Testing the usefulness and predictive power of the adapted Altman Z-score model for Bulgarian public companies

The applicability and predictive power of the models for predicting future corporate bankruptcy depend on both the selection of the individual indicators that are chosen to construct them and the coefficients (weights) with which they are included in the model. The latter are heavily dependent on the conditions in which the companies operate and require the adaptation of the models to the economic environment in different countries and/or to different types of companies. The purpose of the paper is to test the adapted Altman Z-score model for Bulgarian public manufacturing companies (D), derived by Georgiev and Petrova (2015) in order to verify its applicability and predictive power 5 years after its release. The empirical study is designed in three parts to meet the following objectives: (1) to verify the predictive power of the adapted model (D) using the original sample of companies; (2) to test the continued applicability of the model (D), using a new sample of companies that became insolvent five years after the adapted model was derived; and (3) to compare the accuracy of the adapted model (D) to the original Altman Z-score and Z’-score models for Bulgarian public manufacturing companies. The empirical data about the status (bankrupt/non-bankrupt) of the companies in the original sample confirms that the adapted model D (Georgiev and Petrova 2015) correctly predicted their future likelihood of bankruptcy with an even higher success rate than the one achieved in the initial study (85% for 2 years and 70 % - for 5 years into the future). The results from testing the model with new data prove that it is still effective in predicting bankruptcy and can be used in practice 5 years after its release, though its success rate shows a slight (about 5%) decrease. The results also show that the adapted model D performs better than the original Altman’s Z-score and Z’-score for Bulgarian public companies.