Value-at-Risk Models Applied to Taiwan's Stock Market

In this paper, the parametric normal method, the historical simulation method and the Monte Carlo simulation method are applied to Taiwan's stock marekt estimating the one-day 95% and 99% VaRs for the electronic and banking & insurance sector indices. Then, the basic frequency backtesting and the conditional coverage testing are used to examine the performances of these VaR models. We find that all the approaches in question generate the accurate 95th percentile risk measures for both sector indices, while only the 250-day and 1000-day historical simulation methods produce the reasonable 99th percentile risk measures for both sector indices