| This paper explores the possibility that investors' judgment biases, when observing patterns in data, help explain momentum in security returns. We find that investors tend to overweigh and extrapolate the recent patterns (trends) in returns and earnings. When the trend is broken, they do not update their expectations fully to reflect the implications of the new evidence and thus under-react. Momentum is strong when the historic trend in data is broken, but it is weak and short-lived when the trend is intact. Moreover, we observe price reversals for stocks with trending earnings innovation patterns or trending return paths, with much of the reversals coming from value stocks. Finally, we are able to refine naive price and earnings momentum strategies. |
