Paper Title
A Multivariate Logit Model for Corporate Bankruptcy Forecasting

Abstract
This study develops a bankruptcy prediction model by combining financial ratio analysis and logistic analysis from a sample of 258 bankrupt and non-bankrupt public companies in the United States. First, Mann-Whitney tests reveal a significant difference in the mean values of bankrupt and nonbankrupt companies for 48 financial ratios. Second, based on these financial ratios, a logistic modeling procedure is used to develop bankruptcy classification model. Finally, validation of the bankruptcy model is by holdout-sample Type I accuracy, Type II accuracy, and overall correct classification rates. The research result shows that the logit model possesses high classification accuracy and relatively small differences in classification rates between in-sample and out-of-sample as compared to industry-relative analysis. As such, the model help managers more accurately estimate bankruptcy risk and hence, have a better opportunity to take corrective actions early. Keywords - Bankruptcy, Financial failure, Forecasting, Logit models.