A default prediction model for Italian SMEs: the relevance of the capital structure

Abstract

Small and medium enterprises (SMEs) play a fundamental role in the economy of many countries, including Italy. The simplicity of the financial structure of Italian SMEs, relying mostly on bank loans, makes them riskier than their larger counterparts and this induces banks to develop default prediction models specifically addressing the management of SMEs’ credit relationship. Building on this framework, this article aims to analyse the determinants of the default probability of a sample of 9208 Italian limited liabilities SMEs in a time frame of 3 years, over the period 2006 to 2010. Specifically, this article adopts a logistic regression model estimated on a database provided by the Centrale Rischi Finanziari (CRIF), an Italian credit rating agency. The results show that, among the observable financial and economic characteristics of the firms, the capital structure (both in terms of internal and external funds and in terms of the source of external financing) and interest expenses are more relevant than economic variables as determinants of SMEs’ default.

Keywords:

• default prediction
• capital structure
• small and medium enterprises
• logistic regression

JEL Classification:

• G21
• G32

Acknowledgements

First of all, we are extremely grateful to Renato Maino for his valuable supervision and useful comments on a previous version of this paper. We would also like to thank Marco Muscettola for his support in data set preparation.

Notes

¹ However, since we adopt a 3-year window, with the available data, we cannot distinguish between the effects of these characteristics before and after the financial crisis.

² An exception is provided by Ciampi and Gordini (2013), adopting a lag time of 4 years.

³ For a comprehensive review of the literature, see, e.g. Ravi Kumar and Ravi (2007).

⁴ As far as the use of different statistical methods is concerned, the relationship between generalization and specificity in these models is another widely debated theme. Several authors discuss the need to develop industry-specific models rather than general ones. In most cases, industry-specific research (Altman et al., 1994; Sironi, 2003) is more accurate than research using general models. This is probably due to greater homogeneity of the financial indicators within specific industries.

⁵ In our database, the definition of SMEs refers to 2009.

⁶ The literature (see, for instance, Gai, 2008) provides several definitions of the default event. This choice is not neutral since it might influence the results of the empirical analysis, other than the number of firms classified as default firms. These definitions include legal ‘bankruptcy’ or ‘liquidation’, which are more restrictive in that the moment of legal failure does not reflect the real failure event.

⁷ This method maximizes the variance of the squared loadings of each factor on all the variables and allows the differentiation between the original variables and the factors extracted. In other words, the varimax rotation method minimizes the number of original variables that have a high weight on a specific factor and produces orthogonal factors, that is, ones that are not related to each other.

⁸ Another selection algorithm, the backward selection method, starting from the full model (including all the explanatory variables), iteratively eliminates the explanatory variables that are not significant. Furthermore, the application of the backward methodology provides the same results, in terms of both coefficients and significance, as well as in terms of the performance of the model. These results are available on request.