Loading...
Please wait, while we are loading the content...
Similar Documents
Using BS-PSD-LDA approach to measure operational risk of Chinese commercial banks
| Content Provider | Semantic Scholar |
|---|---|
| Author | Wang, Zongrun Wang, Wuchao Chen, Xiaohong Jin, Yan-Bo Zhou, Yanju |
| Copyright Year | 2012 |
| Abstract | The research of operational risk management among Chinese commercial banks is still in its preliminary stage. Operational risk events are rare and data is hard to collect. This leads to very small data samples. Besides, a large number of empirical researches show that the distributions of operational losses are often skewed with fat tails. To address these issues, this paper puts forward a loss distribution approach (LDA) based on bootstrap sampling and piecewise-defined severity distribution (BS-PSD-LDA). The approach divides data samples into two distinct parts (high-frequency low-severity losses and low-frequency high-severity losses), and fits the two parts by lognormal distribution and Generalized Pareto distribution respectively. Using hand-collected samples of 426 operational losses in Chinese commercial banks during 1994–2010, we estimate the magnitude of operational losses using the BS-PSD-LDA method. We show that our method provides a better fit than the traditional parametric methods. Besides, the method using historical simulation of nonparametric method seems to offer a good fit to the sample as well. However, we believe that the BS-PSD-LDA approach offers improvement from the perspective of satisfying risk control requirement of the regulatory authority and ensuring the efficiency of funds' utilization. |
| Starting Page | 2095 |
| Ending Page | 2103 |
| Page Count | 9 |
| File Format | PDF HTM / HTML |
| DOI | 10.1016/j.econmod.2012.06.031 |
| Alternate Webpage(s) | http://isiarticles.com/bundles/Article/pre/pdf/50057.pdf |
| Alternate Webpage(s) | https://doi.org/10.1016/j.econmod.2012.06.031 |
| Volume Number | 29 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
