Black-Litterman Model
The Black-Litterman model is a portfolio construction framework that combines equilibrium market returns (derived from global market capitalisation weights) with an investor's subjective views using Bayesian updating to produce a blended, stable expected return vector.
The Black-Litterman (BL) model, developed by Fischer Black and Robert Litterman at Goldman Sachs in 1990, addressed a core problem with classical Markowitz mean-variance optimisation: tiny changes in expected return inputs produce wildly different portfolio weights, often generating extreme and unintuitive positions. BL solves this by anchoring expected returns to market equilibrium (implied by CAPM) and then adjusting them only to the extent that the investor holds high-confidence views.
The starting point is the equilibrium return vector, derived by reverse-optimising from observed market cap weights. If Nifty 500 stocks constitute the investment universe, the equilibrium implied returns are extracted by asking: what expected returns, fed into a mean-variance optimiser, would reproduce the observed market cap weights as the optimal portfolio? This imposes market consensus as the neutral prior.
The investor then specifies views: absolute (e.g., Infosys will return 14% annually) or relative (e.g., banking sector will outperform IT sector by 2%). Each view is accompanied by a confidence level (expressed as its variance). Bayes theorem blends the prior (equilibrium returns) and the likelihood (views weighted by confidence) to produce a posterior expected return vector. This posterior is then fed into a standard mean-variance optimiser to derive portfolio weights.
The output is significantly more stable and intuitive than pure Markowitz. Portfolio weights shift from equilibrium weights only in the direction of active views, and the magnitude of shift scales with view confidence. This prevents the optimiser from making large unintended bets in areas where no view exists.
In India, institutional fund managers at large AMCs, insurance companies, and sovereign funds applying factor or macro investing frameworks have incorporated BL-style overlay to tilt sector weights. However, full BL implementation requires quantitative infrastructure — covariance matrix estimation, Bayesian solvers — limiting its formal use to larger institutions. Several PMS managers and category III AIFs with quantitative mandates used BL-inspired blended return forecasts in their portfolio construction process.