The “endowment model” — heavy allocations to private equity, venture, real assets, and hedge funds — has a well-documented return advantage and a less-discussed structural vulnerability: liquidity. An endowment or foundation (E&F) must fund an annual payout regardless of market conditions, but a large share of its assets cannot be sold on demand, and the very moments when liquidity is most needed are the moments when illiquid assets are least sellable. This article lays out a general, simulation-based framework for modeling that risk and the portfolio design choices that govern it.
The structural problem
Survey data on the E&F universe shows a consistent pattern: larger portfolios allocate more to illiquid assets — private equity, venture capital, and real assets — than smaller ones. The willingness to lock up capital in exchange for an illiquidity premium is, in large part, what the endowment model is.
The tension is mechanical. Spending is a fixed claim. Capital calls from private programs arrive on the general partner’s schedule, not the investor’s — and they tend to accelerate in downturns as GPs deploy into dislocation. Distributions, the natural offset, slow down at exactly the same time. So in a stressed market the liquid sleeve is simultaneously asked to fund spending, meet capital calls, and avoid being force-sold at the bottom. Modeling that interaction is the whole game.
A simulation framework
The natural tool is Monte Carlo simulation over a long, multi-year horizon. Within each simulated year, a disciplined sequence of operations captures the real cash mechanics of an E&F:
- Compound the portfolio to that year’s simulated asset-class returns.
- Honor capital calls, including ad hoc calls triggered in market downturns — the inconvenient, pro-cyclical calls that make liquidity planning hard.
- Fund the spending payout — typically modeled on a ~5% rule, smoothed over a trailing multi-year average of portfolio value so that spending doesn’t whipsaw with markets.
- Rebalance toward the strategic benchmark — but with a crucial real-world constraint: rebalancing is suspended in down markets, because you cannot sell illiquid assets to rebalance, and you don’t want to sell depressed liquid assets to top up illiquid targets.
Cash for the payout is drawn from liquid assets by liquidity ranking — most liquid first. That ordering is not a detail; it is the mechanism that determines whether the program survives a bad sequence of years.
Laddering: self-funding the private program
The single most important design idea for managing private-market liquidity is laddering commitments across vintages. By staggering commitments over time, incoming distributions from mature funds help offset outgoing capital calls from newer ones. Done well, the private program becomes partially self-funding — an immunization of sorts, where the cash flows of the book net against each other.
Two failure modes bracket the design:
- An over-large alternatives target that the liquid sleeve cannot realistically support through a stress episode.
- Filling commitments too quickly, which front-loads calls before the offsetting distribution stream has matured.
The literature on this is well established — Siegel’s work on illiquid assets and liquidity, and Sheikh and Sun’s defense of the endowment model, are useful conceptual anchors.
Proxying private-market risk
A simulation is only as good as the risk it assigns to the illiquid book — and reported private-market returns are notoriously smoothed, understating true volatility and correlation. Several standard approaches address this:
- De-smoothing reported returns to recover the underlying volatility (Ross-Zisler, Yule-Walker, and Geltner/Ang-style unsmoothing methods).
- Listed private-equity indices, which trade daily but have a weak link to the underlying assets.
- A leveraged small-cap value proxy (in the spirit of Stafford), which approximates the economic exposure of buyout via public-market analogues.
None is perfect. The point is to avoid feeding the simulation a flattered risk estimate that makes the illiquid book look safer than it is.
Defining a liquidity crisis
To compare designs you need metrics. Three are sufficient:
- Liquidity crisis — a year in which the required payout (plus calls) exceeds the liquid / fixed-income allocation available to meet it.
- Severity — conditional on a crisis occurring, the average amount by which obligations exceed the liquid sleeve. (Likelihood and severity are distinct: a design can lower the odds of a crisis while making the rare crisis worse.)
- Annualized return over the horizon — tracked alongside the two risk metrics, because every liquidity safeguard has a return cost.
The discipline is to evaluate any design choice on all three axes at once.
What the scenario tests teach
Running the framework across design choices yields several robust, qualitative lessons.
1. How much to hold in liquid assets. The benefit of a larger liquid reserve shows diminishing marginal returns. Beyond a moderate buffer — on the order of a high-single-digit percent of the portfolio — the reduction in crisis likelihood flattens while the return drag and tail severity keep growing. There is an efficient “sweet spot,” not a “more is always better” rule.
2. What to hold inside the liquid sleeve. Tilting the liquid reserve toward cash rather than corporate bonds reduces crisis likelihood efficiently, because cash holds its value precisely when it is needed. The trade-off is the give-up in yield.
3. How much of the private program to deploy. Holding a portion of private capital un-deployed (in liquid form) lowers crisis likelihood — but a moderately-high deployed share tends to balance returns against manageable liquidity risk. Counterintuitively, holding too much in reserve can actually raise tail severity in some configurations, because the return drag forces a larger drawdown elsewhere.
4. What to hold the committed-but-uncalled capital in. The choice of vehicle for capital earmarked for future calls mostly affects returns (riskier holding vehicles such as global or EM equities raise expected return) while barely moving crisis likelihood or severity. The real concern there is sellability under stress, not headline volatility.
5. Where the extra capital-call funding comes from. When ad hoc calls hit, the source of funding matters for both severity and return — making it a genuine design lever rather than an afterthought.
The takeaway
Liquidity risk in the endowment model is not a single number; it is the interaction of spending, capital calls, rebalancing rules, and the sellability of each sleeve through a downturn. A Monte Carlo framework that honors the real sequence of year-end cash flows — and that proxies private-market risk honestly — turns those interactions into something you can stress-test and design around. The headline lessons are reassuringly intuitive: ladder commitments so the private book partly funds itself, hold an efficient (not maximal) liquid buffer weighted toward genuine cash, and judge every safeguard on likelihood, severity, and return together.
Educational commentary only. This article describes a general modeling framework and qualitative conclusions using public concepts and references. It deliberately omits any proprietary capital-market forecasts, simulation parameters, thresholds, or specific result figures. Nothing here is investment advice.