Stock-Bond correlation term structure

There has been so much discussion on the correlation between stocks and bonds that I was surprised that this topic has missed something fundamental. The stock bond correlation will differ by the maturity of the bonds analyzed. That is, there is a term structure of correlation. The stock bond correlation for the long bond will vary from that for the intermediate and short-term bonds. This seems obvious given there is a duration difference in bonds, yet given that interest rates across the yield curve will not always show a parallel move, the specifics for this relationship have to be analyzed. 

First, there is a difference in high and low-frequency stock-bond covariance. Second, the average covariance was upward-sloping during the Pre-GFC period, only to be downward-sloping during the post-GFC period. Third, there is a change in the relationship between stock and bond returns. It used to be upward sloping and is now downward sloping, and finally, there is a substantial change in the cyclical covariance relationship. This is all described in the paper, "The Term Structure of Stock-Bond Covariances." 

These changes are linked with the change in quantitative easing, yet there needs to be more work on why there would be such significant changes in the stock-bond covariance relationships.