A BL-MF fusion model for portfolio optimization: Incorporating the Black–Litterman solution into multi-factor model

Abstract

We study a Black–Litterman and multi-factor (BL-MF) fusion model that integrates equilibrium expected returns and investor views information from the Black–Litterman framework with the return-factor correlation information captured in the multi-factor model. The optimal estimator derived from our model improves accuracy in estimating expected returns and covariance matrix. We build optimal portfolios using our BL-MF model and benchmarks, adhering to both standard and criteria tailored for capturing tail risk with non-normal return distributions. Out-of-sample tests show our BL-MF portfolios outperform various benchmarks, and robustness checks validate this performance advantage, regardless of changes in sub-period, estimation window length or data frequency.