Black–Litterman portfolio allocation
Start with market-implied equilibrium returns, add absolute or relative views, and assign a confidence to each one. The posterior allocation updates in real time so you can see exactly how each assumption changes the portfolio.
| Asset | Annual volatility | Market weight | Implied return | Posterior return | Target weight | Active tilt |
|---|---|---|---|---|---|---|
| US Equity | 16.00% | 40.00% | 3.70% | 7.03% | 21.94% | -18.06% |
| Intl Equity | 17.00% | 18.00% | 3.60% | 6.25% | 3.98% | -14.02% |
| Treasuries | 6.00% | 20.00% | 0.04% | 2.14% | 58.51% | +38.51% |
| Credit | 8.00% | 10.00% | 1.15% | 1.67% | 2.21% | -7.79% |
| Gold | 13.00% | 6.00% | 0.47% | 2.91% | 12.04% | +6.04% |
| Commodities | 18.00% | 6.00% | 1.04% | 2.41% | 1.33% | -4.67% |
Returns are annualized excess returns. The headline portfolio return adds the risk-free rate to the weighted posterior return.
Π = δΣwmkt · market-implied excess returns
μBL = Λ⁻¹[(τΣ)⁻¹Π + PᵀΩ⁻¹Q] · posterior expected excess returns
ΣBL = Σ + Λ⁻¹ · posterior covariance, where Λ = (τΣ)⁻¹ + PᵀΩ⁻¹P
wBL = normalize[(δΣ)⁻¹μBL] · target portfolio weights
Market equilibrium, refined by your views.
Black–Litterman combines a market-implied return prior with the views you choose to express. The model derives its starting point from market-cap weights, then blends in absolute or relative views in proportion to their confidence. Those posterior returns feed the allocation, making each portfolio tilt traceable while keeping equilibrium as the reference.
Model workflow
- Review the assets, market weights and volatilities. Together they define the portfolio used to establish the market equilibrium prior.
- Set risk aversion (delta), prior uncertainty (tau) and the risk-free rate. These assumptions shape the implied returns and reported portfolio metrics.
- Inspect the reverse-optimized returns: the excess returns that make the market-weighted portfolio optimal under the selected risk model.
- Add absolute return views or relative spread views. A relative view compares two assets without assigning either one a standalone forecast.
- Assign a confidence to each view. The model translates confidence into view uncertainty and blends the views with the equilibrium prior.
- Compare prior and posterior returns, then review the resulting weights, expected return, volatility and Sharpe ratio.
Core equations
Reference: Black & Litterman (1992), "Global Portfolio Optimization," Financial Analysts Journal; He & Litterman (1999) on calibrating the view-uncertainty matrix; Idzorek (2007) on confidence-based Omega.
A worked allocation
Six-asset allocation with three views (default inputs)
- Assets and market weights: US Equity 40%, Intl Equity 18%, Treasuries 20%, Credit 10%, Gold 6%, Commodities 6%; volatilities in the same order are 16%, 17%, 6%, 8%, 13% and 18%.
- Risk settings: delta 2.5, tau 0.05 and a 2% risk-free rate.
- Views: US Equity at 8.5% (72% confidence), Treasuries at 3.5% (66%), and Gold outperforming Treasuries by 2.5% (58%).
- 1
Reverse optimization estimates equilibrium excess returns from risk aversion, covariance and market weights. Here, US Equity starts at 3.70% and Treasuries at 0.04%.
- 2
The three views move posterior returns to 7.03%, 2.14% and 2.91% for US Equity, Treasuries and Gold respectively.
- 3
Converting the posterior through the covariance matrix produces weights led by Treasuries (58.5%), US Equity (21.9%) and Gold (12.0%). The other assets account for about 7.6%.
- 4
At these inputs, expected return is 5.46%, volatility is 5.54% and the Sharpe ratio is 0.63.
5.46% expected return · 5.54% volatility · 0.63 Sharpe. Posterior excess returns range from 1.67% for Credit to 7.03% for US Equity.
Reading the posterior
Posterior returns are a weighted compromise between the market-implied prior and your views. Higher confidence gives a view more influence; lower confidence keeps the posterior closer to equilibrium. Portfolio weights reflect both those returns and the covariance structure, so low-volatility assets can receive large allocations even when their expected premium is modest. Treat concentration as a consequence of the chosen assumptions and unconstrained optimizer, and compare it with your own allocation limits before acting on it.
Model assumptions
- The model treats the market-implied equilibrium returns as a useful prior and assumes returns and views can be represented by the selected covariance structure.
- Covariance is built from the entered volatilities and a stylized correlation matrix. It is not estimated from a specific return history.
- View uncertainty (Omega) is diagonal, derived from confidence, and treats views as independent. Related or overlapping views may therefore be represented imperfectly.
- Weights use an unconstrained mean-variance solution normalized to sum to one. The model does not enforce long-only, position-size or turnover limits.
- Risk aversion (delta) scales equilibrium premia; tau scales prior uncertainty and how much views can move the posterior. Results are sensitive to both.
Common questions
What problem does Black-Litterman solve?
It provides a structured way to combine market-implied expected returns with investor views. Starting from equilibrium returns can reduce the sensitivity of a mean-variance allocation to standalone forecasts, while confidence makes the influence of each view explicit. It does not remove estimation risk or replace portfolio constraints.
How do absolute and relative views differ?
An absolute view assigns a return level to one asset, such as US Equity at 8.5%. A relative view specifies a difference between assets, such as Gold outperforming Treasuries by 2.5%, without forecasting either return on its own. The pick matrix (P) and view vector (Q) encode both forms.
Why is so much of the portfolio in Treasuries?
The unconstrained solution scales posterior returns by the inverse covariance. Treasuries have the lowest entered volatility (6%) and a positive posterior, which can produce a large weight. The result depends on the model inputs and does not apply long-only, position-size or turnover limits.
What are implied equilibrium returns?
They are the excess returns that make the selected market-weighted portfolio optimal under the model's risk assumptions. Reverse optimization calculates them as Pi = delta × covariance × market weights. They form the prior that investor views update.
What does the confidence on a view do?
Confidence controls the uncertainty assigned to a view through Omega. Higher confidence means lower uncertainty and more influence on posterior returns; lower confidence leaves them closer to the equilibrium prior. The effect also depends on the view structure, prior uncertainty and covariance.
What do the portfolio return, volatility, and Sharpe ratio represent?
Expected return is the portfolio-weighted posterior return plus the risk-free rate. Volatility is the portfolio standard deviation implied by the covariance matrix. The Sharpe ratio is excess return divided by volatility. At the default inputs, those values are 5.46%, 5.54% and 0.63 respectively.
Model notes last reviewed 2026-10-06. For research and education only. See our disclaimer.