Three Essays on Contest Theory and Experiments



Journal Title

Journal ISSN

Volume Title



Contests are ubiquitous in daily life. Examples range from job applications, political elections, sports competitions to resource contentions. Because of their prevalence, it is important to understand how individuals behave in contests. By learning contestants’ behavior, contest designers can choose appropriate competition mechanisms to better motivate contestants. This dissertation focuses on the theoretical and experimental investigation of contestants’ behavior in contests. In Chapter 1, I develop theory and design a lab experiment to investigate individual behavior in a specific format of contest: rank-order tournament. As is common in this literature, in my model an individual’s output equals the effort she or he exerts plus an idiosyncratic shock. My research is novel in that I focus on the case where the idiosyncratic shock is asymmetrically distributed. This is valuable as asymmetric productivity shocks can capture certain important types of tournaments. Theoretically, under optimal tournament design, individuals are expected to exert the same effort regardless of the shape of the shock distribution. I am interested in whether this theoretical prediction holds in the lab; that is, whether subjects’ effort provisions are identical across different shock distributions. I find tournament winners demonstrate economically and statistically stronger reactions to asymmetrically distributed shocks than to shocks from a symmetric distribution. In addition, those who often win tournaments react more strongly to shocks than to those who tend to lose. These results highlight the importance of accounting for the influence of the shock distribution’s shape on a tournament participant’s effort. In Chapter 2, I provide the first experimental investigation of a non-exclusive group (NEG) contest in which individuals actively participate in multiple groups simultaneously. Unlike theory predictions, I find that, in relation to a group contest where each individual belongs only to one group, subjects in the NEG contest provide less aggregate effort and they are more likely to free-ride on their group members’ effort. The discrepancy between theoretical predictions and laboratory findings is partially due to the complex strategy space in the NEG contest. In particular, in the NEG contest, it is difficult for subjects clearly to identify their “enemy” and “ally”; therefore, it influences behavior in ways that are not easy to capture with theory. This study complements previous theory and experiments on group contests and yields novel insights for collective decisions in multi-dimensional competition environments. In Chapter 3, I extend the war of attrition by studying a three-period dynamic contest game. In my model, individuals can fight against their opponents at a certain period of the contest and can flee at any time. Waiting is costly. I focus on the role of waiting costs and show that the value of waiting costs is a key factor in determining the type of equilibrium in such dynamic contests. Specifically, as waiting costs increase, contests end earlier, battles are less likely to occur, and weaker players become more likely to flee first. A lab experiment verifies most key features of our model. However, unlike theoretical predictions, I find that as waiting costs increase, the duration of contests and the frequency of battles fail to drop significantly. Moreover, I find that in each treatment, subjects exit the contest significantly earlier than predicted. This theoretical model, combined with laboratory findings, have important policy implications relevant to reducing costly battles in dynamic contests.