DCC-MIDAS modelWithout exogenous variables, we first examine the long-term correlation between the stock and bond markets affected by therealized correlation (RC). Table 1 shows the results of estimating the DCC-GARCH model for the stock and bond market returns.Obviously, θ is significant for the realized volatility (RV) and realized correlation (RC). This is a baseline model which does notincorporate investor's sentiment.From Table 1, we can find that the results in the table show that almost all parameters are significant. The estimates of α and β taketypical values. The sign of θv indicates the response of long-run volatility to RVs. Another feature of the GARCH–MIDAS model in Table 1is that the sums of α and β are 0.9808 and 0.9915 for the GARCH–MIDAS of stock and bond markets volatility. These numbers arenoticeably less than but close to 1, which implies high persistence of financial volatility. As DCC-MIDAS is concerned, the estimatedparameters of a and b are 0.0653 and 0.8522, whose sum is also less than 1. The results of the DCC-MIDAS dynamic correlation estimation,ωc reflects how the stock-bond correlation responds to RC shocks. We find a statistically insignificant but positive ωc, suggestingthat the impact of RC may be misspecified. Different results can be found in Table 2 where investor's sentiment is involved.