The Efficient Horizon of Expectation and Stock Prices

Investors' expectations on firms' cash flow growth can be biased (e.g. Bordalo et al. (2019)), yet we know little about how these biases and their asset pricing implications vary with forecast horizons. In this paper, I show that extreme expectations at all horizons beyond the current period inversely forecast abnormal stock returns, but some with a delay --- extreme expectations at long-horizons persist until they reach the imminent horizon, causing persistent mispricing. Consistent with managers' expectation management altering the efficient expectation horizon, the pattern is stronger after Regulation-FD. A model based on ``natural expectation'' by Fuster et al. (2010) generates the short- and long-horizon forecast error dynamics that match the empirical patterns. Surprisingly, this extrapolative belief model also predicts underreaction.