Sticky Expectations and Stock Market Anomalies

We propose a simple model in which investors price a stock using a persistent signal and sticky belief dynamics à la Coibion and Gorodnichenko (2012). In this model, returns can be forecasted using (1) past profits, (2) past change in profits, and (3) past returns. The model thus provides a joint theory of two of the most economically significant anomalies, i.e. quality and momentum. According to the model, these anomalies should be correlated, and be stronger when signal persistence is higher, or when earnings expectations are stickier. Using I/B/E/S data, we measure expectation stickiness at the analyst level. We find that analysts are on average sticky and, consistent with a limited attention hypothesis, more so when they cover more industries. We then find strong support for the model's prediction in the data: both the momentum and the quality anomaly are stronger for stocks with more persistent profits, and for stocks which are followed by stickier analysts. Consistently with the model, both strategies also comove significantly.