Even though Altman’s Z-Score models are considered common tools for evaluating the financial health of companies, we present several reasons to demonstrate that the use of these models may be obsolete nowadays. In this paper, we discuss the urgent need for an up-to-date bankruptcy prediction model and investigate the extension of the Z”-Score models in predicting the health of Italian private companies by utilizing multiple discriminant analysis (MDA) with a pair-matched sample of 88 failed and healthy firms from 2007 to 2015. Firstly, based on prior literature, we formulate and test a set of hypotheses to identify new potential variables that could have discriminant power and find that two of them, namely, trade credit demand and losses caused by trade credit extension, represent powerful discriminators. Next, we assess the classification performance of the original and the revised Z”-Score models in predicting failure, with the aim of examining the models’ applicability in the Italian context by means of an accuracy matrix, highlighting that they cannot be generalized to countries other than the US. After that, we measure the overall predictive power of the proposed model (78.41%) and evidence that it outperforms the original and the revised Z”-Score models (62.50% and 68.18%, respectively). Thus, in this paper, we propose a new five-factor MDA model that surpasses the Z”-Score models and is fairly accurate in predicting the health of privately held Italian companies one year prior to failure. However, the proposed model is not without limitations due to the decline in performance two years prior to bankruptcy and to a moderate Type I error. Furthermore, this work contributes to the emerging body of literature in accounting and finance by adjusting and extending the Z-Score models developed by Altman. The proposed model could assist financial statement users, such as potential investors and other practitioners, as well as empirical accounting researchers, in identifying potentially bankrupt firms. Compared to works that have focused on verifying the discriminant capability of existing bankruptcy prediction models, this paper lays the ground for future research and sheds light on the importance of deeply investigating the topic of failure prediction by publishing further studies, which will improve the classification accuracy of the models by adding alternative variables that are useful in the Italian context and/or utilizing alternative modelling techniques.
KEYWORDS: bankruptcy prediction models, trade credit, discriminant analysis.
D'Amico, E., Dello Strologo, A. & Matozza, F. (2020). Predicting corporate bankruptcy in Italy by adopting Z”-Scores: an assessment and an extension, RIREA, n.2, pp. 216-232