A Forecast Combination Approach to Equity Factor Timing

We investigate the benefits of forecast combination for timing equity factors based on predictive regressions using macro predictors. Relative to standard predictive regression models, forecast combination reduces the noise of forecasts and hence improves their out-of-sample predictive accuracy. Given the nature of macro predictors, the ensuing dynamic model reacts when major macro events happen. Before transaction costs, portfolio simulation results show considerable outperformance of the factor timing model over a static factor allocation. But much of this performance wedge is eroded when transaction costs are taken into account, rendering this article a cautionary tale about the benefits of factor timing