In this paper, we empirically examine the possibility that asset pricing models hold more closely at times of heavy marketwide trading. Our cross-sectional tests show that expected returns for portfolios and individual stocks are positively related to market beta at the turn-of-the-month, when institutions first sell shares to meet cash needs and investors subsequently buy shares. The expected return-beta relation is stronger when macroeconomic news is scheduled for release during this window. Periods of high market volatility and heavy trading are also followed by a positive expected return-beta relation. Our analysis suggests that market makers pay closer attention to the risk-return tradeoff when they must accommodate intense marketwide trading, causing asset pricing models to perform better.