Adaptive Expectations and Commodity Risk Premia
48 Pages Posted: 28 May 2016 Last revised: 9 Mar 2018
Date Written: March 5, 2018
We investigate the determinants of the commodity (ex-ante) risk premia for different maturities through the lens of a model of adaptive learning in which expected future spot prices are revised based on past prediction errors and changes in economic fundamentals. The main results show that risk premia are highly time varying and their dynamics is predominantly driven by hedging pressure and time-series momentum, conditional on a set of common predictors. Cumulative impulse-response functions from a panel VAR model show that these effects are persistent and not short-lived. Further, we provide evidence that the ex-ante spot premia is positively (negatively) correlated with the variance (skewness) of past realized returns, consistent with existing theoretical evidence. Finally, we show that adaptive expectations are broadly consistent with the cross-sectional average of a subset of Bloomberg professional analysts' forecasts.
Keywords: Commodity Markets, Adaptive Expectations, Empirical Asset Pricing, Hedging Pressure, Time-Series Momentum
JEL Classification: G12, G17, E44, C58
Suggested Citation: Suggested Citation