Verification of the effectiveness of discrimination models for forecasting bankruptcy of enterprises

Abstract

Purpose: An attempt to evaluate the effectiveness of financial analysis tools used to assess financial health and to forecast bankruptcy. Methodology: The study used 31 of the most popular discriminatory models for bankruptcy prediction. The effectiveness of early warning models has been evaluated on the basis of financial data of economic entities operating on the territory of the Republic of Poland. The sample of the enterprises has comprised a total of 172 entities – both bankrupt and operating in good financial condition, located in 16 provinces. The data period was 2011-2020. The companies have represented various sectors of activity. Data was obtained from the Emis.com website. Findings: Most of the models used have been highly effective in forecasting. However, none of the methods has been 100% effective. It has also been noted that the models estimated on the basis of a sample of enterprises from only one sector, the so-called sector models allocated to the evaluation of a specific industry, have not had a significantly higher percentage of correct diagnoses than universal models. In the last analyzed period, the three most effective methods are the so-called universal models that can be used in the evaluation of companies from various industries. Practical implications: The results of the audit can be used, among others, in the assessment of the going concern of enterprises by business managers, business analysts, investors, and above all, statutory auditors when auditing financial statements. Originality/value: The study provides an answer to the question of which models are characterized by high prognostic effectiveness. In addition, the obtained results allow to resolve the issue of the usefulness of models created in the 1990s. The large size of the research sample allows for the generalization of the results and their wider application than has been the case in the literature on the subject so far. Keywords: bankruptcy, financial condition, discriminatory models, early warning models. Category of the paper: Research paper.