碩士 / 輔仁大學 / 應用統計學研究所 / 94 / In recent years, the listed and OTC companies in Taiwan unexpectedly declared reform, default or leave listed for the shareholders, they could only rely on the regular financial reports rather than selling the stock on the peaks. Consequently, How to construct an efficient distress warning systems in order to offer the alert information of the distress in advance is an important issue. In this paper, we propose three kinds of two-stage financial distress warning models to improve the predict ability of the financial crisis and let shareholders evade the risk. The data of listed and OTC companies in Taiwan ranging from 2000 to 2004 are selected as studied samples. First of all, we utilize the data of the banking and manufacturing industry to find out possible financial variables and corporate governance variables that account for financial distress. Second, we apply the cross-validation method to compare the performances of logistic regression, artificial neural networks and three kinds of two-stage financial distress models.
The result of experiment shows that for the banking the affected variables are the non-performing-loans (NPL) ratio and certificate of depository ratio and for the enterprises, return on assets (ROA), return on equity (ROE), operating margin ratio, acid-test ratio, debt ratio, cash flow ratio, equity growth ratio, inventory turnover rate, fixed assets turnover, percentage of board of directors, percentage of controlling shareholder. For the banking we can use the non-performing-loan (NPL) ratio and certificate of depository ratio construct financial distress models. In the other side, for the enterprises we can use percentage of board of directors, percentage of controlling shareholder and equity growth ratio measure the possibility of the financial distress risk. This study shows that for the banking, the third two-stage financial distress model has highest accurate rate among five models. As regards manufacturing industry, the first two-stage financial distress model has highest rate among five models. Accordingly, it seems that the accurate rate of the two-stage financial model is larger than the one of the one stage-financial model.
| Identifer | oai:union.ndltd.org:TW/094FJU00506035 |
| Date | January 2006 |
| Creators | Gi Pei-Hua, 紀佩華 |
| Contributors | Hou Chia-Ding, Lin Shu-Ling, 侯家鼎, 林淑玲 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 86 |
Page generated in 0.0787 seconds