Is reciprocity really outcome-based? A second look at gift-exchange with random shocks

Brent J Davis; Rudolf Kerschbamer; Regine Oexl

doi:10.1007/s40881-017-0041-2

Is reciprocity really outcome-based? A second look at gift-exchange with random shocks

J Econ Sci Assoc. 2017;3(2):149-160. doi: 10.1007/s40881-017-0041-2. Epub 2017 Nov 20.

Authors

Brent J Davis¹, Rudolf Kerschbamer¹, Regine Oexl¹

Affiliation

¹ University of Innsbruck, Innsbruck, Austria.

Abstract

By means of a laboratory experiment, Rubin and Sheremeta (Manag Sci 62(4):985-999, 2016), study a bonus-version of the gift-exchange game, including two treatment variations. First they vary whether the effort provided by the agent directly translates into output for the principal, or whether it is distorted by a shock. Second, for the condition with a shock they vary whether the shock is observed by the principal, or not. The authors' main findings are that (1) the introduction of an unobservable shock significantly reduces welfare; and (2) informing the principal about the size of the shock does not restore gift-exchange. In a replication study we largely reproduce finding (1), but we fail to confirm finding (2). Our data suggests that small behavioral differences in the initial rounds lead to a hysteresis effect that is responsible for the differences in results across studies.

Keywords: Gift-exchange; Incomplete contracts; Laboratory experiment; Outcome-based reciprocity; Principal agent model; Random shocks; Replication study.