Collusion (defined as side contracting between agents) and renegotiation (defined as side contracting between a principal and an agent) coexist in many real-world situations. For example, a union can be viewed as a form of collusion among workers, and firms usually renegotiate with their workers about their compensations or benefits when the firms face financial problems. While there are many studies considering collusion or renegotiation separately, very few take both into account. By working on a model with one principal, one agent, and one supervisor who observes the agent's private information (effort or type), we find various interesting interactions between collusion and renegotiation that have different efficiency implications, depending on the exogenous features of the collusion technologies. More specifically, although in most contracting situations the presence of either collusion or renegotiation is known to be costly to the principal, we find that the principal can never be worse off with both collusion and renegotiation than with neither of them (the second best). We have verified that this result holds under two major types of collusion under moral hazard (when the agent's effort is not observable to the principal), namely mutual insurance and effort coordination; we expect it to hold as well with report manipulation under adverse selection (when agent's marginal cost is not observable to the principal). Specifically, under moral hazard the first best can be achieved when the mutual insurance is moderately cooperative, whereas the second best can be achieved when the mutual insurance is most cooperative or least cooperative. This suggests an interior optimal strength of collusion under moral hazard. In an abstract sense, the research suggests that collusion and renegotiation have to be considered in tandem in organizational and contract design.