以區別分析法預測銀行倒閉

Abstract

The financial crisis of the late 2000s caused hundreds of bank failures in the United States and around the globe. Consequences also included a large amount of money spent by governments around the world to bail out countless banks, the worldwide stock market crash, and the significant global economy decline. While the financial crisis seems to come to an end, it remains questionable if more banks would fail in the near future. This paper tries to find the characteristics of banks’ performance, indentify the differences between failed and existing banks, and employ the discriminant analysis technique to derive functions to classify banks into two groups-failed and existing banks. The comparison of financial ratios of two groups of banks indicates remarkable differences between them. The level of importance of financial ratios to the prediction of bank failures is indentified. The discriminant functions exhibit satisfactory predictive power of bank failures.

The financial crisis of the late 2000s caused hundreds of bank failures in the United States and around the globe. Consequences also included a large amount of money spent by governments around the world to bail out countless banks, the worldwide stock market crash, and the significant global economy decline. While the financial crisis seems to come to an end, it remains questionable if more banks would fail in the near future. This paper tries to find the characteristics of banks’ performance, indentify the differences between failed and existing banks, and employ the discriminant analysis technique to derive functions to classify banks into two groups-failed and existing banks. The comparison of financial ratios of two groups of banks indicates remarkable differences between them. The level of importance of financial ratios to the prediction of bank failures is indentified. The discriminant functions exhibit satisfactory predictive power of bank failures.