Construction of bankruptcy prediction model using discriminant analysis and financial ratios

Abstract

The main aim of this research is to develop a statistical model that can reliably predict bankruptcy of Serbian companies one year before bankruptcy proceedings start. The main motive for the research is the fact that there are not many scientific papers focusing on this important issue in Serbia. Bankruptcy prediction model may be useful for future researchers, but also for business owners and other stakeholders. Research was conducted using financial ratio indicators and discriminant analysis in IBM's SPSS v.26 program. Initially 100 companies from the territory of Serbia were included in the research, but after data screening and meeting all the assumptions for discriminant analysis, 74 of them were included in the final modelling process. It was confirmed that the commonly used financial ratios and discriminant analysis can be useful in creating a bankruptcy prediction model, since the classification power of the developed model is 71.6% for original grouped cases, and 70.3% for cross-validated cases.