Kinlaw, Kritzman & Turkington [2015] continues:
"We test this hypothesis by analyzing the returns of 10 U.S. sectors from December 1978 through December 2013. Each month, we regress the cross-section of monthly, three-year, and ten-year sector returns on the cross-section of changes in earnings during the same periods. We also regress the cross-section of monthly, three-year, and ten-year sector returns on...a proxy for sector discount rates." The empirical results "reveal rather starkly that high-frequency returns are more strongly related to changes in discount rates than to changes in earnings, whereas the opposite is true for relatively lower-frequency returns."