Research Papers
Dynamic Communication with Trading Commissions - Slides, Online Appendix
This paper studies a dynamic cheap-talk communication game to investigate the role of trading commissions in broker-investor relationships. An investor’s optimal portfolio mix is fixed and privately known to a broker. The stochastic evolution of the portfolio-mix can be adjusted by the investor at a fixed cost paid to the broker. I consider two cases - whether the broker (sender) can or cannot send uninformative messages to the investor (receiver). In the Markov-Perfect Equilibria, the portfolio mix has an informational value to the investor because it influences the broker’s incentives for truth-telling. The presence of uninformative messages further restricts the set of such ‘truth-telling’ portfolio-mix-values. As the investor’s payoff from being informed becomes very large, and the sender is able to send uninformative messages, the only pure-strategy equilibrium involves no informative communication, and any informative equilibrium involves delays in communication. I also study an extension with proportional adjustment costs.
Research In Progress
Dynamic Information Acquisition with Trading Commissions - Slides
I study a dynamic game between an investor and a broker with delegated information acquisition. The investor wants to maintain her portfolio of risky assets at an optimal benchmark level, which is stochastic and unobservable. The broker has the ability to generate public information about the benchmark at a cost, and earns a fixed commission whenever the investor changes her portfolio. I compare this to the case where the investor is able to acquire information herself, and show that with information power, the broker delays information acquisition and the investor increases her trade likelihood. This characterizes the inefficiency in the broker’s actions as compared to the case with the optimal contract, which is a direct payment for research.
Reputational Cheap Talk
I study a dynamic model of reputational cheap talk in which a constant state determines whether a crisis will or will not occur at some random time in the future. Conditional on his ability, a sender becomes increasingly informed about this constant state. A low-ability sender does not receive any information but can send false reports to appear informed. The sender gets a high payoff if the passive receiver's beliefs put more weight on him being of high-ability. I characterize the distribution of the time of false reports by the low-ability sender and show that the distribution of false reports, which the sender makes to pretend that he has become informed, put more weight on earlier times than the true distribution for a high-ability sender. An application on reports by financial experts suggests that they may tend to over-enthusiastically report possible events such as stock booms or crises in order to appear highly skilled.