A structural credit risk model based on purchase order information

Abstract

This study proposes a credit risk model based on purchase order (PO) information, which is called a gPO-based structural model,hand performs an empirical analysis on credit risk assessment using real PO samples. A time-series model of PO transitions is introduced and the asset value of the borrower firm is obtained using the PO time-series model. Then, we employ a structural framework in which default occurs when the asset value falls below the debt amount, in order to estimate the default probability of the borrower firm. The PO-based structural model enables us to capture borrower firms' precise business conditions on a real-time basis, which is not the case when using only financial statements. With real PO samples provided by some sample firms, we empirically show the effectiveness of our model in estimating default probabilities of the sample firms. One of the advantages of our model is its ability to obtain default probabilities reflecting borrower firms' business conditions, such as trends in PO volumes and credit quality of buyers.