The decision in brief
Freeze the portfolio and assumptions, apply coherent shocks to assets and funding sources, and calculate both economic losses and dated cash requirements. Combine historical analogues, hypothetical scenarios, and reverse tests. Explain model limitations and identify feasible actions before presenting any scenario as evidence that the family's obligations remain fundable.
Begin with a concrete failure condition
Define what the committee wants to avoid: missing essential payments, breaching a collateral requirement, exceeding an accepted loss, or becoming dependent on an unavailable asset sale. Different failure conditions require different tests. A portfolio might remain solvent by estimated net worth while lacking cash for a payment, or meet every near-term obligation while experiencing a loss the family is unwilling to tolerate.
CFA Institute's Measuring and Managing Market Risk distinguishes sensitivity measures, scenario analysis, and statistical risk estimates, with limitations for each. Use those distinctions to choose the tool. A single-factor sensitivity helps explain one exposure; a multi-factor scenario explores combined effects. Neither automatically provides a probability, and a value-at-risk number should not be labeled the portfolio's maximum possible loss.
Freeze inputs and construct coherent scenarios
Record holdings, valuation dates, currencies, debt, pledges, and outstanding commitments at a common decision date. Document gaps in look-through exposure rather than silently treating missing positions as risk-free. Distinguish observed information from analyst assumptions and manager estimates. A test that mixes new public prices with old private values can still be useful, but the valuation mismatch must be visible.
Build several internally explained scenarios: an economic contraction, an inflation and financing shock, or an issuer-specific impairment combined with lost family income. State the horizon and how rates, spreads, currencies, and distributions interact. Historical episodes can inform assumptions without requiring that their exact sequence repeat. Hypothetical shocks should be severe enough to test a decision, not assembled solely to make the preferred portfolio look resilient.
Calculate a synthetic portfolio loss transparently
Assume an illustrative USD 100 million portfolio: USD 45 million equities, USD 30 million bonds, USD 15 million private assets, and USD 10 million cash. Apply synthetic losses of 40%, 10%, and 30% to the first three categories, leaving cash unchanged. The estimated losses are USD 18 million, USD 3 million, and USD 4.5 million, for USD 25.5 million in total.
Ending economic value is USD 74.5 million before payments and additional transactions. The assumptions are neither forecasts nor a historical backtest. The private-asset shock is an analytical estimate rather than guaranteed realizable proceeds. Show the arithmetic and exposure mapping so that a reviewer can reproduce the total. A polished chart cannot compensate for including the same underlying holding twice or omitting an important liability.
Add a separate funding stress
In the same synthetic scenario, suppose dated demands include USD 6 million of family payments, USD 4 million of capital calls, and USD 3 million of collateral. Against USD 10 million of cash, the initial funding gap is USD 3 million if all demands precede additional receipts. Identify which unencumbered asset can be sold, at what stressed value, and whether proceeds settle before the earliest shortfall.
Do not add all USD 13 million to the investment loss. A capital contribution transfers capital into an investment, while collateral may remain an asset but become unavailable. Spending has a different effect. MSCI's LiquidityMetrics material distinguishes liquidity analysis by transaction costs, market depth, and liquidation time. That supports testing sale feasibility separately from estimated portfolio value; a modeled valuation is not a promise of executable liquidity.
Reverse the test and challenge model weaknesses
A reverse test begins with the failure condition and asks which combinations of shocks could produce it. For example, determine how far collateral values and business receipts would need to fall before payment coverage disappears. Explore several combinations rather than declaring one mathematically smallest shock the only danger. The result describes a vulnerability boundary under assumptions, not the likelihood of reaching it.
Challenge diversification, stale valuations, and nonlinear instruments. A simple percentage haircut may be inadequate for options or leveraged exposures, which can require fuller repricing. Vary liquidation timing and costs as well as price shocks. Keep proposed management actions separate from the initial test, then rerun the portfolio after those actions with realistic constraints. Assuming perfect execution at unstressed prices can erase the very risk being examined.
Turn results into accountable decisions
Present the main loss drivers, earliest funding breach, assumptions with the greatest influence, and feasible responses. Actions might include changing a reserve, reducing a concentration, slowing commitments, or revisiting discretionary payments. Name an owner and deadline for each action. Keep unresolved data limitations visible and preserve the original scenario so later changes can be distinguished from corrections to inputs.
Implementation questions include who independently checks the arithmetic, who confirms collateral terms, and when changed holdings require a rerun. A second reviewer can test computational correctness without independently validating the scenario's realism. Say which review occurred. Passing the selected scenarios is evidence about those conditional cases only; it is not a guarantee against losses, a forecast of recovery, or proof that every important risk has been captured.
Decision points
- Define the failure condition before choosing shocks or producing loss charts.
- Separate market-value losses from collateral, capital calls, and spending demands.
- Make scenarios reproducible and link each material vulnerability to an owner.
Sources and further reading
The linked sources provide conceptual background. The worked examples and planning assumptions on this page are this publication's own illustrations.