Introducing bounded rationality in a standard consumption-based asset pricing model with 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...