Firm Bankruptcy Forecast Using Survival Analysis Approach

Abstract to scientific paper “Firm Bankruptcy Forecast Using Survival Analysis Approach” The aim of this study was to analyze the probability of bankruptcy for the Russian companies in the manufacturing sector, taking into account time-dependent covariates. To achieve this goal the following tasks were performed: (1) an analysis of the literature on bankruptcy prediction; (2) an investigation into main theoretical models on bankruptcy prediction; (3) a construction of a probability survival function corresponding to the duration of the companies’ existence and an identification of the main features of bankrupt firms; (4) a construction of survival analysis model, namely Cox regression with time-dependent covariates; (5) a detection of statistically significant predictors of the risk of bankruptcy; (6) an estimation of the predictive power of the model by identifying its accuracy. According to the results of Cox’s model evaluation it was revealed that nine financial indicators have a significant impact on the risk of becoming bankrupt. Moreover, the resulting model has an accuracy of 81 percent, which serves as an indicator of a properly chosen specification.