Information and learning in markets with sequential auctions

Sequential auctions are used for wholesaling agricultural produce, fish, securities, and other commodities. In sequential auctions, each bid generates information. This is crucial in determining optimal bidding behavior. The focus of this dissertation is the effect of learning in these markets.In the second chapter it is assumed that bidders maximize expected utility. Risk averse bidders reduce the risk of losing an auction by placing higher bids. Again, there is a static auction that yields the same expected final allocation. Thus, although in the sequential auction bidders learn, this does not lead to an increase in expected payoffs.The third chapter introduces choice. The winner can choose how much to purchase. The expected payoff of the bidder is lower than when compared to the auction without choice. An implication of this is that goods may be allocated inefficiently.In Chapter 4 a model with a continuum of bidders' types is presented. This assumption leads to a different derivation of the equilibrium strategies and payoffs and hence a different type of equilibrium strategy. Specifically, all bidders use pure strategies in all auctions.In the first chapter, two identical goods are auctioned in sequence. Risk neutral bidders have a high or low valuation of the goods. They know their own valuation, but not their opponents'. After the first auction the winning bid is revealed, allowing the updating of beliefs regarding bidders' types. The updated beliefs are used to determine the second auction strategies, which affect the bidders' expected payoffs. Regarding the first auction, the anticipation of information affects bidders' values. Since bidders who lose the first auction have an expected payoff in the second auction, their values depend on this payoff. However, since the second auction payoff depends on the information generated in the first auction, bidders must anticipate the information generated in the first auction. It is shown that learning results in higher payoffs for the bidders. Also, the equilibrium is contrasted with a static auction, in which learning has no significance. The static auction leads to the same expected outcome.