Voting by Hands Promotes Institutionalised Monitoring in Indirect Reciprocity

TL;DR

This paper models how voting for monitors that overlook certain defections can sustain cooperation in social dilemmas, highlighting the effectiveness of different monitoring strategies and voting mechanisms.

Contribution

It introduces a co-evolutionary model showing that voting for monitors, especially STERN types, can stabilize cooperation more effectively than simple reputation ratings.

Findings

Overlooking defection against ill-reputed players is crucial.

Voting for monitors stabilizes cooperation more than direct strategy change.

MILD monitors achieve higher cooperation levels with lower monitoring costs.

Abstract

Indirect reciprocity based on reputation is a leading mechanism driving human cooperation, where monitoring of behaviour and sharing reputation-related information are crucial. Because collecting information is costly, a tragedy of the commons can arise, with some individuals free-riding on information supplied by others. This can be overcome by organising monitors that aggregate information, supported by fees from their information users. We analyse a co-evolutionary model of individuals playing a social dilemma game and monitors watching them; monitors provide information and players vote for a more beneficial monitor. We find that (1) monitors that simply rate defection badly cannot stabilise cooperation---they have to overlook defection against ill-reputed players; (2) such overlooking monitors can stabilise cooperation if players vote for monitors rather than to change their own strategy; (3) STERN monitors, who rate cooperation with ill-reputed players badly, stabilise cooperation more easily than MILD monitors, who do not do so; (4) a STERN monitor wins if it competes with a MILD monitor; and (5) STERN monitors require a high level of surveillance and achieve only lower levels of cooperation, whereas MILD monitors achieve higher levels of cooperation with loose and thus lower cost monitoring.