Who will watch the watchers? On optimal monitoring networks

Abstract

Consider a moral hazard scenario in which a monitor must detect deviations to provide proper incentives to attain an efficient outcome. What if the monitor himself were to deviate after being bribed by his monitored subject? In this paper, I model a multi-agent public-good provision game in which each player prefers shirking to working in the absence of exogenous enforcement and can bribe those assigned to monitor him. I find that an optimally designed monitoring structure must have players arranged in a core-periphery network, with a small group of heavily monitored players who monitor all the others. In this network, a perfect Bayesian equilibrium (PBE) is supported in which shirking is prevented, bribery among the players ceases and total monitoring costs are minimized. Furthermore, the efficiency of this core-periphery structure is robust under different cost assumptions and monitoring schedules. In addition, the spread of bribery can be prevented as long as all the “core” players do not collude.