Default Prediction Modeling for Small Enterprises: Evidence from Small Manufacturing Firms in Northern and Central Italy

Previous empirical research shows the effectiveness of using sets of economic-financial ratios for company default prediction statistical modeling. However, such research rarely focuses on small enterprises (SEs) as specific units of analysis. In Italy, SEs account for more than 98% of all firms and employ over 70% of the total workforce. The results of our statistical analyses, conducted on a sample of small manufacturing firms in Northern and Central Italy, show that both discriminant analysis and logistic regression are effective tools for designing SEs’ default prediction models based on economic-financial ratios. In addition, the proposed models gain in prediction accuracy when they are specifically constructed for separate business sectors and separate company size groups. Without denying the value of jointly using quantitative and qualitative variables to measure a firm’s rating, this paper confirms that: i) SEs’ credit rating models must adequately weight sets of appropriately selected financial and economic ratios; ii) SEs’ credit rating should be modeled separately from that of large and medium-sized firms; and iii) SEs’ credit rating models need to be specifically designed so as to take into account the diverse economic and financial profiles of firms in different industries and at different stages of growth.