In this paper, I study how to design incentive structures in asymmetric coordination games to enhance coordination success. Players decide whether to invest in a project. Investing is costly, and the cost is heterogeneous across players. The project succeeds and generates benefits only if the number of investors exceeds a certain threshold. Adopting a 'global game' approach, I find that there is a unique optimal allocation of the benefits to incentivize players to invest. Specifically, players with higher investment costs receive a larger share of the benefits, and the disparity in the benefit allocation is smaller than the disparity in costs. Two lab experiments were conducted with three treatments varied in the benefit allocation. The results support the theoretical predictions when cost types were assigned based on participants' performance in an intelligence test. However, when cost types were randomly assigned, participants were more likely to invest under the equal allocation, which was driven by fairness considerations.

Experiments were conducted to investigate the effect of the decision mode on reciprocation and of being positively surprised by trust in a one-shot trust game in a first experiment. Trustees, some of whom were exogenously induced to be surprised by trust, decided directly whether to reciprocate after observing their partner’s choice or according to a contingent response plan made before observing their partner’s choice. In contrast to the prediction of a psychological model of surprise, trustees were more likely to reciprocate under contingent decision making than under direct decision making, and were not more likely to reciprocate, the more surprised they were by trust. A second experiment was then ran in which trust choices were the outcome of a lottery, with the purpose of eliminating the trustors’ intentions and the incentive to reciprocate trust. This made the hot-cold gap disappear, suggesting that the tendency to committ to reciprocation of trust is an important behavioral force.

A lab experiment was conducted to investigate whether people are more inclined to reciprocate trust when their partner faces a larger opportunity cost of trusting. Participants played a series of binary trust games with varying sizes of opportunity cost. Trustees were more likely to reciprocate when the opportunity cost was higher. Additionally, trustors’ beliefs about the reciprocation rate also increased with the opportunity cost. Two psychological mechanisms could explain this effect: Intention—a higher opportunity cost makes the trusting action seem kinder, and Guilt Aversion—trustors may expect to receive more when forgoing a larger opportunity, leading trustees to reciprocate more to avoid disappointing them. In a treatment where trust choices were determined by a lottery, neither reciprocation nor beliefs responded to the opportunity cost, suggesting intention as a key mechanism. In another treatment, where trustees made choices conditional on each possible belief of the matched trustor, their reciprocation was responsive neither to opportunity cost nor to belief, providing no evidence for guilt aversion.