Risk Budgeting Beyond Portfolio Weight Percentages

Choose the risk the committee is budgeting

Begin with the family's capacity for losses, essential spending, debt commitments, and investment horizon. State the unacceptable outcome before selecting a statistic. Volatility describes variability around returns, while a scenario loss, a liquidity shortfall, and a single-issuer limit describe different concerns. A committee should not assume that compliance with one measure implies compliance with the others.

CFA Institute's Introduction to Risk Management treats risk tolerance, governance, measurement, and modification as connected parts of a framework. Translate that into a small set of complementary controls. For example, monitor portfolio volatility as an estimate, concentration as an exposure, and payment coverage as a dated funding measure. Each needs an owner, a definition, and an action when the accepted boundary is exceeded.

Distinguish dollars invested from modeled risk

AQR's Understanding Risk Parity explains why a conventional capital allocation can be dominated by equity risk. The general insight is that an asset's capital weight does not fully describe its contribution to portfolio variability. Volatility and co-movement with other assets matter. Apply that insight without importing the paper's historical performance results as promises about a new family portfolio.

For a covariance-based model, portfolio variance is the weighted sum of variances and covariances. An asset's share of that modeled variance is its weight multiplied by the corresponding element of the covariance matrix times the weight vector, divided by total portfolio variance. Under this standard decomposition, the shares also equal shares of volatility contribution. State the estimation period, return frequency, and treatment of missing or stale prices.

Use a transparent synthetic two-asset example

Assume an illustrative unleveraged portfolio with equal capital in equities and bonds, annualized volatilities of 20% and 5%, and zero correlation. These are synthetic inputs, not current estimates or forecasts. Equity contributes 0.5 squared times 0.20 squared, or 0.0100, to variance. Bonds contribute 0.5 squared times 0.05 squared, or 0.000625. Total variance is 0.010625 and volatility is about 10.31%.

Equities therefore account for about 94.1% of modeled variance contribution despite receiving only 50% of the capital. With the same zero-correlation assumptions, 20% equities and 80% bonds produce equal weighted standalone volatilities and equal contributions, with portfolio volatility about 5.66%. That different mix is not automatically preferable: it changes economic exposure and may not meet the family's objectives. The example demonstrates measurement, not a recommended allocation.

Test model dependence and hidden shared exposures

Repeat the calculation with different correlations and volatility inputs. Estimated contributions can move when markets change, even if holdings do not. A short estimation period may react quickly but be unstable; a long period may underrepresent recent conditions. Private marks and incomplete look-through data can make an apparently precise covariance matrix less informative than its presentation suggests.

Review shared economic drivers across nominal asset classes. Public equity, private buyouts, credit, and the family business may all weaken in a financing shock. Risk diversification is not established merely by counting asset labels or managers. Use scenario attribution alongside statistical contributions and identify where the model cannot reliably estimate exposure. Do not reduce a complex concern to a convenient proxy without explaining what the proxy misses.

Keep funding and leverage outside the volatility shortcut

A lower modeled volatility does not automatically mean the portfolio can safely support more borrowing. Leverage creates repayment, collateral, and refinancing demands that may intensify during stressed markets. A cash-funded portfolio and a leveraged overlay can have similar estimated market volatility while imposing different liquidity obligations. Show gross exposure, debt, collateral, and counterparty dependence separately from the risk statistic.

Similarly, an asset with a small measured volatility contribution can still be inappropriate for the payment reserve if withdrawals are restricted. Keep uncalled commitments and potential hedge funding requirements in a cash-flow stress. Risk budgets should constrain the decisions that can damage family objectives; they should not reward instruments for appearing stable when their economic risks are difficult to observe or their cash demands are omitted.

Specify who can use and change the budget

Approve limits at the portfolio level before dividing responsibility among managers or sleeves. Explain which limits are hard constraints and which are review triggers. A manager can remain inside a mandate while increasing a risk shared with other managers, so aggregation must occur centrally. Where a limit is breached, document whether the cause is a trade, a market move, or revised measurement.

Ask who updates the model, who challenges assumptions, and whether a proposed investment improves the portfolio after all relevant risks are considered. Revisit the budget when obligations or ownership change, not only when measured volatility rises. A useful report connects each material risk to a decision, its expected benefit, and the remaining uncertainty. It does not treat a favorable model score as permission to ignore liquidity or concentration.

Sources and further reading

1. CFA Institute: Introduction to Risk Management
2. AQR: Understanding Risk Parity

The linked sources provide conceptual background. The worked examples and planning assumptions on this page are this publication's own illustrations.

Continue the research