Loading...
Loading...
Why foundation endowments misallocate capital.
The perpetual endowment is treated as the neutral container of serious philanthropy. It is not neutral.
Invest a corpus, distribute a small fraction each year, never spend it down — this is the default architecture of institutional giving, so familiar it is rarely seen, let alone defended. But perpetuity is not a vessel into which charitable purpose is poured and from which it emerges undistorted. It is a capital structure with its own logic.
The problem: once a foundation commits to existing forever, it commits to preserving the real value of its corpus indefinitely. That mandate forces the institution to optimise for its own survival over the resolution of the problem it exists to serve — regardless of how good its trustees are.
This is not an indictment of trustees. It is a claim about what the structure does to capital no matter who administers it.
A donor of significant means who wishes to give at scale is presented, almost without deliberation, with a single dominant form: the perpetual endowment, designed explicitly never to be spent. To propose anything else — to spend the corpus down, to recycle it, to synchronise its release to the life of a problem — is to invite the suspicion that one lacks "the long view" real philanthropy supposedly requires.
But the long view has been confused with a particular capital structure, and the confusion is consequential. Some problems genuinely are long; institutional continuity genuinely has value. These are real considerations. The argument is not that they are false but that they have been allowed to settle a question they cannot settle — whether the mission is best served by an institution built to last forever, or whether the institution's permanence has quietly become the end and the mission the means. Perpetuity is a choice with structural consequences, not the absence of one.
Every perpetual foundation operates under two obligations usually spoken of as one. A duty to the corpus — to invest prudently and preserve the endowment's real value across generations. And a duty to the mission — to advance the charitable purpose for which it exists. Governance treats these as complementary: the corpus is preserved so the mission can be served, indefinitely. Under a perpetual mandate they are not complementary but rivalrous — and the structure resolves the rivalry in a fixed direction. The corpus wins.
The mechanism is the real-terms preservation mandate. To preserve real value while distributing a fixed share, a foundation must earn, every year on average, a return equal to its distribution rate plus inflation plus administration. The distribution is not the purpose expressed in dollars; it is the residual left after preservation has been satisfied. The mission is funded from what remains once the corpus has been made whole. This is a fiduciary inversion: in the stated logic the corpus serves the mission; in the operative logic the mission is funded by whatever the corpus does not need to preserve itself.
Why does the corpus win? Because when two fiduciary duties compete and only one is precisely measurable, the measurable one dominates. A foundation that erodes its corpus by three percent faces a reckoning at the next investment committee; a foundation that under-serves its mission by some unknowable margin does not, because the margin is unknowable. The asymmetry of measurement becomes an asymmetry of obligation. The corpus — legible, auditable, defended by a committee whose entire remit is its preservation — wins by default.
Corpus primacy is the conflict expressed as an operating principle: preservation of the endowment functions as the first-order objective and the mission as a second-order one — not in rhetoric, where the ordering is always reversed, but in the actual structure of decision-making, where preservation is the constraint that binds and the mission is the variable that flexes to fit.
Ask which of the two duties is permitted to fail. A foundation facing a downturn will cut distributions to protect the corpus. It cannot, under a perpetual mandate, do the reverse — sustain mission spending and let the corpus erode toward depletion. The corpus is defended; the mission absorbs the shock. This is not how badly run foundations behave under stress; it is how the perpetual structure is supposed to behave to remain perpetual. Institutional continuation structurally outranks the problem's urgency, and the mission becomes the residual.
Over time the institution acquires an interest in lasting forever that is independent of — and eventually competitive with — its charitable purpose. Survival becomes a value in itself. This reproduces regardless of the quality or sincerity of trustees: it is the resting state to which the institution returns whenever active counter-pressure stops. That is what it means for corpus primacy to be structural rather than behavioural — it is the gravity of the system, not a choice anyone keeps making.
Most jurisdictions impose a minimum distribution requirement — a floor beneath which annual payout may not fall. It is conceived as a protection. The failure mode is that the floor becomes the ceiling. The minimum is read as the target, the target becomes the norm, and the norm hardens into a sectoral expectation that distributing more is imprudent — even profligate.
The number reveals the design. The United States five-percent minimum was calibrated against an assumed long-run total return of roughly eight percent — about three for inflation, five for distribution. The payout floor was reverse-engineered from the preservation mandate. It was never set by asking what charitable need required; it was set by asking what a foundation could give while still lasting forever. The floor is a preservation number wearing the costume of a generosity number.
Australia furnishes a live demonstration: following the 2024 Productivity Commission review, the government moved to raise ancillary-fund minimums to six percent, and the advisory sector warned this would harm philanthropy by maximising giving "too soon." That objection is coherent only if one has already accepted that the institution's perpetuation — not the mission's resolution — is the thing being optimised. A minimum is not a legal maximum, but the structure makes it a behavioural one: a corpus-primary institution gravitates to the preservation-compatible number and stays there.
The first two failure modes concern how much a foundation gives. The third concerns when. Horizon inversion is the condition in which the perpetual mandate holds capital longest precisely in those domains where the social return to early deployment is highest — where the structure's preference for deferral is most directly at odds with the mission's need for speed.
A perpetual endowment is, financially, a machine for converting present capital into future capital through compounding. Every dollar not distributed is a dollar that grows; the preservation logic rewards holding over releasing. The structure applies, in effect, a negative social discount rate to its own capital — it values future deployment more than present deployment, because future deployment is what perpetuity is for.
Now place that against the actual shape of social returns. For a large class of problems — early-childhood intervention, public health, disease eradication, climate mitigation — the return to early action far exceeds the return to late action, because early action forecloses the compounding of harm. The structure's optimal timing and the mission's optimal timing point in opposite directions, and the perpetual mandate resolves the conflict in favour of the structure. The foundation books its return on the corpus as prudent stewardship while the problem compounds its harm at a rate the foundation never measures — and can preserve the institution perfectly while allowing the mission to become impossible.
The sector has developed three reforms, each a sincere response to these tensions. Each is a within-structure adjustment to a structural problem — softening a symptom while leaving corpus primacy intact:
Higher payout rates. Lifting the floor moves more money, but within the same logic — the corpus stays primary, the distribution stays the residual, merely set slightly larger. The structure absorbs the higher rate and reasserts preservation around it.
Mission-related & program-related investment. Putting the corpus to work for the mission relaxes the conflict only to the extent the investment earns a preservation-compatible return. The moment it behaves like a grant, it re-triggers the preservation constraint. It serves the mission best where the mission needed help least.
Spend-down commitments. Renouncing perpetuity dissolves the engine — but as a one-time, all-or-nothing exit, not a repeatable architecture. The capital does its work and is gone; there is no recycling. And corpus primacy makes the decision feel, to trustees acculturated to preservation, like a dereliction.
This is the precise sense in which the perpetuity trap is a trap and not merely a flaw. A flaw can be corrected within the system that contains it. A trap is a configuration in which the available corrections are themselves constrained by the thing that needs correcting. The reforms address a temporal mismatch at the level of parameters — the rate, the use, the lifespan — while leaving untouched the conflict that generates the failure modes. They depend on sustained counter-pressure from unusually committed trustees, and decay back toward the default the moment that pressure relaxes.
If the three failure modes descend from a single conflict, then an aligned architecture is one in which that conflict does not arise — because the structure is not built to preserve a corpus in the first place. Three conditions follow directly from the three failure modes:
Deploy capital fully against the problem and have it return, so the same capital redeploys against the next instance. No corpus is held back to be preserved — capacity is maintained through deployment, not against it.
Match the timing of release to the problem's return profile — fast where returns are front-loaded, patient where they are genuinely long — rather than imposing one infinite horizon on every mission.
Measure problem-resolution across the full cycle of deployment and return, not the fraction of a corpus distributed in a twelve-month window. The accountability that anchors trustees to the floor would instead anchor them to the problem.
The reframe is from the stewardship of a corpus over the life of an institution to the stewardship of a problem over the life of its resolution. The perpetuity trap is a property of one architecture, not of the act of giving at scale — and the recoverable, recycling instruments developed elsewhere are how that door is walked through.
How recycling capital is housed in charitable lawFor genuinely multi-generational missions, a structure that holds and compounds may well be the right servant of the problem. But the long view is not the same thing as a perpetual corpus. The objection was never to patience — it was to patience imposed regardless of fit. The perpetual structure applies one horizon, infinity, to every mission it touches, because timing was never an input to its design.
Yes — many are acutely aware and work skilfully against the tensions described here. That is precisely the point: they must work against the structure. Excellence in the role is spent partly on resisting the design rather than executing it, and the moment that resistance relaxes, the preservation constraint reasserts its priority automatically. This is not a charge of bad faith — it is a claim about the structure, not the people inside it.
No — it adjusts a number within corpus-primacy logic. A higher payout is still a fraction of a preserved corpus, still derived from the preservation calculation rather than the problem's need, still released on a schedule set by the corpus. Raising five to six percent moves more money within the same machine; it does not change what the machine is built to do.
The full argument: the two fiduciary duties, the three failure modes, and the conditions an aligned capital architecture must satisfy.
View PapersCapital that recycles rather than depletes — live deployments with active counterparties, not theory.
Live Deployments