Publications

Cooperation in Infinitely Repeated Games of Strategic Complements and Substitutes (joint with Wieland Müller and Sigrid Suetens) Journal of  Economic Behavior and Organization, vol . 188, 1191-1205, 2021 

We report the results of an experiment conducted to study the effect of strategic environment on cooperation in indefinitely repeated two-player games. We find that choices in the first rounds of the repeated games are significantly more cooperative under strategic substitutes than under strategic complements and that players are more likely to choose joint-payoff maximizing choices in the former than in the latter case. We argue that this difference is driven by strategic risk: it is less risky to cooperate under substitutes than under complements. We find that choices are on average significantly higher in the first rounds of the repeated games with strategic substitutes, while the treatment effect disappears if one focuses on all rounds of the repeated games. We relate this finding to the interaction of two countervailing forces: strategic risk and best-reply dynamics. On the one hand, subjects initiate cooperation less under strategic complements by choosing lower choices due to the higher strategic risk of cooperation. On the other hand, in all but the first rounds best-reply dynamics come into the picture: players are more inclined to follow cooperative moves of the partner under complements, offsetting the treatment effect observed in the first rounds.

Working Papers

Over-workers and Drop-Outs in Competitions: Contests with Expectations-Based Loss-Averse Agents (Previously circulated with the title: Effort Provision and Optimal Prize Structure in Contests with Loss-Averse Players)

Competition is often presumed to boost performance, however, empirical evidence shows bi-modal effort provision: over-workers and drop-outs. To address this discrepancy and its potential implications, I analyze a multiple-prize contest with expectation-based loss-averse contestants; following the evidence that individuals evaluate outcomes both in absolute terms and relative to their anticipated outcomes. The model's predictions align with observed behavior: high-ability players, holding high expectations, exert effort aggressively while low-ability players, holding low expectations, exert little or no effort compared to the standard theoretical predictions. This has important implications for the contest design: awarding multiple prizes can become optimal in cases where standard preferences predict the optimality of a single prize. The reason is that awarding an additional prize motivates drop-outs to exert effort, while it de-motivates over-workers by reducing competition. The optimal prize allocation hinges on the interplay between the degree of loss aversion, ability distribution, and the number of competitors. These findings suggest that muting competition may enhance total output in the presence of agents with expectation-dependent reference preferences.

Can Communication Mitigate Delays in Strategic  Investment Timing? (joint with Sander Onderstal and Joep Sonnemans) submitted  

In economic environments, decision-makers can strategically delay irreversible investments to learn from the actions of others. This creates free-riding incentives and can lead to socially suboptimal outcomes. We experimentally examine if and how communication mitigates this free-riding problem in an investment-timing game. In our baseline investment-timing game, participants choose when to invest in a nonrival project with uncertain returns, in groups of two or four players. The earliest investor of the group bears the costs of investment while everyone in the group benefits if the project reveals high returns. If more investors invest at the same time, they share the costs. In the communication treatment, subjects can freely communicate before choosing the investment time. We find that in groups of two players, communication increases cooperation and leads to significantly earlier investments. In groups of four players, however, communication significantly reduces delay only in the first period of interaction, but not in the aggregate over all periods.

Online Experimental Auctions  (joint with Anouar El Haji, Jayson L. Lusk, Sander Onderstal)

For entrepreneurs, it is essential to obtain information about potential customers’ preferences to successfully market their products. Experimental auctions have been used widely to obtain such information. While most research on experimental auctions relies on laboratory experiments, more recently, experimental auctions in the field have become increasingly important. In this paper, we explore a novel way of conducting experimental auctions in the field using online communities. In particular, we discuss some design challenges and present applications from four different online auctions using data with Veylinx, an online experimental auction platform. We test various marketing theories using our data with a representative Dutch sample.

Information Demand and Cooperation in a Trust Game:  (joint with Sigrid Suetens) 

In repeated social dilemma games, uncertainty about other players' types might help players to sustain high payoffs. In a repeated trust game or centipede game, for example, if matched with a payoff-maximizing second mover, the first mover may be better off when he does not know the second mover is payoff-maximizing than when he does know. We design an experiment to study whether first movers choose to know when they are given the chance to do so, while having the information would be detrimental in payoff terms. In the experiment, subjects play a repeated centipede game. In one treatment first movers are not informed about the type of the matched second mover and in another treatment, they have the choice to obtain information about the type of the matched second mover. We find that, in the choice treatment, almost all first movers choose to obtain the information about the second mover's type. We also find that the first-move trust rates for the first movers are significantly lower when they have the opportunity to know the type of the matched player. 

Work in Progress

How to Procure Innovation: An Experimenting Competition or a Competition to Experiment? (joint with Sander Onderstal and Joep Sonnemans)

We study procurement mechanisms for innovation, where a principal is interested in spurring innovation among a set of agents and the innovation feasibility is unknown. A subclass of such mechanisms is innovation contests where the principal chooses a prize-sharing scheme and a disclosure policy to maximize innovation. Theoretical results suggest that a public-information winner-takes-all (WTA) contest dominates any other public contest and any other WTA contest, while hidden-equal-share (HES) contests dominate a public WTA contest under certain conditions (Halac et. al., 2017). We propose an alternative mechanism, “auctioning the right to experiment”: the principal runs a second-price sealed-bid auction to sell the right to experiment alone. We show that (i) the auction can implement the optimal mechanism, (ii) for a given prize value, the auction dominates public WTA contests (i.e., it has the same success probability, while it creates higher revenue for the principal), and (iii) the auction dominates HES contests for a sufficiently high prize value. We test these theoretical predictions in a laboratory experiment in which we compare agent behavior in the public-information WTA contest, the HES contest, and the auction.

Information and Discrimination  (joint with Dianna Amasino and Yeşim Orhun)

Status in Contests: An Experimental  Study