建構企業財務危機預警模型——考慮財務與公司治理因素

Establish a Warning Model for Financial Distress—Considering Financial Ratios and Corporate Governance

本研究運用羅吉斯迴歸分析，建構財務危機預警模型，除了傳統的財務比率變數之外，並加入了公司治理變數，與學者提出之Z-score模型與ZFC財務危機指標作比較，探討以上模型在區別正常與危機公司的準確度何者較佳，實證結果顯示：1.危機前一、兩年各危機預警模型之整體預測準確率大多可達到80%以上，各模型皆具有一定的預測能力，但在判別危機公司之準確率上，加入公司治理因素的財務變數預警模型表現最佳。2.對危機預警模式具有顯著解釋能力之財務變數為負債比率;具有顯著解釋能力之公司治理變數為董事長兼任總經理、董監質押率、董事長異動次數。3.存貨週轉率過高，意味著存貨管理可能生問題，使模型內財務變數產生不符合預期正負關係的結果，影響模型的解釋能力。

The purpose of this study will establish the financial distress warning model considering not only the traditional financial ratio variables; corporate governance variables will be included too. The logistic regression is used to establish the warning model. Copared the predictive ability of financial distress with Z-score and ZFC financial distress warning models, the empirical results express: 1. The overall predictive accuracy approximately exceeds 80% in each warning model inclunding one and two years prior to the distress. The adequate predictive ability exists in each model. But the model considering both financial ratio and corporate governance variables obtains the best predictive accuracy in discriminating distressed companies. 2. In the financial distress warning model, Debt Ratio, is the financial variable with the significantly explanation ability. CEO Duality, Collateralized Shares, and CEO Change, are the corporate governance variable with the significantly explanation ability. 3. Excessively high inventory turn-over ratio suggests some problem in inventory management. It may lead to some relationship among valiables not complying with theory. So the explanation ability of model will be less significant.