| Numerous researchers and practitioners have studied bankruptcy prediction for the past sixty years. Paul J. Fitzpatrick, in 1932, appears to be the first to perform a detailed study comparing bankrupt firms to healthy firms using financial ratio analysis. Edward Altman, in 1968, was the pioneer in using multivariate analysis, or multiple weighted ratios, to arrive at a single Z-score for distinguishing between bankrupt and nonbankrupt firms. Altman's original Z-score model used five financial ratios and served as a very accurate predictor of financial distress or health for large manufacturing firms.;Over the years, Altman refined his original model and developed new models in response to changes in economic characteristics and differences in industry environment. In 1993, he published a revised four-variable model which eliminated the variable of sales to total assets in order to minimize the potential industry effect when such an industry-sensitive variable as asset turnover was included. Also, new discriminant weights were established for the remaining ratios.;This study utilized Altman's 1993 revised four-variable model to evaluate the financial condition of publicly traded retail firms. A validation sample and an evaluation sample were involved in the study.;Sample A contained 32 bankrupt and 32 nonbankrupt firms drawn from the retail industry and matched by SIC code and asset size. Sample A served as the validation sample. The model yielded accuracies of 92 percent, 69 percent, and 62 percent for the periods one, two and three years prior to bankruptcy using Altman's cut-off values of less than 1.10 for bankrupt firms, 1.10 to 2.60 as a grey area, and above 2.60 for healthy firms. This study established revised cut-off values of less than 2.31 for bankrupt firms, 2.31 to 2.35 as a grey area, and above 2.35 for healthy firms. These revised cut-off values yielded predictive accuracies of 98 percent, 84 percent, and 77 percent for the periods one, two and three years prior to bankruptcy. The narrower range of the revised cut-off values indicates improved model accuracy.;The second sample, Sample B, contained 31 bankrupt firms matched with 31 nonbankrupt firms and was used as an evaluation, or holdout, sample. Three hypotheses were tested using these samples.;Three hypotheses were tested at the 95 percent confidence level. H 1: The mean Z″-Scores for bankrupt. and nonbankrupt firms are not equal. H2: There is a significant relationship between the Z″-Score and the asset size of firms. H3: The population distributions of Z″-Scores for growing firms and declining firms are significantly different.;Altman's 1993 revised four-variable Z″-Score model with revised cut-off values can be used to accurately distinguish between bankrupt firms and nonbankrupt firms in the retail industry. There is no significant relationship between Z″-Scores and asset sizes of retail firms. There is no significant difference in the Z″-Scores for retail firms that are growing and retail firms that are declining in sales revenues. |
