Overbidding and Heterogeneous Behavior in Contest Experiments

2012

This paper examines experimentally a contest in which the organizer takes a share of the prize if there is a tie. Our hypothesis is that contestants perceive this as unfair and that this perception has a non-proportional effect on the bids. The results indicate that despite the very low probability of a tie, contestants over-weight its importance and sharply decrease their bids if the organizer takes a share. We show that a fair contest is the optimal strategy for the organizer since the increase in the average bid in a "fair" contest is higher than the organizer's decrease in expected value from losing his share in a tie.

2014

We study experimentally the effects of cost structure and prize allocation rules on the performance of rent-seeking contests. Most previous studies use a lottery prize rule and linear cost, and find both overbidding relative to Nash equilibrium prediction and significant variation of efforts, which we term 'overspreading.' In a 2 × 2 design, we investigate the effects of allocating the prize by a lottery versus sharing it proportionally, and of a convex cost function versus linear costs of effort, while holding fixed the Nash equilibrium prediction for effort. We find that the share rule results in average effort closer to the Nash prediction and lower variation of effort. Combining the share rule with a convex cost function further enhances these results. Our findings indicate that a significant amount of subjects' non-equilibrium behavior in contests can be explained by features of the experimental design. These results contribute towards design guidelines for contests based on behavioral principles that take into account implementation features of a contest that may not affect the Nash equilibrium prediction.

Journal of Economic Behavior & Organization

We investigate how individual risk preferences affect the likelihood of selecting the more able contestant within a two-player Tullock contest. Our theoretical model yields two main predictions: First, an increase in the risk aversion of a player worsens her odds unless she already has a sufficiently large advantage. Second, if the prize money is sufficiently large, a less able but less risk averse contestant can achieve an equal or even higher probability of winning than a more able but more risk averse opponent. In a laboratory experiment we confirm both, the non-monotonic impact and the compensating effect of risk aversion on winning probabilities. Our results suggest a novel explanation for the gender gap and the optimality of limited monetary incentives in selection contests.

SSRN Electronic Journal, 2006

We study tournaments with many ex-ante asymmetric contestants, whose valuations for the prize are independently distributed. First, we characterize the equilibria in monotone strategies, second, we provide sufficient conditions for the equilibrium uniqueness and, finally, we reconcile the experimental evidence documenting the 'workaholic' behavior in contests with the related theory by introducing heterogeneity among participants. It is a 'weak' participant that might become a 'workaholic' in an equilibrium, that is, his effort density might increase at the highest valuation-weak, either because he is more risk averse or because his rivals consider that it is very unlikely that he has a high value for the prize. In contrast, effort densities are always decreasing in case of symmetry with identically distributed values for the prize and identical attitudes towards risk in case of CARA, as well as in contests with only two participants. Moreover, we show that for low valuations more risk averse agents are less likely to exert low effort than their 'strong' rivals, while those with dominated distribution of the prize valuation are more likely to do so. An explicit solution for the uniform distribution case with contestant-specific support is provided as well.

SSRN Electronic Journal, 2000

In modern firms the use of contests as an incentive device is ubiquitous. Nonetheless, recent experimental research shows that in the laboratory subjects routinely make suboptimal decisions in contests even to the extent of making negative returns. The purpose of this study is to investigate if changing how agents are endowed with resources can increase the efficiency in contests. To this end, we conduct a laboratory experiment in which subjects are asked to allot costly resources (bids) in an effort to attain an award (prize). In line with other laboratory studies of contests, our results show that subjects overbid relative to theoretical predictions and incur substantial losses as a result. Making subjects earn their initial resource endowments mitigates the amount of overbidding and thus increases overall efficiency. Overbidding is also linked to gender with women bidding higher than men and having lower average earnings. Other demographic information such as religiosity and individual preferences towards winning and risk also contribute to excessive bidding.

SSRN Electronic Journal, 2000

SSRN Electronic Journal, 2000

This study examines behavior of subjects in simultaneous and sequential multi-battle contests. In simultaneous contests, subjects make positive bids in each battle 80% of the time and bids fall within the predicted boundaries. However, 35% of the time subjects make positive bids in only two, instead of all three, battles and they significantly overuse moderately high bids. In sequential contests, theory predicts sizable bids in the first battle and no bids in the subsequent battles. Contrary to this prediction, subjects significantly underbid in the first battle and overbid in subsequent battles. Consequently, instead of always ending in the second battle, contest proceeds to the third battle 38% of the time. Finally, in both simultaneous and sequential settings, subjects make higher aggregate bids than predicted resulting in negative expected payoffs.

Journal of Mathematical Psychology, 2005

We report the results of an experiment on two-stage contests with budget-constrained agents competing to win an exogenously determined prize. We study a class of two-stage contests where in stage 1 agents first compete within their own groups by expending resources, and in stage 2 the winners of each group compete with one another to win the prize by expending additional resources subject to the budget constraint. Winners in each stage are determined by Tullock's proportional contest success function. We present the subgame perfect equilibrium solution for this model, derive predictions for our experiment, and then test them experimentally. The equilibrium model is flatly rejected mostly due to over expenditure in stage 1. A descriptive model that extends the equilibrium solution by allowing for 1) non-pecuniary utility of winning and 2) misperception of the probability of winning better accounts for some but not all of the behavioral regularities. We then turn to an adaptive learning model that accounts for several features of the dynamics of play but still significantly under-predicts the stage 1 expenditures.

The Economic Journal, 2008

Applying behavioral considerations proposed by psychologists, it is shown that distortion of probabilities is yet another possible reason for rent under-dissipation in contests with relatively small number of participants. Such distortion may also result, however, in overdissipation of the contested rent. Focusing on contests with n homogeneous contestants and the most commonly studied contest success function that depends on a single parameter r, our main results clarify under what circumstances, namely, combinations of n and r, (i) rents are more under-dissipated relative to the standard situation where probabilities are not distorted (ii) rents are under-dissipated, yet less intensely relative to the standard situation where probabilities are not distorted (iii) rents are over-dissipated and (iv) the contest does not possess a symmetric interior equilibrium in pure strategies. Our analysis points to a new possible source of reduced or increased waste in contests. In the case where r=1 and assuming distorted probabilities as in Cumulative Prospect Theory, it is shown that, when the number of contestants exceeds (is smaller than) 14, rents are over (under)-dissipated.

We conduct an experimental analysis of a best-of-three contest. Intermediate prizes lead to higher efforts, while increasing the role of luck (as opposed to effort) leads to lower efforts. Both intermediate prizes and luck reduce the probability of contest ending in two rounds. The patterns of players' efforts and the probability that a contest ends in two rounds are consistent with 'strategic momentum', i.e., momentum generated due to strategic incentives inherent in the contest. We do not find evidence for 'psychological momentum', i.e., momentum which emerges when winning affects players' confidence. Similar to previous studies of contests, we find significantly higher efforts than predicted and strong heterogeneity in effort between subjects.