A study of Altman's (1983) revised four-variable Z-score bankruptcy prediction model for asset sizes and manufacturing and service companies

Bankruptcy prediction models have been around since the first half of the twentieth century. These models are still very much needed in today's economy. Managers need to know if their businesses are operating soundly. Investors need to know the same information for their investments. Creditors want to make sure that loans can be recouped. All stakeholders should have a bankruptcy prediction model to be alerted if bankruptcy is on the horizon since they are all affected by it.In 1968, Edward Altman started the process of transforming univariate analysis to multiple discriminant analysis. Multiple discriminant analysis is used primarily to classify and/or make predictions where the dependent variable is in qualitative form such as bankrupt and non-bankrupt.This study uses Altman's 1968 and 1983 z-score model to determine if it still predicts bankruptcy and if it applies to our modern economy. The sample consists of 50 failed manufacturing companies and 11 failed service firms for the period 1998-2003. An equal number of non-bankrupt firms were matched to the bankrupt firms. Matching the paired-samples involved using similar SIC codes and similar asset sizes. Three hypotheses were tested. Sample A or the validation sample used 31 bankrupt and non-bankrupt firms. Sample B or the holdout sample included the 30 bankrupt and non-bankrupt selected in Sample A.Each hypothesis was tested at the 95 percent confidence level. H1a: There is a significant difference in bankruptcy prediction among companies using the Altman 1983 four variable z-score model. H2a: There is a significant difference with manufacturing and service companies using the Altman 1983 four-variable z-score model. H3a: Altman's 1968 z-score is a significantly better predictor of bankruptcy than Altman's 1983 four-variable z-score model.Altman's 1993 four variable z-score model has no significant relationship between asset sizes and the four-variable z-scores of bankrupt and non-bankrupt companies. There is a significant relationship in bankruptcy prediction using manufacturing and service companies. The study did not find any significance that Altman's 1968 original z-score model was a better predictor than Altman's 1983 four variable z-score model.