A nested var bootstrapping with fat tail correction for equity portfolio in Taiwan

Abstract

Value at risk (VaR) measurement has shown increasing promise after explosions of financial disasters in developed and emerging markets. VaR is of special necessity for tradings in Taiwan stock market after experiencing the land mine firms events. Though there are plenty of VaR estimation models, historical simulation seems to be a popular technique for the implementation of such a risk measurement approach. However, a main problem revealed in the historical VaR simulation is the potential for up biased estimation caused by the fat - tailed return distribution. Accordingly, this paper proposes a nested VaR bootstrapping with fat - tail correction trying to improve the effectiveness of VaR estimation in the sense of uncoverage ratio. Empirical investigation is performed using the electronic and financial portfolios constructed from the most active traded industries in the Taiwan stock market. The approximate optimal in-sample period is analyzed and the nested VaR bootstrapping with fat- tail correction is evaluated in comparison to the uncorrected one. The empirical results indicate the fat- tail phenomenon in the Taiwan stock market and show the advantage of our fat- tail corrected VaR bootstrapping model. The main contribution of this paper might be the successful design of the fat- tail correction together with the nested VaR bootstrapping.