An Application of Neural Networks for the Financial Early Warning System-A Case of Local Banks
| Autor: | CHEN KANG YU, 陳康宇 |
|---|---|
| Rok vydání: | 2010 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 98 The establishment of private banks at Taiwan starts from Year 1991, the number of private banks increased from 25 to 53 just after 10 years. However, the profitability fell sharply because the market did not grow proportionally. In order to improve the situation, Taiwan government actively pursues policy to encourage merging and acquisitions for the financial institutions. As a result, the number of private banks decreased to be 38 in year 2009. Financial liberalization and globalization are the goals of the government to relax the establishment of private banks. However, the operation of he banks show problems due to the reasons of loosely corporance governance, illegally over loan and so on under the competitive environment. Furthermore, the Double-card debt crisis had a dramatic impact to Taiwan from Year 2005 to year 2007 that force the banking industry to address their risk management and operational issues. Unfortunately, in year 2008, the Global financial crisis due to the US subprime mortgage crisis arise which severely affect the wealth management business. As a result, both global and national banking environment became more competitive and difficult year 2008 to 2009. By using Taiwan National banks as the research samples, the algorithm of Artificial Neural Network (ANN) is employed to develop a financial early warning system (EWS) for management authority as a tool in order to detect, prevent and to react the financial crisis. The empiricl results show that the correction rate of identifying the problem banks is 100% for the sample data of the Double-card debt crisis in contrast to the the 57% for those of the global financial crisis. The lower correction rate of the the global financial crisis may attribute to the insufficient and undisclosure of the key information, such as various dereviatives products. Meanwhile, the results are expected to be improved if monthly data of the banks are available. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|