It is well known that production-based models face a greater challenge in explaining the fluctuations in asset returns due to endogenous consumption and dividend smoothing. Researchers have long realized that relaxing the assumption of perfect rationality is a potential way to explain patterns observed in both macroeconomic and financial time series. Recently, Fuster, Laibson, and Mendel (2011) argue that quasirational models deserve greater attention.
In this paper, following their advocate, we depart from the perfectly rational expectations framework and introduce extrapolative expectations into a standard dynamic stochastic general equilibrium (DSGE) model featuring recursive preferences. We show that our model greatly improves upon traditional rational models in matching both quantities and prices.
Extrapolation biases are pervasive in human behavior, and there is abundant evidence from both the psychology and nance literature (see, e.g., Hirshleifer (2001), and Fuster, Laibson, and Mendel (2011)). In controlled experiments on human subjects, Tversky and Kahneman (1974) document that individuals' decisions are influenced by various heuristics. The representativeness heuristic is a form of non-Bayesian updating where individuals tend to overweight recent observations and ignore the underlying laws of probability. For example, when investors see that a company displays a series of recent high earnings growth, they may classify this company as a growth rm and ignore the probability that very few companies can keep growing.
Using both survey and experimental data, De Bondt (1993) shows that the forecasts of individual investors adhere to a simple trend-following methodology. Vissing-Jorgensen (2003) fi nds survey evidence of extrapolative expectations even among wealthy investors. In particular, investors who have experienced high portfolio returns in the past expect higher returns in the future. Hong, Stein, and Yu (2007) suggest that individuals tend to gravitate toward simple models when making forecasts. Finally, Fuster, Laibson, and Mendel (2011) suggest that extrapolation is important for understanding macroeconomic fluctuations.