Contract design with limited commitment

We consider the problem of a principal who wishes to contract with a privately informed agent and is not able to commit to not renegotiating any mechanism. That is, we allow the principal, after observing the outcome of a mechanism to renegotiate the resulting contract without cost by proposing a new mechanism any number of times. We provide a general characterization of renegotiation-proof states of such a renegotiation. The proposed solution concept provides an effective and easy-to-use tool to analyze contracting problems with limited commitment. We apply the solution concept to a setting with a continuous type space, private values and non-linear contracts. We find that the optimal contracts for the principal are pooling and satisfy a "no-distortionat-the-bottom" property.