| dc.description.abstract | Financial Crisis Warning Models have been developed for a long time and it main purpose is to find out the financial problems within companies in advance for bankers or investors. Though some complicated warning models, such as Artificial Neural Net Work Model or KMV Model have been investigated recently, in order to assist investors aware the finial crisis in advance by using the brief financial report, this study focuses on whether some traditional financial crisis warning models, such as The Discriminate Model and Logistic Regression Model are still available presently.
Companies with the financial crisis which listed on TSE and OTC between 1990 and 2004 in Taiwan were targets of this study. This study investigates not only the identification ratio of original Altman Z-Score model, but re-regress the parameters of the five financial variables within the original Altman Z-Score models. In addition, this study focuses on the market factors and financial ratios in five main divisions that have been used frequently. Furthermore, the statistics methods were used to find out the most dependable indication within the six main divisions which predicts the financial systems of the enterprise and then to build up the two popular financial crisis warning models that has been used theoretically and practically (i.e., the Discriminate Model and the Logistic Regression Model). This study compares the identification ration of these two warning models and the findings reveal that:
1.The results of T-test: It can be understood that within the six main divisions the net worth/total assets and total liability/ total assets ratio are the most important variables of the indication of company paying ability followed by the variables of EPS and cash flow ratio which means the distinguishable differences between companies with and without financial crisis is the degree of paying and earning ability. The findings coincide with people’s intuition to the companies with financial crisis. Moreover, the cash flow ratio is not only a distinguishable variable which is used to analyze the asset of the companies, but it also coincides with the present scholars’ concept of cast flow.
2.Starting from one to three years prior to the financial crisis, the justified predication rate of the Discriminate Model in order is 88.75%, 80.00%, and 73.75%. On the other hand, the justified predication rate of the Logistic Regression Model in order is 90.00%, 72.50% and 67.50%. Based on these findings, it suggests that one year before the financial crisis, the Logistic Regression Model provides a more correct predication while the Discriminate Model provides a better predication two and three years before the financial crisis. | en_US |