QuarcHome
Model catalogueBack to the model index

Black–Litterman

Blends portfolio-implied expected returns with views and uncertainty about those views.

How it works

Black–Litterman starts from market-implied equilibrium expected returns and updates them using investor views and uncertainty about those views. A relative view can compare two assets or baskets. Confidence determines how strongly the view shifts the expected returns, which are then inputs to portfolio optimisation.

Worked example

In a one-dimensional illustration, suppose a prior return estimate is 4% and a view is 6%, both with equal independent uncertainty variance. The Gaussian update gives 5%. Doubling the precision of the view relative to the prior moves it to about 5.33%. In a portfolio update, covariance can carry part of the adjustment to related assets.

Validation and limitations

Record the benchmark weights, risk-aversion assumption and each view’s mapping and confidence. Stress the confidence matrix and constrain turnover or leverage where applicable. View uncertainty should reflect actual forecasting evidence. Compare mean–variance optimisation for the portfolio step.

Research sources