羅吉斯迴歸與公司治理指標於財務危機預測模型之應用

公司財務預警一直是財務學界與實務界重視的議題之一，過去學者預測企業發生財務危機，主要是利用財務比率作為自變數進行研究。近年之研究論文，除了使用傳統的財務變數指標之外，更加入了公司治理因素指標進行探討。然而這些研究中，公司治理因素的衡量相當複雜且較為主觀，因此本研究採用一個較全面、客觀，更容易取得的公司治理指標來探討加入公司治理指標是否可以提昇公司財務預測之正確率。研究樣本為選取於民國93年1月1日至民國94年12月31日間,排除金融產業後，在公開資訊觀測站資料庫所記載之上市且資料完整之公司，其中包含了22家財務危機公司及44家財務正常公司共66家上市公司。研究方法則運用羅吉斯迴歸（logistic regression）模型，分別建立財務比率單一指標模型及包含財務比率與公司治理指標的雙指標模型，比較財務危機預警之區別正確率。研究結果發現，加入公司治理變數後，模型整體區別正確率由86.4%提昇至93.9%。

關鍵字

公司治理；財務比率；財務危機；羅吉斯迴歸

並列摘要

Corporate financial distress has been an important issue for financial field. Previous articles generally used financial ratios as independent variables to predict corporate financial distress. In recent years, researchers have started to add corporate governance measures as independent variables. However, the corporate governance measures used in these researches are complicate and subjective. This study uses a simple and objective corporate governance measures and examine whether the inclusion of corporate governance measure in a logistic regression prediction model would provide better results than that based on traditional financial ratios alone. The samples of this research are selected from the M.O.P.S (market observation post system) database from Jan 1, 2004 to Dec 31, 2005. The samples have 66 companies including 44 normal companies and 22 financial distress companies. The result of empirical analysis showed that 86.4% of the observations were classified correctly in the model based on traditional financial ratios alone. And the correct rate of classification increased to 93.9% in the model incorporating the corporate governance measure and traditional financial ratios.