Mean-variance dominant trading strategies

•We investigate the relative importance of anomaly-based strategies from a Mean Variance standpoint.•We employ spanning methodologies in a classical unconditional setting and a novel conditional setting.•In the classical setting the majority of the anomalies is needed to maximize return to risk compensation.•In the conditional setting the same maximization is obtained by including only two of the ten anomalies.•The efficient use of conditioning information reduces diversification benefits across anomalies.

The paper examines the relative importance of ten anomaly-based trading strategies. We employ Mean Variance spanning methodologies in a classical unconditional setting and a novel conditional setting. Fixed-weight optimal portfolios stemming from the unconditional methodology indicate that all the strategies are needed to enhance the mean-variance tradeoff. This conclusion is completely reversed when we allow for time-varying portfolio weights as a nonlinear function of lagged economic indicators. The overall results suggest that diversified anomaly-based holdings are of limited benefit to sophisticated investors who employ dynamic trading strategies.