Default prediction of small and medium-sized enterprises with industry effects

Abstract

Financial statements vary between industries. Therefore, economic intuition suggests that industry effects should be an important component in bankruptcy prediction; however, in previous academic literature on default prediction, not much attention has been paid to these effects. In this study a number of questions concerning credit risk modelling of small and medium-sized enterprises (SMEs) and industry are answered. The results indicate that financial ratios of SMEs differ between industries. Industry, measured by the weight-of-evidence (WoE), is a significant variable in default prediction of SMEs. Rating models designed for a specific industry (trade, service, or manufacturing) contain other variables and/or variables have other weights than for a generic PD model. Industry specific default prediction models perform slightly better than the generic model with industry effects, except for the service industry, where the industry specific model considerably outperforms the generic PD model.