摘要：This paper investigates the ability of logistic regression in anticipating corporate failures in Malaysia over a ten year period covering the economic cycles of boom, the Asian financial crisis of 1997 and economic recovery. 64 companies were analyzed with an initial 16 financial ratios. A strong logit model was developed with four ratios found to be significant in its predictive power and classification results showed very high average accuracy rates of 88% and 90% for the analysis and the hold-out samples respectively and for each of the five years preceding the actual failure. This study also showed it does not need many ratios to be able to anticipate potential company failures and that even with more advanced statistical models used recently, logistic regression is still a very effective and reliable statistical tool.
