Abstract: We had an empirical test of the relationship between the realized volatility and tracking error of the ETF with high frequency data. We construct the measurement method for the realized volatility and tracking error of the ETF in Chinese stock market with high frequency data. The two kinds of tracking error are constructed by the absolution value and standard deviation method respectively. The statistical character, Ljung-Q and ADF value of the three kinds of series are analyzed. It has a demonstration that the realized volatility and tracking error of the ETF in Chinese stock market has causality relationship, the latter is the cause for the former. Furthermore, with the cross correlation of the realized volatility and tracking error of the ETF in Chinese stock market, it concludes that the series of the realized volatility and tracking error of the ETF have correlation relationships, and the former is lagged the latter appreciably, and it demonstrateds their causality relationship in the paper. By the method of VAR and impulse response function, it shows that the volatility of the tracking error of the ETF in Chinese stock will has effect on the realized volatility, which has the characteristic of promoter and persistency.

Key words: high-frequency, fund, tracking error, realized volatility