Amber Wang

Ph.D. Candidate

Phone: (858) 232-9811

Email: amberwang@ucsd.edu

Research Statement

Research Statement (PDF)

Working Papers

"Cheaper Talk" (with Peicong Hu) Job Market Paper

Technological advancement has been lowering the cost of information provision. Thanks to new information technologies, everyone with access to the Internet can provide information at minimal cost. In this paper, we introduce a fixed cost of "talking" into the canonical cheap talk model and allow the sender to choose whether to "talk" or not. We explore the following questions: (1) Is a sender with more accurate information or a smaller bias necessarily more valuable to the receiver? (2) Does a lower "talking cost" always benefit the receiver? (3) Will the sender choose ex ante the communication technology that is optimal for communication? The main results are: (1) A sender with less accurate information or a larger bias can be more valuable to the receiver by being more motivated to provide information. (2) Too high a talking cost discourages information provision, whereas too low a cost can reduce the effectiveness of communication. (3) A sender may prefer a "cheaper" but less effective communication technology ex ante.

"Knowledge Sharing in Research Contests" (preliminary draft available upon request)

I study the organization of a two-agent research contest where each agent has to achieve two breakthroughs to achieve the discovery. The first breakthrough generates some intermediate knowledge. The second breakthrough generates the discovery which is only valuable when achieved for the first time. Sharing the knowledge of the agent who has achieved the first breakthrough (the "leader") to the agent who has not (the "laggard") can assist the latter to catch up. A principal designs how to motivate agents using non-negative transfers, and decides whether to share the leader's knowledge to the laggard. A benevolent principal can achieve the social optimal outcome even if she relies on agents' voluntary reports to learn the arrival of breakthroughs. Nevertheless, to prevent the laggard from over-investing in catching-up, a "losing prize" to the laggard is necessary even if he does not achieve any breakthrough in the contest. The total transfers to agents can hence exceed the value of the discovery, leading to inefficiency if the principal is budget constrained. In this case, the principal may find it optimal to exclude the laggard from the contest upon the arrival of the knowledge. As an extension, I also study when a profit-maximizing principal would share (or not share) the leader's knowledge.

"The (De)Formation of Gym-attending Habits"

The history-dependency of habitual behavior, as suggested by the classic habit formation theory, implies the possibility for a one-shot behavior intervention to have effects persistent in the long run. By instrumenting previous gym-attendance with temperature, I first establish the existence of habit formation effects on spontaneous gym-attending behavior. I then argue that, although one-shot intervention does increase people's subsequent exercise frequency after the intervention is over, the exercising habit is also sensitive to small negative shocks (such as bad weather) which can lead to habit deformation. Furthermore, this sensitivity does not diminish over time and the exercising habit does not become stabilized if one exercises more often. Consistent with the fact that paying people to go to the gym only has short run effects in most previous literature, the finding in this paper points out the flaw of this popular intervention approach.

Work in Progress

"Conventional Equilibrium" (with Joel Sobel)

"Knowledge Sharing in Reserach Contests: Evidence from NIH Data Sharing Policy"