Mean–variance optimisation
Choosing asset weights that give the highest expected return for each level of volatility. Highly sensitive to its return inputs, so without constraints it tends to produce concentrated, unstable allocations.
Where you will meet it
Mean–variance optimisation appears in 3 learning outcomes on the CFA syllabus. A few of them:
- Weigh the return potential and risks of private versus public market exposure within a long-term strategic allocation
“Reported volatility and correlations are smoothed, so unsmooth them before any mean-variance optimisation, or the model will overweight private assets.”
- Explain how mean–variance optimisation is used to set an asset allocation and assess its strengths and weaknesses
“Mean–variance optimisation uses expected returns, volatilities and correlations to trace the efficient frontier, then selects the portfolio that maximises utility, often written Um = E(Rm) − 0.005 × λ × σm² with returns in percent and λ the risk aversion coefficient.”
- Discuss the methods available for setting the weight of alternatives in an asset allocation
“Mean–variance optimisation can include alternatives, but inputs must be adjusted: unsmooth appraisal-based returns, include illiquidity, and cap allocations, since otherwise the optimiser overweights assets with understated risk.”
Practise it
Practise mean–variance optimisation until it is automatic: questions that explain every wrong option.
Take the free diagnostic →