Sacrifice, Benefit, and Reciprocity: Evidence from a Trust Game

Social preferences have generated much interest in recent economic literature. While reciprocity has been closely examined by several economists using gift exchange and other games, much of their focus is on the effect of intentions and motivations. Instead, I focus on two strictly outcome-based factors in the context of a trust game: sacrifice made by the giver and benefit received by the recipient. I attempt to isolate these influences by varying the multiplier across variations of the trust game to create comparison groups that differ in one of these factors, but not both. Unable to control for all expected sources of bias, I use two different specifications that I suspect to have negative and positive signs of bias. I find that both sacrifice and benefit seem to matter, meaning that the recipient of a gift seems to account for both the cost and the value of the gift. Specifically, I estimate $1 of additional sacrifice to increase reciprocal giving by $0.35 under the first specification and $0.57 under the second, and that $1 of additional benefit from a gift increases reciprocal giving by $0.28 and $0.34 under the first and second specification, respectively. Robustness checks incorporating other data such as order effects, session effects, and demographic and personality type controls raise some room for doubt, but do not provide a strong refutation of my qualitative results.