Despite their popularity, monitoring methods using only financial statements do not provide a real-time feed of changes in business conditions. To solve this problem, our article proposes a credit risk model based on purchase order (PO) information, also called “structural model based on purchase orders”. It performs empirical analysis of credit risk assessment using actual samples of purchase orders. A time series model of purchase order volumes is introduced and the asset value of the borrowing company is obtained using the purchase order time series model. To estimate the default probability of the borrowing firm, we then use a structural framework where default occurs when the value of the asset falls below the amount of the debt. Finally, we empirically show the effectiveness of the model in estimating the default probabilities of the companies in the sample.