The triadic design to identify trust and reciprocity: Extensions and robustness

Our paper reconsiders the triadic design proposed by Cox (2004) to identify trust and reciprocity in investment games. Specifically, we extend the design in two directions. First, we elicit expectations by a fixed-fee incentive scheme and test the coherence of them with the triadic outcomes. We expect that if trust is reported by the triadic design, investors’ expected gains should be also observed. Second, we collect information about participants’ choices by using both direct-response (as Cox) and strategy method. By the latter we are able to control reciprocity for initial inequality, which is endogenous when reciprocity is investigated. Finally, we test the existence of an emotional bias, i.e. we test if expectations mismatches induce participant to change actual choices from the planned ones.