文摘
Essay 1: A Model of Dynamic Information Disclosure. We study a dynamic game in which a financial expert seeks to optimize the utilization of her private information either by information disclosure to an investor or by self-using. The investor may be aligned or biased: an aligned investor always cooperates on the disclosed information, whereas a biased investor may strategically betray the expert. We characterize the joint dynamics of the experts information disclosure and the investors type revelation and show that, by the process of gradual information disclosure, the expert can significantly alleviate the hold-up effect exerted by the biased investor. In particular, we show that the equilibrium dynamics of the players interactions is unique. We also examine how the expert can further improve her utilization of information by committing to a deadline or by committing to a particular pattern of information disclosure. Essay 2: Reputational Concern with Endogenous Information Acquisition. We develop a reputational cheap talk model in which an expert acquires and conveys information and a decision maker takes a payoff-relevant action. The expert may be aligned or biased: an aligned expert cares about the decision makers payoff and would like to be known as aligned, whereas a biased expert always distorts information toward a particular direction. Our main finding shows that the aligned experts reputational concern may have a non-monotonic effect on his incentive to acquire information; that is, he acquires better information if and only if his reputational concern is moderate. Another finding shows that, although the biased type of expert only distorts information transmission, the existence of this type may actually increase the decision makers payoff. We also examine how delegation may affect the players decisions and payoffs in this essay and show that even with the rights to better use the information ex post, the aligned experts information acquisition incentive may be weakened ex ante. Finally, we show that the decision maker prefers communication to delegation whenever informative communication with information acquisition is feasible. Essay 3: Learning, Belief Manipulation and Optimal Relationship Termination. We study a dynamic agency problem in which a principal and an agent interact on a project with initially unknown quality. A key feature in this problem is that the agents hidden actions can give rise to hidden information about the project quality, which enables the agent to benefit from manipulating the principals learning process. In particular, the agents attempt on belief manipulation varies in his own assessment about the project quality. We examine how the principal can structure the provision of incentives by resorting to relationship termination. Relationship termination has two opposing effects: it destroys the surplus that the principal can obtain from the relationship continuation, but it also lowers the informational rents that the agent can capture from the belief manipulation. We show that in equilibrium the optimal rule of relationship termination follows a cut-off strategy: it is introduced in the contracts only when the expected relationship value is higher than a threshold value. In consequence, the dynamic agency cost presents a non-monotonic relationship with the project quality. We also examine how a limitation on the principals payment ability shifts the agents incentive on belief manipulation backwardly. Essay 4: Multilateral Bargaining with an Endogenously Determined Procedure. We consider a multilateral bargaining game in which a manager negotiates sequentially with several workers to share the units of surplus. The novel feature of our setup is that the manager can determine the ordering of her bargaining opponents endogenously. We show that double-sided hold-up effects arise in this game: the workers can hold up the manager by coordinating their moves, whereas the manager can hold up the workers by switching between the opponents. The interaction of these two effects gives rise to multiple equilibria, some of which present inefficient delays. Moreover, the delay may be bounded away from zero even if the time interval between two offers becomes arbitrarily small.