As suggested by Altman (1968), companies are much more likely to fail if they are unprofitable, highly leveraged, and suffering from cash flow difficulties. For this study, companies that have been liquidated and those that have ultimately gone bankrupt are both included as failing companies. Although many early bankruptcy prediction models used discriminant analysis (DA) to predict financial failure, this technique has become less frequently used because of two assumptions made to use DA. First of all, to use DA, the variables used are assumed to have multivariate normal distributions, and it is well known that the variables normally used in bankruptcy studies are not normally distributed. Also, the samples of companies that are failing and non-failing are assumed to be randomly drawn, an assumption that does not allow for the commonly used technique of matched pairs. These problems, along with many others, have caused more researchers to turn to using probit and logit models instead of discriminant analysis.
This paper will focus, in particular, on using logistic regression modeling in order to predict bankruptcy. The unique incremental contribution of this paper is the focus on companies only from the retail industry for a recent sample period. Previous studies including Lennox (1999) have included a variable in the model for industry sector in order to rule out industry bias in the sample, but most studies have not examined the use of bankruptcy prediction models in the retail industry. The results of this study may provide valuable information to retail companies and investors in the retail industry.
The bankruptcy prediction model used in this study uses five variables that have been significant predictors of bankruptcy in previous studies. The variables to be used in this paper as predictors of bankruptcy are number of employees (EMP), return on assets (ROA), cash flow margin (CFLM), debt-to-equity ratio (DTEQ), and cash to current liabilities (CHELCT). Number of employees is expected to be a significant predictor of bankruptcy because of the increased likelihood that smaller companies with fewer employees will fail. Return on assets is included as a measure of profitability of the companies, and it is assumed that less profitable companies are more likely to fail. Cash flow margin and cash to current liabilities are included in the study as measures of company liquidity. The debt-to-equity ratio is a measure used to analyze the solvency of a company, and it is included in this study because it appears to be a good measure of company performance. The variables are best illustrated in Table 1.