The Global Portfolio Optimization approach of Black and Litterman (1992) combines economic models with statistical learning and is popular because it generates estimates of expected returns that are plausible in many situations.

The technique departs from the assumption that the market is a mean-variance portfolio implied by the CAPM equilibrium model, and builds on the fact that the observed market capitalization can be considered as optimal weights assigned by the market. Market weights reflect market prices that, in turn, embody the market’s expectations of future returns.

Hence, the approach can reverse-engineer the unobservable future expected returns from the assumption that the market is close enough to equilibrium, as defined by the CAPM, and allow investors to adjust these estimates to their own beliefs using a shrinkage estimator. The model can be interpreted as a Bayesian approach to portfolio optimization. We will introduce Bayesian methods in Chapter 9, Bayesian Machine Learning.