Heterogeneity in Dynamic Contests: Theory and Experiments

The proposed research project focuses on studying theoretically and in laboratory experiments contests characterized by either a dynamic structure, heterogeneity of the contestants, or both. Contests are competitions in which agents invest resources in order to win a prize, each forfeiting the investment made in any case. Our results have potentially important implications in many domains. For instance, in personnel economics applications the maximization of the workers effort is typically the aim. On the other hand, if efforts or investments in the contest are seen as wasteful (as e.g. in rent seeking contests) optimal contest design can help to reduce society`s loss. The project consists of 4 parts, which differ with regard to the contest structure, number of agents and the prize structure. In the first part we analyze theoretically and in lab experiments two-stage elimination contests with 4 agents, where the winners from first-stage pairwise competitions proceed to the second stage and the winner of the second-stage receives a prize. Winning probabilities in each interaction are given by a Tullock lottery contest success function. Our focus is on the effect of heterogeneity on overall effort provision of agents. In addition we plan to analyze the effect of different seeding variants in the first stage of the contest on the behavior of agents. The second contribution analyzes the behavior of homogenous and heterogenous agents in a simultaneous or sequential two-player contest with a proportional prize. Theory predicts that symmetric agents behave exactly the same way in the simultaneous and in the sequential contest. When agents are heterogenous a first mover with low costs has an incentive to overinvest thereby crowding out the investment of the follower. By contrast, a first mover with high costs should invest less in the theoretical benchmark to induce the follower to invest less too which would lead to a pareto improvement. We test those theoretical predictions in the lab. Our main research interest is in the effect of heterogeneity and sequentiality on total expenditures. In addition we plan to address the question of collusion, especially which structure facilitates collusion. In the third contribution we analyze again a two-player contest with a sequential game structure but now we endogenize the move order. Agents have to decide whether they want to invest in the rst period of the game or wait and invest in the second period. This is especially interesting in the lab, namely whether participants are able to coordinate on a certain move order. In the fourth contribution we extend the setting from contribution 2 to an endogenously determined prize which is a function of chosen effort levels.

Loosely defined, contests are competitions in which agents invest resources in order to win a prize, each forfeiting the investment made in any case. Situations fitting this definition are countless, ranging from naturally arising contests (e.g. violent military conflicts between countries, struggles between animals for a food resource) to deliberately designed ones (e.g. promotion contests, procurement auctions, sports tournaments, etc.). This project has contributed to the literature on contests and strategic interactions in three important ways: i) it has added to the literature on dynamic contests, where some agents can observe the outcome of a previous stage or the decision of another agent before making a decision; ii) it has added to the literature on heterogeneous contests, where different types of individuals interact in a given context; and iii) it has added to our understanding of the impact of behavioral aspects and in particular social preferences in strategic interactions. In this summary I cover three central contributions, one to each of three domains mentioned. In a contribution to i) we investigate whether the order of moves affects behavior even in the absence of material incentives for pre-commitment. It suggests that it is an inherent advantage to move second rather than first. The reason is that first movers face strategic uncertainty, while second movers have the power to ultimately determine the outcome through their investment choices. This power seems particularly valuable in environments where behavior is heterogeneous - for instance, because players care to different degrees for the payoffs of others. In a contribution to ii) we highlight an important tension in the incentive systems of many firms: While firms want to incentivize team production by providing rewards for overall output, they also want to incentivize individual effort with relative performance-based bonuses. Rewards for overall output induce a free-rider problem, while relative performance schemes provide incentives for team members to refrain from helping others or even to sabotage their colleagues. The study examines these issues in theory and lab experiments in the presence of individual heterogeneity. The results help to understand better how to balance out individual versus team rewards and how to structure teams when employees have heterogeneous abilities. In a contribution to iii) we investigate whether decision makers bracket their choices narrowly, to facilitate complex decision problems. Evidence from a framing variation of rewards in experimental two-stage pairwise elimination contests indicates that decision makers neglect the option value of participation in future stages of the contest if the reward frame facilitates the separate consideration of stages, but not if the reward frame induces forward-looking behavior. The paper presents novel evidence on the prevalence and determinants of choice bracketing as a means to cope with strategic complexity.