- Restrictions on Access:
- Restricted (PSU only).
- The dissertation focuses on understanding contests, a widely used mechanism that allocates scarce resources among competing individuals. The first chapter "Creative Contests: Theory and Experiment'' draws on contests in which contestants are uncertain about the contest organizer's exact preferences. Such uncertainty appears routinely in design and research competitions, as well as in job-promotion competitions wherein innovation is appreciated by the CEO. In this paper, I develop a model of creative contests in which two firms compete by adjusting their designs when they are uncertain about the contest organizer's ideal design. As an application of the model, I investigate whether the contest organizer would be better off if she eliminates this uncertainty by disclosing her ideal design to the participants. I find that disclosure is not always optimal. I then conduct a laboratory experiment to test whether my model's predictions are consistent with real world observations, when decision makers are not necessarily risk neutral and fully rational. This paper advances research on contests in the following directions. First, it introduces a tractable way to model creativity in contests and analyze how contestants compete therein. Second, a model of creative contests that accounts for such uncertainty enables us to study many new questions. The disclosure problem of the contest organizer demonstrated in the paper is one of such examples. Lastly, this paper also contributes to the experimental literature on contest by examining directly how and whether information about the organizer's exact preferences affects human behaviors in a contest experiment. The second chapter "On Optimal Favoritism in All-Pay Contests'' looks at another design problem of contests: when contestants have different abilities, how could the contest organizer take advantage of this asymmetry to her benefit? I analyze the optimal favoritism in a complete-information all-pay contest with two players, whose costs of effort are weakly convex. The contest designer could favor or harm some contestants using two instruments: head starts and handicaps. I find that any given player's effort distribution is ranked in the sense of first-order stochastic dominance according to how symmetric the players are. Consequently, regardless of which instrument is used and what the designer's objective is, "leveling the playing field" is optimal. My findings in this paper extends the conventional wisdom of "leveling the playing field" in two dimensions. First, instead of focusing on particular forms of objective functions, my analysis shows that joint efforts increase in the sense of first order stochastic dominance as contestants become more symmetric. Therefore, as long as the organizer values individual effort, she would benefit from "leveling the playing field". Second, this paper does not restrict contestants' cost functions to be linear, which is widely assumed in the literature. I show that "leveling the playing field" is optimal when cost functions are weakly convex, and prove that it may not be the case when weak-convexity are not satisfied. In the third chapter "Single Prize Contests'', we turn our focus on single prize contests and explore a more fundamental question: how to solve for equilibria when more than one player receives zero payoff while participating in the contest? Specifically, we study single prize contests with three players. Existing solution techniques are not readily applicable because it is hard to find a unique initial condition to start the algorithm, when two or more players receive zero payoff. We develop an innovative algorithm which has a unique initial condition and can generate all equilibria. We show that there might exist multiple equilibria and provide a sufficient condition for a unique equilibrium. The technical innovation in this paper serves a good starting point in the search for equilibrium construction methods of all-pay contests with arbitrary number of prizes and active participants.
- Dissertation Note:
- Ph.D. Pennsylvania State University 2020.
- Technical Details:
- The full text of the dissertation is available as an Adobe Acrobat .pdf file ; Adobe Acrobat Reader required to view the file.
View MARC record | catkey: 31804075