Two analysts run MVO with identical asset classes and identical constraints. One estimates U.S. equity returns at 8.5%, the other at 9.0%. The 50-basis-point gap (a rounding error) shifts the equity allocation from 35% to 58%. That is the core fragility the curriculum's improvements exist to fix.
LEARNING OUTCOME STATEMENTS (CFA Institute): describe and evaluate the use of mean-variance optimization in asset allocation; recommend and justify an asset allocation using mean-variance optimization; interpret and evaluate an asset allocation in relation to an investor's economic balance sheet; recommend and justify an asset allocation based on the global market portfolio; discuss the use of Monte Carlo simulation and scenario analysis to evaluate the robustness of an asset allocation; discuss asset class liquidity considerations in asset allocation; explain absolute and relative risk budgets and their use in determining and implementing...
Investors want more expected return but penalize variance in proportion to how much they personally dislike risk. Markowitz formalized this trade-off by treating variance as the proxy for risk, letting any investor trace a frontier of best-possible return...
Common mistakes
- Using arithmetic mean for multi-period projections. The arithmetic mean overstates compound growth. Geometric mean = arithmetic mean minus one-half variance. A 10% arithmetic return with 20% volatility compounds at roughly 8%. Candidates who project terminal wealth using the arithmetic mean overstate outcomes.
- Forgetting that reverse optimization produces implied returns, not forecasts. Implied returns tell you what the market prices in. They are a starting point, not a recommendation. Using them without overlaying views produces a passive market-weight portfolio.
- Treating Black-Litterman views as free. Every view shifts the allocation. A wrong view with high confidence produces a large, wrong tilt. The model does not validate views, it implements them faithfully.
Bottom line
- MVO is an "error maximizer": a tiny input change (e.g. +50 bps to one return) produces extreme, concentrated allocations
- Corner portfolios define where asset classes enter/exit the frontier; any efficient portfolio is a blend of two adjacent corners
- Reverse optimization extracts implied returns from market-cap weights, avoiding estimation error; the global market portfolio is the no-view baseline
- Black-Litterman blends implied equilibrium returns with confidence-weighted investor views, producing stable, intuitive allocations
Exam shortcut
When the question asks what is wrong with an MVO output, the answer is almost always "estimation error in expected returns." When Black-Litterman appears, check whether the view is expressed as relative or absolute and whether confidence is high or low, these determine the magnitude of the tilt. For risk budgeting calculations, ACTR = weight x beta x portfolio vol. Sum of ACTRs = total portfolio risk.
The full lesson (about 6,750 words, 45 min read) adds 2 worked examples, all 12 common mistakes, a self-check, free in the app.
Learning objectives
- principles
Browse all free CFA L3 Private Markets lessons or jump into free CFA L3 Private Markets practice questions.