Establishment of Trading Strategies with Value-at-Risk Models

This study used value-at-risk (VaR) to construct four trading strategies and simultaneously applied seven indicators to assess the suitability of the model for Taiwan Eight industries Index. To estimate the VaR, we employed three models, including the exponentially weighted moving average (EWMA), generalized autoregressive conditional heteroskedasticity (GARCH), and the Monte Carlo model, to construct suitable model parameters. Our results showed that 1) the optimum VaR model for the expansion and contraction periods of the various share indices presented significant differences in suitability and model parameters; 2) the construction of trading strategies using downside VaR can be applied to the Taiwan Eight Industries Index; 3) a comparison of the three models regarding their suitability for the eight industries index indicated that EWMA (75%) and Monte Carlo (25%) are applicable to a bull market, and EWMA (37.5%), GARCH (37.5%), and Monte Carlo (25%) are applicable to a bear market; and 4) the proposed moving daily VaR (MDV) trading strategy shows that although the return rate of bull periods is lower than the return rate under the buy-and-hold (BH) strategy, the negative return rate of bear periods is significantly lower than the return rate under the BH strategy.