Application of Taxonomic Measures to Bankruptcy Prediction
Purpose: The paper aims to propose new method of bankruptcy prediction. In our research we construct composite measures of financial efficiency using taxonomic distance to the distinguished pattern. Design/Methodology/Approach: In our study we use the sample of 136 Polish manufacturing non-public companies. Half of them are bankrupts (i.e. filed for bankruptcy with the court in years 2019 – 2022), whereas the rest of them run their business and are companies with a similar amount of assets as bankrupts. Data used in research has been acquired from the Emerging Markets Information Service EMIS, which contains financial reports information one year prior to the bankruptcy filing. According to the value of these measures calculated for all analyzed companies they are classified to two classes. Findings: The study shows that taxonomic measures are useful for predicting corporate bankruptcy. Identifying a grey zone improves classification accuracy within specific clusters, even though it slightly lowers overall model performance. The results also highlight company size—measured by asset value and structure—as a key factor distinguishing bankrupt firms from those that remain solvent. Practical Implications: The results of our experiments show that level of recognition of both groups of companies is quite high but it depends on the selected pattern. Originality/Value: The study highlights that the taxonomic measures applied are simpler than many other bankruptcy prediction methods, making them more accessible. Their straightforward nature enables use by managers of smaller firms that do not have dedicated financial staff. As a result, these measures offer a practical tool for monitoring bankruptcy risk in resource-constrained organizations.