An internal control approach to construct a litigation warning model

• Main Authors: Wen-Pi Yen, 嚴文筆
• Other Authors: Chin-Shien Lin
• Format: Others
• Language: zh-TW
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/83128477235476593888

碩士 === 國立中正大學 === 會計與資訊科技研究所 === 92 === An Internal Control Approach to Construct a Litigation Warning Model Abstract Since 1996, a series of financial scandals such as Enron and WorldCom’s bankruptcy have not only surprised local and international capital market, but also led to serious losses by investors and creditors. Suddenly, the credibility to the function of management and internal/external auditing collapses and the attention to the corporate governance is greatly being addressed. The previous researches on fraud were mostly divided into two categories. One is the exploration of “red flags”; the other is the comparison among applied statistical tools. However, the former with its enormous emphasis on analyzing the relevant importance of “red flags” lacks a practical connection with employing real sued cases into the enterprise’s internal control systems. This study initially utilizes sued cases to analyze the characteristics and weaknesses of internal controls. 74 sued cases and 148 healthy companies were selected as research examples. According to SAS No. 82, four re-grouped risk factors were found from factoring 27 fraud risks. Logistic regression is applied to establish a litigation warning model. The results were listed: 1. The high risk to management characteristics is consistent with the family-run feature of Taiwanese corporations. 2. The logistic regression results show that the risk variables of management characteristics and industrial environment were coefficients significant, but the operating and financial characteristics and assets peculation were not. Hence, in evaluation of internal controls and fraud risks, more attention should pay to those significant risk factors. 3. In the test of correctness of model prediction, logistic regression reaches 85% correctness and shows type II error 12%. Therefore, the relatively high correctness implies that such a tool is proper to be used in fraud detection. 4. A further analysis between litigation and healthy companies, ANOVA results indicate that there are explicit differences in risk factors between these two groups. It means, the risk factors included in the study can be implemented to the internal controls’ evaluation. Key Words：internal control、fraud、litigation、warning model