What is the Black-Litterman model and why is it better than standard mean-variance optimization?

The Black-Litterman (BL) model solves the major practical problems of standard mean-variance optimization (MVO). Let me explain why it exists and how it works.

The problem with standard MVO:

1. Garbage in, garbage out: Small changes in expected return inputs cause huge changes in portfolio weights
2. Extreme positions: MVO often produces concentrated, unintuitive allocations (e.g., 80% in one asset, short positions)
3. Error maximization: MVO effectively maximizes the impact of estimation errors

Black-Litterman's elegant solution:

The BL model starts with equilibrium returns (what the market implies about expected returns based on current capitalization weights) and then blends in the investor's views to produce a modified set of expected returns.

Step-by-step intuition:

Step 1: Extract equilibrium returns

Using the global market portfolio (based on market cap weights), reverse-engineer the expected returns that would make rational investors hold those weights. This is the neutral starting point — if you have no views, you hold the market.

Step 2: Express views

The investor specifies views with confidence levels:

• Absolute: "I believe US equities will return 8% (with 60% confidence)"
• Relative: "I believe EM equities will outperform European equities by 3% (with 80% confidence)"

Step 3: Blend

The BL model uses Bayesian statistics to combine the equilibrium returns with the investor's views, weighted by their confidence. Higher confidence → views dominate. Lower confidence → equilibrium dominates.

Step 4: Optimize

The blended returns are fed into a standard MVO. Because the starting point is equilibrium and views are incorporated smoothly, the resulting weights are much more stable and intuitive.

Practical example:

Aurora Capital starts with global market cap weights:

• US Equities: 55%, Europe: 20%, EM: 10%, Bonds: 15%

View: "EM equities will outperform European equities by 3% over the next year" (70% confidence)

BL output: Tilt slightly toward EM (say 14%) and away from Europe (say 17%), with modest adjustments elsewhere. No extreme positions.

BL vs. standard MVO:

Exam tip: CFA Level II tests the conceptual understanding — why BL is preferred, how views are incorporated, and the role of equilibrium returns. You're unlikely to be asked to perform the full matrix algebra.

Dive deeper into portfolio management on AcadiFi's CFA Level II course.