Long Horizon Predictability: An Asset Allocation Perspective.
43 Pages Posted: 7 May 2014 Last revised: 6 May 2018
Date Written: April 30, 2018
Abstract
We analyze the effects of asset return predictability at various horizons on an individual's portfolio strategy and welfare gains as measured by a certainty equivalent return rate, for long term investors. We use a method to account for long horizon predictability that does not make violence to the data, and two alternative OLS procedures that allow investors to capture the differential information contained in various period returns. More specifically, our second procedure exploits the information present in the term structure of "forward" equity risk premia. We show that, adopting this procedure, the investor's welfare gain may be substantial relative to that obtained from using short horizon predictability only. Consequently, investors are better off by simultaneously using information in short and long horizon returns.
Keywords: dynamic portfolio decision, predictive regression, long horizon predictability, GMM estimation, inter-temporal hedging, continuous time
JEL Classification: G14
Suggested Citation: Suggested Citation