An Examination of the Stationarity of Multivariate Bankruptcy Prediction Models: A Methodological Study

Several issues are raised concerning previous bankruptcy prediction studies using multiple discriminant analysis of financial ratios: 1. inconsistency both in the values of the coefficients reported and the relative importance of various financial ratios used, 2. pooling of data across years without considering underlying economic events in those years, and 3. whether to control for multicollinearity. A study is undertaken to examine these issues in an empirical setting. Three external economic factors are investigated: 1. inflation, 2. interest rates and credit availability, and 3. the business cycle. Logit analyses were employed as the primary statistical tool. Analysis leads to 3 general conclusions: 1. The accuracy and structure of predictive models differ across different economic environments. 2. Different prediction models seem appropriate for firms in different industrial sectors, even for the same economic environment. 3. More useful results may be obtained by specifically considering multicollinearity in the intertemporal and intersectoral development of the models.