Volatility of fund performance are used to indicate overall risk fund, and active risk is defined as the volatility of market’s benchmark index tracking error [5]. First, we use each fund and HS300 index of weekly return to calculate the fund relative to the HS300 index of the tracking error: ER_{i, t} = R_{i, t}-HS300_t then put the sample standard deviation of the tracking error as the initiative of funds in the first quarter risk measurement: VolExcessR300Yearl_i,T = Std (ER_i,t). Figure 1 shows the overall risk and active risk time series of sample funds during 2004-2012. It can be seen that the overall risk of fund in the 2008 financial crisis has a significant rise, while funds active risk in the whole sample period keep a relatively stable trend.

This paper set up the following model to examine the investors of capital flow and the previous fund risk bearing behavior of the relationship between:

Model (1), respectively using VolRawYearly overall risk and active risk VolExcessR300Yearly, the retunr Performance is expressed in the past 12 months the average fund return. The interpretation of the other control variables are given in part 2 of this article. If investors hold the denial manner to fund risk-taking behavior, the open-end fund investors will choose to vote with their feet, evacuation their investment funds, leading to fund as a net outflow fund, in the late before the corresponding Risk coefficient β₁ accords with negative. On the other hand, if the β₁ symbol is positive, it means that managers of risk-taking behavior can attract more investors’ money, so fund managers own reward will crease.

The estimates results of model (1) are given in Table 1. Overall, adopting both overall risk and active risk, the sign of β₁ isn’t negative. There has no minus means fund risk-taking behavior at least not lead investors to pull away. Further observation found that active risk VolExcessR300Yearly coefficient is 0.6726, and highly significant

at the 1% level. From the angle of economic significance, in the case of other factors unchanged, fund manager at every 1% increase in the risk actively, will bring 0.67% new capital of investors that are also highly significant in terms of economy. It suggests that the fund initiative risk-taking behaviour is indeed brought better returns for fund managers, their risk-taking is worth. Observing other control variables, the fund performance performance coefficient Performance is positive, suggests that investors prefer to good fund performance; Fund size TNA and fund age Age coefficients is negative and highly significant, showed that fund investors prefer to choose the smaller and younger fund. These results are consistent with previous literature research conclusion, also conform to the general expected.

Next, in this paper, the fund risk is decomposed into systemic risk and idiosyncratic risk, further investigating the relationship between the fund risk and fund Managers Reward. Figure 2 shows the sample fund time series of systemic risk and idiosyncratic risk. It can be seen, idiosyncratic risk during the sample period on fund performance is stable, and always maintain in low level, while the level of fund systemic risk in the period 2008-2009 has a significant rise, which corresponds with the 2008 global financial crisis, China’s financial market systemic risk has a substantial rise.

The regression model is still based on the model (1), risk Risk here is decomposed into systemic risk and idiosyncratic risk. Among them, according to the single factor and three factors of performance differently, one factor SysRiskF1 systemic risk and systemic risk and systemic risk SysRiskF3 three factor two ways to measure, idiosyncratic risk using single factor idiosyncratic risk IdoRiskF1 idiosyncratic risk and three factors IdoRiskF3 said in one of two ways. The empirical results given in Table 2, the systemic risk in the corresponding model I

and model II, risk Risk coefficient were not significant. In model III and IV, idiosyncratic risk coefficients are negative at a significant at 5% significance level, show that the fund manager’s choice of idiosyncratic risk behavior is not recognized with investors. Investors choose fund idiosyncratic risk to vote with their feet, a significant net outflow of funds, the fund’s choice of idiosyncratic risk behavior will bring negative returns to fund managers. And in terms of systemic risk, fund manager rewards show the relationship between positive ratio and risk, but the relationship doesn’t get statistical test scores. The other control variables of regression results in Table 2 are similar to Table 1, and conform to the general expected, so we don’t describe specificly.