| Summary: | 碩士 === 國立成功大學 === 政治經濟研究所 === 90 === The purpose of this article is to develop a dynamic model for predicting problem banks in Taiwan. The dynamic predicting model is based on Theodossiou（1993）, which can strengthen the ability to detect the deterioration of bank's financial condition by incorporating information sequentially over time.
In this study, net worth ratio, overdue loan ratio, operating expenses ratio, pre-tax earning ratio, and loans to deposits ratio are adopted as indicator of bank's performance. We manifest the difference of time-series behavior of attribute variable for problem and non-problem banks, and then combine VARMA and Multivariate CUSUM model to construct a real-time monitor system of problem banks.
Empirical results show that the model would have discriminated those getting-worse banks from healthy banks and provided warning signals in early stage of degeneration. Therefore, the dynamic model is competent to predict problem banks in Taiwan.
|
|---|
