Departures from fully rational behavior due to cognitive limitations or psychological phenomena are typically referred to by economists as boundedly rational behavior. In this dissertation, I study how bounded rationality impacts cooperation in repeated games, herding behavior and bidding in online auctions. The methods I use include theoretical modeling and empirical analysis of data collected in controlled laboratory experiments as well as data from the field. This research contributes to the understanding of the consequences of bounded rationality in strategic interactions.
In the first chapter, I investigate whether cooperation in an indefinitely repeated prisoner's dilemma is sensitive to the complexity of cooperative strategies. I use an experimental design which allows manipulations of the complexity of these strategies by making either the cooperate action or the defect action state-dependent. Subjects are found to be less likely to use a cooperative strategy and more likely to use a simpler selfish strategy when the complexity of cooperative strategies is increased. The robustness of this effect is supported by the finding that cooperation falls even when the defect action is made state-dependent, which increases the complexity of punishment-enforced cooperative strategies. A link between subjects' ACT scores and the likelihood of cooperating is found, indicating that greater cognitive ability makes subjects more likely to use complex strategies. Behavior when subjects play multiple simultaneous games is compared to their behavior in isolated single games, providing evidence that the additional cognitive cost of playing multiple games also limits cooperation within this environment.
Despite numerous applications, the importance of capacity constraints has so far received little attention in the literature on herding behavior. I attempt to address this issue in my second chapter by constructing a simple model of herding with capacity constraints and studying behavior in this environment experimentally. The model predicts and experimental results confirm that capacity constraints can attenuate herding, with the size of the effect dependent on the penalty of choosing an option after its capacity has been reached. For subjects earlier in a sequence of choices, behavior without a capacity constraint does not differ markedly from that observed in comparable experiments despite the fact that preceding choices are made by computers with fixed, commonly known choice rules rather than other humans. For subjects later in a sequence of choices, I find evidence that whether they respond rationally to the capacity constraint is dependent on factors such as the depth-of-reasoning involved in the fully rational equilibrium and the subject's cognitive ability.
The third chapter of this dissertation is a study of data on bidding behavior in eBay auctions of Amazon.com gift certificates. I find that 41.1% of winning prices in these auctions exceed the face value, which is an observable upper bound for rational bidding because Amazon.com sells certificates at face value. Alternative interpretations are explored, but bidding fever seems to be the most plausible explanation for the observed behavior.