ABSTRACT
This chapter describes the institution, summarizes a theory that supports Marshall's intuition, and discusses how those theoretical predictions may be confronted with experimental evidence. The theoretical analysis of the k-double auction (k-DA) can be radically complicated if even minor modifications of its rules are made so that it no longer uses a market-clearing price to mediate trade. Following Harsanyi, the equilibrium concept that has been used to study the k-DA theoretically is symmetric Bayesian Nash equilibrium. The chapter describes a minor modification, which John H. Kagel and William Vogt used during the pilot phase of an experimental investigation of the k-DA, and discusses the analytical problems it causes. It reviews the testable propositions that the theory implies and discusses the possibility of using experimental methods to test these implications. The Bayesian theory of double auctions posits sophisticated behavior on the part of traders within a game of incomplete information.
