Fluctuation Reduction of Value-at-Risk Estimation and Its Applications

• Main Authors: CHIAN, GUAN-YI, 錢冠邑
• Other Authors: HUANG, SHIH-FENG
• Format: Others
• Language: en_US
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/e7ja5q

碩士 === 國立高雄大學 === 統計學研究所 === 105 === Value-at-Risk (VaR) is a fundamental tool for risk management and is also associated with the capital requirements of banks. Banks need to adjust their capital levels for satisfying the Basel Capital Accord. On the other hand, managements do not like to change the capital levels too often. To achieve a balance, this study proposes an approach to reduce the fluctuation of VaR estimation. The first step is to fit a time series model to the underlying asset returns and obtain the conventional VaR process. A new VaR (NVaR) estimation of the conventional VaR process is then determined by applying change-point detection algorithms and a proposed combination scheme. The return processes of 30 companies of S&P 500 from 2005 to 2016 are employed for our empirical investigation. The results of our in-sample investigation indicate that the capital levels computed from the NVaR process are capable of satisfying the Basel Accord and reducing the fluctuation of capital levels simultaneously. To calculate out-of-sample capital requirements, an innovative approach for one-step-ahead NVaR prediction is also proposed by incorporating the concept of CUSUM control charts. Numerical results indicate that the proposed one-step-ahead NVaR prediction is also capable of satisfying the Basel Accord and reducing the fluctuation of capital requirements simultaneously by using a comparable average amount of capital requirements to the conventional VaR estimator.