Dynamic Hedging, Expected Returns and Factor Timing

A simple equilibrium restriction identifies time-varying components of expected returns. Using data from the Fama and French threeand fivefactor models and momentum, we use this restriction to predict factor returns out of sample. We derive a policy for timing exposures to these components and find the modified portfolio generates returns with higher out-of-sample Sharpe ratios than the underlying mimicking portfolio. Exposure to returns to the market timing portfolio is priced and the residual variation in the market is not. Gains from market, value and momentum timing policies are compensation for risk. We provide evidence that timing size exploits mis-pricing rather than precision of conditional risk exposures.