We demonstrate that the estimates of the Capital Asset Pricing Model (CAPM) parameters significantly differ across samples, which are based on different days of the week (representing different seasons). Our evidence suggests that the “noise” in the data is not an issue. We also show that...