台灣上市公司財務危機預警模式之建立-瀑布羅吉斯模型之應用

Abstract

民國八十七年的金融危機，對台灣產業界及投資人造成莫大的傷害。本研究試圖將近年來發生財務危機之上市公司進行分析，期能找出公司財務危機之預警模式。在此我們將採三大構面去剖析公司是否會發生危機之指標，分別為財務比率構面、產業因素構面、及其他質性及量性因素構面，並進一步採用動態分析之方式，深入瞭解公司發生危機前財務狀況之變化時點。本研究之特色係首先採用瀑布羅吉斯模式( Cascaded logistic model )進行預警模式之建構，再進一步驗證此模式之實用性及預測能力。研究結果指出，整合得到的瀑布羅吉斯模型，其正確分類率及預測能力皆較各類型變數單獨建立之模型來的好，也顯示出此模型具有不錯的實用性。在此期能協助投資大眾及產官學界儘早預知公司財務危機之發生，而做好監督及補救的工作。

The financial crisis in 1998, it caused a great damage to Taiwan industries and investors. In this research we analyze the financial distress companies’ data in order to develop a model which can predict the financial failures. In constructing the model, three broad categories of indicators are investigated, which are financial ratio category, industry relative ratio category, and other qualitative or quantitative measurement category. Furthermore, we add the change rate of financial ratios into the model in order to figure out if there is a big change in financial ratios before companies go financial distress. The special feature of this research is that we first employ the cascaded logistic model in developing the model for predicting financial failures. A holdout sample test enables us to show that the model has its validity. We find that the cascaded logistic model has better prediction power than other models, so we can tell that the cascaded logistic model has its applicability. The model can help investors, industries, government agencies, and academics identify the financial distress of companies early. Therefore, the supervision and preventive actions can be taken accordingly.