本研究以傳統的危機預警模型為基礎,加入現金流量風險值(cash flow at risk, CFaR) 作為額外的解釋變數,以評估CFaR是否能預測公司危機發生的機率。實證上採用縱橫羅吉斯迴歸模型進行估計,以同時考量時間序列或橫斷面資料之間重要性,強化模型配適度及解釋能力。在所建構的危機預警模型中,以文獻上最常使用的財務比率為基礎,做為預警模型中的解釋變數,包括:資產報酬率(ROA)、速動比率、負債比率、應收帳款週轉率(次)、存貨週轉率(次),以及淨值報酬率。此外,並加入現金流量風險值為解釋變數。實證對象為12家危機公司與36家正常公司,實證期間為2002年下半年至2007年上半年,共計480筆半年資料。實證結果顯示,財務指標的顯著性與文獻相同;在加入CFaR後,整體迴歸估計式的配適度提升,顯示加入現金流量風險值後,有助於解釋公司發生危機的機率。
This study adds the variable-cash flow at risk (CFaR) into traditional crisis warning model to investigate whether this variable can improve the forecasting performance of crisis warning model. To execute the empirical study, we adopt a panel logit regression model, which considers cross-sectional and time series data simultaneously and can improve the forecasting ability of model. In the constructed panel logit regression model, we select six financial indices frequently used in the literature as the regressors, including return on asset (ROA), liquidity ratio, debt ratio, account receivable turnover rate, stock turnover rate, and return on equity. In addition, we add the CFaR into the model as a new regressor. The sample period spans from 2002 to 2007 and the sample objects cover 12 crisis companies and 36 healthy companies. The data set comes from the TEJ database. Empirical results show that traditional financial ratios provide powerful ability in explaining the probability of a company’s crisis. In addition, adding cash flows at risk into conventional crisis warning model can improve the predicting ability of model.