Predicting Construction Contractor Default with Barrier Option Model

Abstract

This is the first study to apply the barrier option model to predict defaults of construction contractors and to assert that the path-dependent characteristic of the model is very suitable for describing the behavior of contractor default. Different from existing contractor-default prediction models, this research uses a much larger contractor sample in empirical analyses to alleviate sample-selection biases, and employs a Receiver Operating Characteristics (ROC) curve to assess the model performance. Empirical results of this study show that the proposed model outperforms traditional financial ratio models in differentiating the risk of defaulted and nondefaulted construction contractors. Additionally, the barrier option model has markedly better discriminatory power than when applied to non–construction-related industries. The results of this paper support the postulation that the barrier option model has significant advantages for the construction industry.