New results on the predictive value of crude oil for US stock returns

The purpose of this study is to clarify the nature of the predictive relationship between crude oil and the US stock market, with particular attention to whether this relationship is driven by time-varying risk premia.

The authors formulate the predictive regression as a state-space model and estimate the time-varying coefficients via the Kalman filter and prediction-error decomposition.

The authors find that the nature of the predictive relationship between crude oil and the US stock market changed in the latter half of 2008. After mid-2008, the predictive relationship switched signs and exhibited characteristics which make it much more likely that the predictive relationship is due to time-varying risk premia rather than a market inefficiency.

The authors apply a state-space approach to modeling the predictive relationship. This allows one to watch the evolution of the predictive relationship over time. In particular, the authors identify a dramatic shift in the relationship around August 2008. Prior research has not been able to identify shifts in the relationship.