We use a Logit Model such as the one developed by Scarpel & Milioni (2001), designed to forecast corporations bankruptcy, together with an Integer Programming Model developed by Gehrlein & Wagner (1997). We aim at supporting decisions of credit concession considering the corporation's solvency probability estimate and minimizing the sum of opportunity and failure to pay costs. As we show, the conjoint utilization of both models eliminates limitations found in each of them, when used in isolation.
credit scoring; Logit; integer programming
