Style-Based VaR Model For Open-End Equity Funds In China

In recent years,as the government encourages the development of institutional investors, China's securities investment funds have a good chance to continuously develop and grow, and the development of open-end funds is particularly rapid.At the time when open-end fund is established,it also specifies its own investment style,in order to meet different investors' risk-income preferences.In operation,many funds' actual investment styles deviate from their claimed styles.With the development of the fund industry,investors have deepened their understanding of fund,from simply concerning about the fund's yield, to recognize that the risk of fund is equally important.Investors really need some way to measure the risk of fund to assist their investment decision-making.This paper first applies Sharpe's return-based style analysis model to monitor fund's actual investment style,and then use rolling-window technique to check the sustainability of fund's style.The empirical results showed that through March 2004 to March 2008,most funds' actual investment style deviated from their claimed style,and mostly funds shifted to be growth style from claimed value style.We also found that the fund's style were not constant through the period analyzed.Funds shift their styles according to the change of market conditions.Among 29 sample funds,there were 21 funds that changed their styles during the period analyzed.The average figure of changing style was 0.97 times per fund. During the period covered,about 85%funds' return could be attributed to return on a passive style portfolio and the funds' manager made the contribution to the left.Then we assess the fund's risk by applying the style-based VaR model.We divided the fund's risk into two compartments:the systematic risk --VaMR,and its specific risk--VaSR,also known as the fund manager's selection risk.Empirical results showed that the market risk VaMR composed a large part of fund's overall risk.Among 29 sample funds,VaSR_99%,1w made a contribution to the overall value-at-risk in the rang from 0.3% to 19%,with the mean 7.5%.As during the period studied,the market overall risk was rising,so all the sample funds' VaR were rising in the same period,but to different extent for different funds.It appeared that funds with small VaR were likely to have small increase in their VaR through the period studied.Fund's VaR did not differentiate among different styles.Funds with the same style could have very different VaR values.To some degree,funds with high VaR value tended to have higher return,but this tendency was not obvious.And funds with the same style could have very different yields.We applied the rolling windows technique to backtest the style-based VaR model.And exception rate was 3.6%,which was much higher than the significance level 1%we set.It was mainly due to rising market volatility during the period studied.As the volatility of CITIC Standard & Poor's style indices was constantly increase,and we gave equal weight to every observation to calculated the indices' standard deviationσ,which were again used to calculate the indices' VaR f_i^VaR.This method would underestimate current market risk while the market risk was constantly rising.So we suggested calculating the indices standard deviation by giving more weight to the more recent observations.