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

Area of Science:

• Behavioral economics
• Experimental economics
• Game theory

Background:

• The gift-exchange game models reciprocal relationships between principals and agents.
• Rubin and Sheremeta (2016) investigated a bonus-version with effort distortion and shock observability.
• Their findings indicated reduced welfare with unobservable shocks and no restoration of gift-exchange when principals were informed.

Purpose of the Study:

• To replicate the findings of Rubin and Sheremeta (2016) on the gift-exchange game.
• To investigate the impact of unobservable shocks and principal's information on agent's effort and overall welfare.
• To identify potential reasons for discrepancies in findings between the original study and the replication.

Main Methods:

• Laboratory experiment using a bonus-version of the gift-exchange game.
• Two treatment variations: effort distortion (shock) and shock observability (observed vs. unobserved).
• Replication of the original study's design with data analysis focusing on welfare and gift-exchange restoration.

Main Results:

• The replication largely confirmed that unobservable shocks significantly reduce welfare.
• Unlike the original study, informing the principal about the shock size did not restore gift-exchange.
• Behavioral differences in early game rounds potentially created a hysteresis effect, explaining the divergent results.

Conclusions:

• Unobservable shocks in gift-exchange games demonstrably decrease economic welfare.
• The observability of shocks does not appear to restore gift-exchange dynamics in this replication.
• Hysteresis effects stemming from initial behavioral patterns may significantly influence long-term outcomes in repeated experimental games.