Purpose. This study aimed to understand whether a general bankruptcy prediction model for small Italian companies having any legal forms has a different predictive capacity than specific bankruptcy prediction models for those having specific legal forms. On the one hand, it focuses on cooperative companies, and on the other, joint-stock and limited-liability companies. Design/methodology/approach. A general bankruptcy prediction model and two specific bankruptcy prediction models (one for cooperative companies and one for joint-stock and limitedliability companies) were constructed and compared regarding predictive capacity. Findings. The overall accuracy levels of the general and specific models were the same, but the percentage of companies correctly predicted to be in crisis out of the total number of companies effectively in crisis (sensitivity) of the latter (in particular, referring to joint-stock and limited-liability companies) was higher than that of the former. Considering the high economic and social costs that can derive from the predictive errors of companies in crisis, specific models should be preferred to the general model. Practical and social implications. This study offers to those who may be interested in evaluating the financial health of a company (stakeholders, such as banks, suppliers, customers, etc., as well as the management and control bodies of the company) bankruptcy prediction models having a high predictive capacity differentiated according to its legal form. Originality of the study. No previous study has verified whether a general bankruptcy prediction model for companies having any legal forms has a different predictive capacity than specific bankruptcy prediction models for companies having specific legal forms. At the same time, in the Italian context, no previous study has proposed a bankruptcy prediction model for cooperative companies.
Relevance of the legal form of companies for bankruptcy prediction / Poli, Simone; Giuliani, Marco; Baccarini, Luca. - In: PICCOLA IMPRESA. - ISSN 0394-7947. - STAMPA. - 2:(2023), pp. 81-102. [10.14596/pisb.3787]