Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


ZZL_upload2_Thesis.pdf (942.7 KB)

ETD Abstract Container

Abstract Header

Essays on Cheap Talk

Li, Zhuozheng
http://rave.ohiolink.edu/etdc/view?acc_num=osu1461277952

Abstract Details

2016, Doctor of Philosophy, Ohio State University, Economics.
In game theory, cheap talk model is used to describe and analyze communication between different players. The ``talk is cheap" in the sense that the content they say does not directly affect their payoff in the game. There is usually an informed player who can choose what to say to the decision maker. The decision maker is not informed but his decision will affect both players' payoff. And their payoffs are not aligned. This is contrast to the signaling game since the message is ``cheap". As long as both parties' preference are not too far apart, information revealing is still possible in this environment. In this dissertation, I focused on cheap talk models where the decision maker has binary choice which has many applications in project selection. It consists of three papers. The first chapter studies a one-period project selection model. The second chapter studies a dynamic search and project selection model. The third chapter studies a model of pre-communication before making coordination decision. In chapter 1, We study a one-period model in which each agent has a project and the two agents compete with one another to have their own project implemented. Under simultaneous communication, we show that all equilibria are partition equilibria, and the partitions of the two agents are intimately related: the interior partition points of the two agents have an alternating/staggering structure. In the most informative equilibrium, the agent with a smaller bias always has the sure option/veto power and weakly more messages. Moreover, the most informative equilibrium might not be the one with the maximum number of partition elements (which is typically the case in other cheap talk models). Surprisingly, in our setting both sequential communication and simple delegation are essentially outcome-equivalent to simultaneous communication. Finally, when the number of agent increases, each agent transmits more information in symmetric equilibrium. In chapter 2, I study a dynamic model in which agents sequentially search for projects. The novelty of the paper is that it combines cheap talk with sequential search. In particular, in each period only one agent is active or has a project, and he sends a message to the DM, who then decides whether or not to implement the project. If a project is adopted then the game ends, otherwise the current project is discarded and the game moves to the next period. Both agents and the DM share the same discount factor, embodying the cost of delay. In equilibrium, the active agent only has two messages: recommend the current project if the quality is higher than a cutoff and do not recommend otherwise. The DM always follows the recommendations. Due to the own project bias, each agent tends to exaggerate the quality of his own project, implying that the equilibrium cutoff is lower than the one that is optimal for the DM. Ex ante, the DM can design the rule that determines the period in which an agent will become active. The main result of the baseline model is that it is optimal for the DM to let only one agent search (the other agent is always inactive). This is because, over time, competition between the two agents increases agents’ incentives to exaggerate, which harms the DM. I also compare two super-symmetric rules: random rule (in any period an agent is randomly selected to be active) and alternating rule (two agents take turns to be active). I find that the DM gets a higher payoff under the random rule, as the competition between two agents is fiercer under the alternating rule. In an extension, I endogenize search effort. That is, an active agent needs to exert effort in order to successfully find a project. The main result in this extension is that sometimes the optimal rule for the DM is to let both agents search. The underlying reason is that competition, while it aggravates communication, encourages agents to search harder. For the same reason, the alternating rule can outperform the random rule with endogenized search effort. Chapter 3 is my current working progress where we studies pre-communication before players making coordination decision. The feature of the paper is that the two players are both senders and receivers. For example, in the battle of sexes, the husband prefers boxing while the wife prefers ballet. If they go to their favorite place, they get a positive payoff, otherwise, they get 0. So their preference are completely conflicted. If they don't coordinate, then they will face a positive mis-coordination cost. The payoff for their favorite place is private information. We show that the communication equilibrium is influential and with communication, there is an improvement at ex ante. When the mis-coordination cost is higher, their payoff is less. We also study the environment with mediator and show that the mediator can manipulate the probability of mis-coordination to improve the social welfare.
Huanxing Yang (Advisor)
Peck James (Committee Member)
Ye Lixin (Committee Member)
122 p.
Economics
cheap talk; sequential search; coordination; mechanism design

Recommended Citations

Citations

  • Li, Z. (2016). Essays on Cheap Talk [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1461277952

    APA Style (7th edition)

  • Li, Zhuozheng. Essays on Cheap Talk. 2016. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1461277952.

    MLA Style (8th edition)

  • Li, Zhuozheng. "Essays on Cheap Talk." Doctoral dissertation, Ohio State University, 2016. http://rave.ohiolink.edu/etdc/view?acc_num=osu1461277952

    Chicago Manual of Style (17th edition)

osu1461277952
882
© 2016, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.