Working Paper: NBER ID: w18305
Authors: Ravi Bansal; Dana Kiku; Amir Yaron
Abstract: The long-run risks (LRR) asset pricing model emphasizes the role of low-frequency movements in expected growth and economic uncertainty, along with investor preferences for early resolution of uncertainty, as an important economic-channel that determines asset prices. In this paper, we estimate the LRR model. To accomplish this we develop a method that allows us to estimate models with recursive preferences, latent state variables, and time-aggregated data. Time-aggregation makes the decision interval of the agent an important parameter to estimate. We find that time-aggregation can significantly affect parameter estimates and statistical inference. Imposing the pricing restrictions and explicitly accounting for time-aggregation, we show that the estimated LRR model can account for the joint dynamics of aggregate consumption, asset cash flows and prices, including the equity premia, risk-free rate and volatility puzzles.
Keywords: Long-run risks; Asset pricing; Time aggregation; Consumption dynamics
JEL Codes: E21; G12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| risk aversion (D81) | asset pricing (G19) |
| shocks to expected growth in consumption (E20) | long-lasting effects on consumption dynamics (E21) |
| ignoring time aggregation (C43) | substantial biases in estimates of risk aversion (D91) |
| decision interval of agents (D79) | temporal aggregation of data (C41) |
| LRR model (C59) | captures cross-sectional variations in asset returns (G11) |
| LRR model (C59) | replicates failure of CAPM (C59) |