Mean-variance optimization
Um = E(Rm) − 0.5 × λ × σm²
MVO chooses weights that maximize utility given expected returns, volatilities and correlations, typically with constraints (no short sales, weights sum to 1).
Criticisms:
- Highly sensitive to expected returns; outputs are concentrated.
- Ignores liabilities and the investor’s non-financial assets.
- Uses only mean and variance (ignores fat tails and skew).
- Single-period.
Improving MVO
- Reverse optimization: infer the returns implied by market-cap weights.
- Black-Litterman: blend those equilibrium returns with the investor’s views according to confidence.
- Constraints: minimum and maximum weights, ranges around a reference portfolio.
- Resampled MVO: simulate inputs many times and average the resulting portfolios; more diversified but less theoretically grounded.
- Non-normal optimization: use conditional VaR or other downside measures.
- Monte Carlo simulation of multi-period outcomes, including spending and taxes.
Risk budgeting
- Marginal contribution to total risk (MCTR) = asset beta to the portfolio × portfolio σ.
- Absolute contribution = weight × MCTR.
- Optimal: (excess return ÷ MCTR) is equal across assets.
Risk parity equalizes risk contributions from each asset class. It ignores expected returns and often needs leverage.
Factor-based allocation
Allocate to risk factors (market, size, value, momentum, rates, credit, inflation) rather than to asset classes. Clarifies hidden exposures; requires care with long-short implementation.
Liability-relative allocation
- Surplus optimization: MVO on surplus returns.
- Hedging / return-seeking portfolios: hedge liabilities first, then invest the rest for return.
- Integrated asset-liability approach: optimize assets and liabilities together (banks, insurers).
Goals-based allocation
- Identify goals, horizons and required probabilities.
- Build a module (sub-portfolio) for each goal.
- Size each module by the present value needed at its required success probability.
- Combine modules into an overall allocation.
Exam focus
Compute utility and pick between portfolios, explain the weaknesses of MVO and the fixes, apply risk budgeting and describe liability-relative and goals-based methods.