In this paper the applicability of two well-known financial distress models namely multiple discriminant analysis and logistic regression analysis have been examined. By using a sample of 298 firms it is found that cash flow and working capital are important predictive variables irrespective of models selected and these models are capable of predicting with minimum error one year in advance, which is vital for the bankers, restructuring agencies and the management to initiate revival process before the company actually gets into financial distress.

The Indian industry was exposed to large-scale domestic and international competition following her economic liberalization in 1992. While few firms were able to take up the challenge, a large number of firms were affected by the competition. The level of non-performing assets of Indian banking system have increased several times during this period. Indian banks today are reluctant to extend credit to industries on the fear of default and the level of investments in government securities has gone up beyond the statutory requirement. The detection of company operating and financial difficulties are subjects which have been particularly amenable to analysis with financial ratios (Altman, 2000). Though at one extreme, many learned academicians question the validity of financial distress prediction models using financial ratios, there is continuing interest in refining and testing financial distress prediction models. Beaver (1966) initiated the interest of academic world to the financial distress prediction models using univariate analysis methodology for classifying bankruptcy and non-bankruptcy firms. The importance of the subject attracted the interest of several authors from different countries. Dimitras et al. (1996) listed 47 studies conducted in 12 countries.

Altman (1968) extended the work and used multivariate discriminant analysis to predict bankruptcy. Several authors have replicated the Altman model in different countries and reported the usefulness of financial ratios in predicting bankruptcy. These models were also extended to examine the quality of existing loan accounts and also in deciding whether to extend the credit or loan to the borrowers. In order to overcome some of the limitations of discriminant models, subsequent studies in 1980s have used logistic analysis. Logit analysis was first proposed for bank failure prediction by Martin (1977) and for the prediction of business failure by Ohlson (1980). Subsequently many others have followed logistic analysis.