Credit scoring method using estimated forward financial statements based on purchase order information

Abstract

In this article, we propose a credit scoring method using purchase order information on target borrower firms. First, we introduce a time-series model which captures purchase order volume transitions of a target firm. Then, we execute credit scoring based on estimated financial statements reflecting expected values of future purchase orders obtained from the purchase order models. We demonstrate the applicability of our method to practical credit risk monitoring with a case study. One of the advantages of our method is its abilities to capture changes of credit risk timely reflecting the firms' business conditions.