiated. Little is known about this case. The principal investigators will also apply the same basic model to analyze how firms collude over time in markets with the same kind of cost shock. One goal is to better understand why firms start price wars when their actions are fully observable by opponents. One possibility is that firms may be signaling the fact that they have especially low costs and thus require a larger share of the market in order to reduce their incentive to cheat on the collusive agreement. A third project will consider what happens in a market where the buyer and seller both have private information. The goal is to show that when the buyer and seller interact repeatedly, strong institutions and commitment power can guarantee that the market can overcome the information problem to reach efficient outcomes. A fourth project analyzes the effect of market design, in particular the auction format, on competition in auctions when entry is endogenous, such as timber auctions. These auctions include both small logging operations and large mills. Economic theory suggests that small bidders participate more in first-price auctions. The project will analyze data from timber auctions to see if this is in fact the case. The project will yield a variety of broader impacts to society, including include an improved ability of antitrust authorities to distinguish competition from collusion. Further, the results will give guidance on how governments and firms should design auction markets, in particular how the auction format and the information revealed about firm bids may affect the types of bidders who enter the auction.