中國上市公司財務危機預警模型之研究－以滬深A股上市公司為例

Abstract

中國股市一直為人詬病的是其資訊的不透明，加上中國缺少具公信力的企業信評系統，導致投資人難以掌握中國上市公司實際的經營狀況和風險。因為本研究之目的在於建置中國A股上市公司的財務危機預警模型從而做為投資人判斷上市公司經營狀況與風險的參考依據。 本文研究期間從2007年至2012年，財務危機樣本的選定標準主要分成兩種：特別處理（Special Treatment，ST）公司和出現逾期借款，且逾期借款金額超過該年度淨利潤的公司。本文選定六種變數，並採用Logistic迴歸模型分別建置危機發生前一至三年的危機預警模型，選定的變數為：資產報酬率、流動比率、負債比率、總資產周轉率、其他應收款占總資產比重和審計意見。 Logistic迴歸結果顯示，以借款逾期公司為樣本所建置的預警模型解釋力較為不理想；由全體危機公司為樣本建置的模型判別準確率在前兩年都達到84.6% 的準確率。本文也發現不同的變數具有不同的預測時效性，「資產周轉率」和「其他應收款比重」不論在長短期預測模型都能做為重要的預測變數；「流動比率」和「總資產報酬率」較適合用於長期預測模型；「負債比率」和「審計意見」則適合用於中短期的預測模型。

Since the lack transparency of information on China listed companies, Investors can’t easily to find out the “real” company’s financial situation. Therefore, Identification of potential financial distress and offering early warnings to investors has become important. This study applies the Logistic Regression with six variables to build financial distress prediction model of China A-shares listed companies. The distress companies in this article including the companies under special treatment (ST companies) and the companies which have defaulted debts (Defaulted debts companies), and the samples were selected from 2007 to 2012. The practical result shows that the distress prediction model which estimated with whole distress companies works much better than the model which estimated with defaulted debts companies only. In addition, six independent variables have different effects during the different periods, some have a great predictive ability in the short-term distress prediction model; some are suitable for the long-term prediction model.