Monitoring Corruption and Overcoming the Collective Action Problem

Presented at French Evolutionary Society for Human Sciences Annual Conference (with Torben Behmer).

Recent corruption scholarship emphasizes that monitoring consistent with the principal-agent model mostly fails under self-reinforcing corruption with non-benevolent principals. Yet incremental monitoring sometimes works in these challenging settings of systemic corruption. We explain why using a collective-action model with multiple equilibria and an assurance-game structure. It captures the limits of top-down supervision under systemic corruption while specifying the citizen-level payoffs, beliefs, and strategic reasoning that can sustain or shift equilibria. Instead of treating monitoring payoffs as monolithic, we separate them into process-level collective benefits and actor-level private benefits. Without the need to eliminate the high-corruption equilibrium, our model shows that both types of monitoring benefits can lower the threshold for taking costly action against corruption. Nevertheless, private benefits are more effective per unit because collective benefits also increase the payoff to free-riding. Overall, our model enables collective-action accounts of corruption to better explain incremental change and heterogeneous outcomes, and our monitoring benefits distinction provides a policy design heuristic. [Draft Paper]