Waiting for News in The Dynamic Market for Lemons (Job Market Paper)
July 2008, Joint with Brendan Daley
[Download]
Trade breaks down in the market for lemons because high-type sellers have a reservation value greater than expected market value. Unraveling occurs and only the lowest types trade. Two related questions arise: What happens the next day? And, from where does the reservation value come? We model these considerations in a dynamic setting with gradual arrival of noisy information. We characterize the unique equilibrium in a continuous-time framework. The equilibrium involves a region of no trade or market failure. The no-trade region ends in one of two ways: either enough good news arrives restoring confidence and markets re-open or bad news arrives making buyers more pessimistic forcing market capitulation i.e., a sell-off of low value assets. Reservation values arise endogenously from the option to sell in the future. Our model also encompasses dynamic signaling environments. In a dynamic setting with sufficiently informative news, Spence's Job Market Signaling and Akerlof's Market for Lemons have the same unique equilibrium. The predictions help explain "irrational" trading patterns in financial markets.
Market Signaling with Grades
December 2007, Joint with Brendan Daley
[In Revision]
We consider a market signaling environment where the sender chooses a perfectly observable
signal that also generates a stochastic grade, which may depend on both the sender's type and the
signal chosen. We fully characterize the set of equilibria satisfying the D1 refinement, which is often
a unique equilibrium. Equilibrium can involve full or partial pooling when grades are informative
and converges to the least cost separating outcome as grades decrease in informativeness. The
equilibrium precisely characterizes the tradeoff the high type faces when deciding how much to rely
on the costly signal and how much to rely on the grade.
The presence of informative grades yields a
Pareto improvement. The high type's welfare increases and his costly (but unproductive) signaling
decreases. Grades never hurt the low type's welfare, and actually increase it when the prior puts
suffcient weight on the high type. We derive conditions under which equilibria converge to full
pooling at the lowest signal as the prior's weight on the high type tends to one offering a resolution
to a long-standing paradox in the signaling literature. Our results are virtually unaffected when
signaling is assumed to be productive.
News Arrivals in Asymmetric Markets: When No News is Good News or Bad News
April 2008, Joint with Brendan Daley
Streakiness in Baseball: The Hot Hand Revisited
April 2007, Joint with Jeffrey Zwiebel