Socioeconomic Distance and the Selectivity in Punishment and Reward[Sanction et récompense par des tiers selon la distance socioéconomique]

Do we punish distant others more harshly and reward close others more readily? We study these decisions in an iterated online Die-under-the-Cup (DUTC) task with n = 720 participants randomly assigned to fixed roles: Die-rollers (P1) and Observers (P2). Over 24 rounds, each P2 is paired with two 12-round blocks of P1s-one socially close and one socially distant-under one of three treatments: Punishment, Reward, or Mixed. P2 privately observes both P1's true and reported values, and can apply -€2 penalty or a +€2 reward that does not affect P2's own payoff. Social closeness is modelled through socioeconomic status (SES) using (1) an objective measure based on average income in the participant's locality of residence and (2) a subjective measure based on self-reported income relative to the locality's average. We construct an observer-level proxy for suspicious reporting using excess self-serving reports relative to fair odds within the 12 reports seen by P2 per SES and test whether social closeness moderates enforcement. Findings show that punishment increases with the suspicion proxy but is weaker for close counterparts, while rewards are granted more often for non-suspicious reports. Third-party penalties and rewards are selective across SES: at comparable suspicion levels, low-SES counterparts are punished more than medium/high-SES counterparts, yet they also receive more rewards for non-suspicious reports. Moreover, we observe that the objective measure of closeness is linked to overall leniency toward close counterparts, whereas the subjective measure shifts reactions to the suspicion proxy: weaker for close counterparts and stronger for distant ones. Overall, our study reveals that counterpart SES shapes responses to reported outcomes.