Application of Statistical Methods in the Post-Hoc Analysis of Bankruptcy Risk Prediction Models

The main objective of this study is to predict bankruptcy risk using statistical methods. The risks generated by the activities of any entity, as well as their mismanagement, can lead to the instability of the company or, in the worst case, to its bankruptcy. To forecast this type of risk, we will use statistical techniques, which can provide concrete financial information that helps both internal decision-makers within the company in implementing the best financial strategies, as well as external parties: clients, investors, etc. The post-hoc analysis is based on a sample of 30 Romanian companies, using the previously developed D model. The study covers a period of 3 consecutive years, from 2021 to 2023, and employs a total of 9 economic and financial indicators that most accurately reflect the probability of bankruptcy. The statistical methodology used is discriminant analysis. The financial variables identified as likely to differentiate between bankrupt and non-bankrupt companies are organized into four groups: activity ratios, liquidity and solvency ratios, debt ratios, and profitability ratios. Throughout the study, based on the results obtained for each company individually, the hypothesis that the D score is effective and applicable to Romanian companies is confirmed.