Asset markets under asymmetric information: An experimental investigation

This dissertation consists of three essays. In the first essay I examine the relative impact of asymmetric information on the performance of two commonly employed asset trading mechanisms, the call auction and a continuous market mechanism. I find that the adverse selection costs suffered by uninformed traders are significantly lower under the call auction, despite no significant reduction in average price efficiency. This result suggests that the assessment of the costs and benefits of insider trading should take place within the context of specific trading mechanism.; In the second essay I note that a characteristic of most models of microstructure is that either the presence of insider(s) or the probability of a trade being initiated by an insider is common knowledge, although this is not a feature of field asset markets. I employ the methodology of experimental economics to explore the importance of this common knowledge assumption by comparing the markets in which the presence (and number) of informed traders is known prior to the opening of trade with markets in which their presence must be inferred from trading patterns. The markets in which the number of informed traders is known are on average more liquid, and prices are more efficient. These results suggest caution in interpreting many extant models of insider trading for policy purposes.; The third essay is joint work with Christopher G. Lamoureux. This essay notes that the market microstructure literature has ignored the fundamental mechanism design problem. We report results from experimental markets with liquidity traders, informed traders, and market makers where in some trading periods we allow bilateral trade to take place in addition to public trade with dealers. In the absence of the bilateral search alternative, dealer profits are large--unlike in models with risk neutral, competitive dealers. However, when we allow traders to participate in the search market, dealer profits are close to zero. There is no evidence that price discovery is less efficient when the market makers are not the only game in town.