This paper investigates the determinants of the arbitration taking place after a corporate defaults. Two ways of resolving financial distress are conceivable: either the creditors privately renegotiate with the debtor, or a formal bankruptcy procedure is triggered. This arbitration depends on the legal context and, more specifically, on the national bankruptcy code. We use original data coming from the recovery units of five French commercial banks. Our sample gathers 735 credit lines allocated to 386 distressed companies.We test four hypotheses. Hypothesis H1 proposes that bargaining power imbalances may have more impact on the arbitration than simple coordination failures, especially under court-administered legislation. Hypothesis H2 suggests that, to initiate the process of renegotiation, a bank needs information on the project’s profitability and on the managers’ reliability. To reach an agreement, the bank must ensure that both the conditions have been met. Hypothesis H3 predicts that the likelihood of renegotiation increases with the bank’s financial involvement. Hypothesis H4 focuses on the level of collateralization: when the bank has inclination for liquidation, collaterals may increase the occurrence of bankruptcy, provided the law facilitates such liquidation and preserves the bank’s priority.For testing the hypotheses H1 to H4, we use sequential LOGIT modeling to split between the variables explaining the decision to engage (or not) renegotiation and the variables explaining the success (or the failure) of renegotiation. Regarding H1, we find that even a court-administered procedure may not be dissuasive provided that the bank’s bargaining power is strong enough. Regarding H2, we show that the profitability of the project and the reliability of the managers are two essential conditions for avoiding bankruptcy. However, it takes some time to discover them. Regarding H3, our estimates show that, when the lending is bigger and/or when the debt contract is longer, then the chances of renegotiation are higher, but this does not predict such renegotiation shall be successful. Last, regarding H4, we do not find any evidence that the level of collateralization influences the arbitration between renegotiation and bankruptcy.