近年全球金融風暴不僅讓金融機構自身難保,就連企業違約、倒閉的例子時有所聞,更造成投資大眾投資的嚴重損失。本研究運用台灣經濟新報(TEJ)資料庫台灣地區1992年1月1日到2008年12月31日違約上櫃公司為樣本,從而找自違約時點前一年的之財務資料,其中包括加權指數、公司市值、收盤價、負債帳面價值、無風險利率等,再利用Merton model所計算出的違約距離、Z-Score與Zmijewski Score作為自變數,得知其相關性後,使此變數建構成一個能最適合羅吉斯迴歸模型運用,再以羅吉斯迴歸模型建構公司財務危機預測模型,檢視模型是否可以提供有效資訊預測公司是否陷入財務危機。 最後其模型驗證得出會計基礎模型(Z-score、Zmijewski)所算出違約機率準確性較高,計算越複雜之市場基礎模型(Merton model)預測效果並不如預期,而本研究加入之系統風險其解釋效果不大。
Global financial crisis made financial institution fail and corporation default. The crisis also shrinks investors’ assets and hurts their confidence. The paper based on TEJ’s database, which includes stock index、market value、price、D/E ratio and risk-free interest rate, since Jan 1992 to Dec 2008. Based on the data, I use default distance and Z-Score by Merton model as independent variables, and calculate the related. After that, I use these variables to establish a financial crisis forecast model. In conclusion, Z-score can precisely forecast default possibility, but Merton model can’t. Besides that, if we put systematic risk in model, it can’t really work out.