| Summary: | 碩士 === 實踐大學 === 企業管理研究所 === 92 === Abstract
Our financial institution''s share of non-performing loans (NPLs) is growing. According to statistics released by the Finance Bureau of the Finance Department, up to the end of September 2003 the ‘NPL ratio’stood at 5.62% of assets. A worsening record of defaulting borrowers suggests bad management..
Our research is based on real life examples of bank lending. Factor Analysis, Discriminate Analysis, and Mahalanobis Analysis are the statistical/analytical tools used to assess a borrower''s creditworthiness.
There are two stages. Data is gathered from "experiment group" and "contrast group". The first stage is to develop a risk-assessment model by using data from the experiment group, which is processed by Factor Analysis, Discriminate Analysis and Mahalanobis Analysis. The second stage takes data from the contrast group. Again, Factor Analysis, Discriminate Analysis and Mahalanobis Analysis are used to determine factor scores. This two-pronged method is used to find out not only what the ratio of bad customer debts is, but also whether the banking sector is managing its loan portfolio with clarity and credibility.
Research has shown that of the 19 variables used by banks to assess risk, the preferred system of credit-approval uses a combination of Factor, Discriminate and Mahalanobis analyses. This method is also sound, resulting in a 97.5% hit ratio, and remains best way to determine whether a borrower will default on his contractual obligations. Moreover, the model helps strengthen the quality of bank lending by increasing the efficiency and lowering the costs of allocating capital. Not least, the model acts as a warning signal to businesses with unhealthy balance sheets. Therefore, precautionary measures should be taken.
|
|---|
