在公司治理观点下.docVIP

在公司治理觀點建立上市公司財務危機預警模型－以施東信1 晏啟華2 1國防大學國防管理學院資源管理研究所研究生 Address:台北縣中和市民安街150號 TEL: 028558 E-mail:openpower1999@.tw 2國防大學國防管理學院資源管理研究所助理教授 Address: 台北縣中和市民安街150號 TEL:028558 E-mail: cheewha@ 傳統的財務危機預警模型都是用財務變數來建構，但是只用財務變數所建立的預警模型似乎無法正確預警出近年來的財務危機事件。因此本研究除了考慮財務變數之外，同時採用最近研究中所指出對財務危機具有影響的公司治理變數來構建預警模型，以改善目前通行之財務危機預警制度，研究發現倒傳遞類神經網路建構「加入公司治理」財務預警模型可即時有效區分出台灣財務危機公司與正常公司，因此，有助於投資人、債權人、貸款銀行、公司管理者與稽核專員，來偵查及預防舞弊，進而提早預警上市公司的財務危機事件發生。另外，研究結果，證明在財務危機預警模型增加「公司治理」資訊：「董監事持股比率」、「全體董、監事質押比率」與「主要經理人持股比率」3個公司治理變數可有效提升財務危機預警模型的預警能力。 關鍵字: 公司治理、財務危機、預警模型、倒傳遞類神經網路ing Artificial Neural Networks to Construct the Financial Distress Alerting Model for Listed Companies in the Perspective of Corporate Governance -A Case Study of Taiwan Shi, Dong-Shin1 Yann, Chee-Wha 2 1Research Student Graduate School of Resources Management, National Defense Management College, Taipei, Taiwan Address:150 Min-An Street, Chung-Ho, Taipei County TEL: 028558 E-mail:openpower1999@.tw 2 Assistant Professor Graduate School of Resources Management, National Defense Management College, Taipei, Taiwan Address:150 Min-An Street, Chung-Ho, Taipei County TEL: 028558 E-mail: cheewha@  ABSTRACT Traditional studies utilize financial ratios to construct financial distress alerting models. However, the models considering only financial ratios couldn’t predict financial distress’s happening in recent years. This study intends to construct financial distress alerting model not only with financial factors, but also with corporate governance, which are demonstrated to affect financial distress dramatically. After combining the information of financial factors and corporate factors, this study creates a “corporate governance” financial distress alerting model which can improve the weaknesses of traditional financial distress alerting modes. The study shows that Back-Propagation Neural Network model can learn from the data of bankrupt corporations and a matched group of survivo