Checking and temporal embeddedness in indefinitely and finitely repeated asymmetric trust games

Abstract

This thesis presents both theoretical and experimental research on the influence of checking in indefinitely and finitely repeated asymmetric trust games of incomplete information. To study information asymmetry we changed the binary trust game, which is the simplest game that reflects the trust problem, into an asymmetric trust game. In this game the trustor cannot directly observe the trustee’s response. The trustor can either earn a high or a low pay-off. The trustee's response determines the probability with which the trustor earns a high or a low pay-off. The trustor can form an expectation about the trustee's decision based on the pay-off he observes, but he never obtains certainty about this. Checking is a control option the trustor can decide to use after both the trustor and the trustee made their decisions, but before they learn the pay-offs they earned in this period. In the situation without the checking option the trustor only receives information about his own pay-off realisation (either a high or a low pay-off). When he checks he will also learn the trustee’s response and therefore the information asymmetry is resolved. ‘Checking’ differs in this respect from the often in the literature found ‘monitoring’ that allows the trustor to control the trustee's behaviour before the trustee makes his decision. Our theoretical models show that co-operation is less easily established and sustained in indefinitely and finitely repeated asymmetric trust games compared to games without asymmetric information. The experimental results show that the checking option has a positive effect on the level of co-operation in repeated asymmetric trust games. In line with earlier experimental results we found no difference in the level of co-operation between indefinitely and finitely repeated asymmetric trust games when the players are inexperienced. In line with theoretical predictions we did find that trustors play different strategies in both types of games. Theory provides some clear predictions on how selfish and honest trustees and trustful and non-trustful trustors should behave in repeated trust games of incomplete information. On the basis of an adjusted isolated encounter trust game treatment we were able to distinguish these trustor and trustee types in our data analysis. We found that the majority of the experimental results on type indicators support the theoretical predictions.