早期學者對於財務危機預警模式的研究,大都只採用財務比率作為變數,並未考慮到產業訊息與業外特性比率等因素。有鑑於此,本研究採用Logit迴歸法建立財務危機預警模式,逐步加入公司財務穩定性比率變數、產業相對財務比率變數,以及業外特性比率變數。實證結果顯示:1.加入產業訊息變數之財務危機預警模式,在正確區別與預測率上均有明顯地提升;2.含財務穩定性變數之模式並不能夠提升模式整體的預測能力;3.以業外特性變數建立之財務危機預警模式,不論是含產業或是不含產業訊息之Logit模式,都能提高模式之正確區別能力與預測能力;4.綜合考慮所有模式之總區別及預測能力,則以採取除以產業平均數之相對財務比率,再加上業外特性變數模式的區別與預測能力最佳。
Previous studies use financial ratios to construct financial distress warning model without considering industry information and the non-operating factors. This study employs logistic regression model to construct the financial distress-warning model, and tries to incorporate financial stability ratios, relative industry benchmark and the non-operating factors. The empirical findings state as follows: The model contains industry information has better distress prediction performance. The model contains financial stability ratios is not able to increase prediction power. Including non-operating factors into the model can increase the model’s prediction ability. Among the various models, the model with industry-relative financial ratio (divided by its industry benchmark) and the non-operating factors has the best discriminating and predicting power.