Insolvency prediction models: an empirical study in Italy

Giacosa Elisa, Mazzoleni Alberto, Teodori Claudio, Veneziani Monica

From the global economic crisis that has also affected Italy arises the global research on that issue and the necessity of developing effective “alert signals” in the context of an existence of a crisis situation. One of the objectives of the current research is to verify effectiveness of the insolvency prediction model the most popular in the literature including recent studies (Jackson and Wood, 2013) regarding the Italian manufacturing companies, in order to modify the Altman insolvency prediction model to develop company insolvency “alert model”, that can be useful for various stakeholders and to compare the effectiveness of the re-estimated model with the original one. The current research was conducted on a sample of Italian companies (27,982 non-failed and 478 failed) whose financial statements for a period 2007- 2012 were available. It came out that using the original cut-off points, the overall error of the original model is significant. The error decreases while using the updated Altman model (only weight of the variables and cut-off points have been changed). Moreover, the adapted model (Altman model has been integrated by other variables and some variables of the original model have been replaced by others) shows better effectiveness comparing to the original model.

Key-Words: Crisi e risanamento d’impresa, insolvency prediction models, economic crisis, Italian companies, financial alert model

Table 5 – First and second type errors of the models

Table 7 –First type and second type errors of the updated model

Table 12 – Comparison of the models