Working Paper: CEPR ID: DP6518
Authors: Klaus Adam; Albert Marcet; Juan Pablo Nicolini
Abstract: Introducing bounded rationality into a standard consumption based asset pricing model with a representative agent and time separable preferences strongly improves empirical performance. Learning causes momentum and mean reversion of returns and thereby excess volatility, persistence of price-dividend ratios, long-horizon return predictability and a risk premium, as in the habit model of Campbell and Cochrane (1999), but for lower risk aversion. This is obtained, even though we restrict consideration to learning schemes that imply only small deviations from full rationality. The findings are robust to the particular learning rule used and the value chosen for the single free parameter introduced by learning, provided agents forecast future stock prices using past information on prices.
Keywords: Asset Pricing; Puzzles; Consumption-based Asset Pricing; Learning
JEL Codes: D84; G12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| learning (C91) | momentum in stock prices (G10) |
| increased expectations about stock price growth (G17) | actual price growth exceeds fundamental growth rates (E31) |
| optimistic beliefs about price growth (D84) | higher actual price growth (E30) |
| learning (C91) | excess volatility in stock prices (G17) |
| learning about stock prices (G12) | persistence in price-dividend ratios (G35) |
| learning (C91) | deviations from the mean (C46) |
| learning model (C52) | ability to match empirical observations (C12) |
| learning model (C52) | excessive volatility of stock returns (G17) |
| learning model (C52) | predictability of excess returns over long horizons (G17) |