A Structural form Default Prediction Model for SMEs, Evidence from the Dutch Market

Abstract

The objective of this research is to develop a structural form probability of default model for small and medium-sized enterprises, dealing with the methodological issues which arise in the modelling of small commercial loan portfolios. Other motivations are to provide an extensive overview of the characteristics of SMEs, and to provide a list of characteristics for an SME PD model, e.g. time and cost efficiency, broad applicability, limited data requirements, and powerful in predicting default. The structural form model is developed and tested on a unique dataset of private firm's bank loans of a Dutch bank. The results are promising; the model output differs significantly between defaulted and non-defaulted firms. The structural form model can be used on its own, or as an additional variable in a credit risk model. A second PD model is developed using logistic regression with a number of financial ratios, including the structural form measure. This variable is significant in default prediction of SMEs and has some additional predictive power, next to the popular financial ratios. Overall, the results indicate that the structural form model is a good indicator for default of SMEs.