Collusion and Renegotiation in Principal-Supervisor-Agent Models∗

We aim at examining the interactions between collusion and renegotiation in a principal-supervisor-agent framework, in which the supervisor and agent can collude while the principal can subsequently renegotiate. Despite common sense and the findings in the literature that collusion and renegotiation are usually costly when considered separately, we find that they play a weakly positive role when considered jointly. The proposed framework unifies and reexamines the claims of some important studies in the literature, suggesting that, in organizational design, the issues of collusion and renegotiation should be studied in tandem.