Assessment of trajectories of non-bankrupt and bankrupt enterprises

Abstract

Purpose: The aim of this study is to show how long-term trajectories of enterprises can be used to increase the forecasting horizon of bankruptcy prediction models. Design/Methodology/Approach: The author used seven popular forecasting models (two from Europe, two from Asia, two from North America and one from Latin America). These models (five multivariate discriminant analysis models and two logit models) were used to develop 17-year trajectories separately for non-bankrupt enterprises and those at risk of financial failure. Findings: Based on a sample of 200 enterprises, the author evaluated the differences between non-bankrupt and bankrupt firms in development during 17 years of activity. The long-term usability of the models was demonstrated. To date, these models have been used only to forecast bankruptcy risk in the short term (1–3 years’ prediction horizon). This paper demonstrates that these models can also serve to evaluate long-term growth and to identify the first symptoms of future bankruptcy risk many years before it actually occurs. Practical Implications: It was proven and specified that long-term developmental differences exist between non-threatened and future insolvent companies. These studies proved that the process of going bankrupt is very long, perhaps even longer than the literature has previously demonstrated. Originality/value: This study is one of the first attempts in the literature globally to assess such long-term enterprise trajectories. Additionally by implementing a dynamic approach to the financial ratios in the risk-forecasting model let visualize the changes occurring in the company.