Escrow and holdback structures that protect LP clawback

A clawback obligation runs against the general partner entity, and distributions received by that entity are ordinarily redistributed immediately to sponsors and other equity holders. The entity therefore cannot be relied upon to hold the resources to satisfy a clawback when one arrives. Everything built to solve that — guarantees by equity holders, a parent-company balance-sheet guarantee, joint and several liability, the individual undertaking — is credit support. It converts an empty entity into a claim against people, which is better than nothing and is not money.

The escrow is the exception. It is cash that never left, sitting in a separate account, and it is the only component of the package whose value does not depend on anyone's solvency or willingness to litigate. An institutional limited partner body's position reflects that: an escrow of at least thirty per cent may suffice as the clawback guarantee, which is a strong statement about the relative worth of cash and promises.

That framing sets up the only question worth arguing about, which is not whether there is an escrow but whether it is sized against the exposure it is meant to cover. An escrow specified as a percentage of distributions and an obligation determined by the mismatch between early and late outcomes are two quantities that move independently, and they diverge exactly when the reserve is needed.

The model agreement's rule is mechanical: the general partner deposits thirty per cent of amounts that would otherwise be distributed as carried interest into a separate fund account, until the applicable partner has received aggregate distributions equal to its commitment plus the preferred return. The prescriptive standard is consistent with it and slightly stronger — accrued carry should be held in escrow with significant reserves, for example thirty per cent of carry distributions or more, with additional reserves for potential clawback liabilities.

Two further figures circulate on the same topic and neither is the same measurement. One jurisdiction's drafting exemplar uses twenty per cent. And market practice is described as reserve accounts representing half of the after-tax carry, with sponsors resisting reserving the entire after-tax portion because doing so defers carry until late in the fund's life.

Four figures, three bases. Thirty per cent and twenty per cent are percentages of gross carry distributions. Half of after-tax carry is a percentage of a post-tax amount, which is a smaller base. Additional reserves for potential clawback liabilities is not a percentage at all. Averaging them or presenting them as a range would produce a number with no referent, which is why they are published here with their bases attached and reconciled arithmetically rather than editorially.

The reconciliation is worth doing because it is exact. Thirty per cent of gross equals half of after-tax precisely when one minus the assumed tax rate is six-tenths — that is, at a forty per cent assumed rate. Above that rate the prescriptive standard reserves more; below it, market practice does. That single line converts an argument about which convention is right into an arithmetic question about the rate the agreement already defines.

The fullest verified release architecture comes from a United Kingdom exemplar, and it is worth reading closely because it is unusually complete. Escrowed sums may not be withdrawn other than in accordance with the deed; to meet a payment obligation under the deed; to the extent amounts held are more than sufficient to meet the carry recipient's obligations on stated assumptions; or on termination of the partnership, or earlier once all investments have been realised and proceeds distributed.

The stated assumptions are the substance of it. Release is tested on the basis that all unrealised investments are realised for zero consideration, that all remaining undrawn commitments are drawn, invested and realised for zero consideration, and that the partnership has no assets available for distribution. That is a deliberately brutal test and it is the correct one for a reserve, because a reserve released on optimistic assumptions is not a reserve.

The softer test — a hypothetical final distribution computed on then-current values of the fund's assets and liabilities — is the one used for interim clawback measurement. Both belong in a model and they answer different questions. The then-current-value test asks whether the general partner is currently holding too much. The zero-consideration test asks how much it could end up holding too much of. The gap between the two is the honest measure of exposure, and the escrow should be sized against the second.

Two mechanical points follow the escrow rather than the test. Escrowed amounts are not treated as partnership assets and are treated as having been distributed to the carry recipient — which matters for allocation and for reporting. And tax advances may be released from escrow to cover irrecoverable tax on allocated carry, which is the mechanism by which a reserve does not create a personal liquidity crisis for the people whose carry it holds.

The net asset value coverage test is the escrow-sufficiency and interim-clawback benchmark, and the prescriptive standard states it as an example rather than a rule: the test should be established to ensure a sufficient margin of error on valuations, for instance at least one hundred and twenty-five per cent of net asset value. Reading that as a market level rather than as an illustrative floor is a common error; it is introduced with an example marker in the source.

The floor exists because the whole escrow analysis depends on valuations that are estimates. A reserve tested against net asset value with no cushion is a reserve that fails precisely when marks turn out to have been optimistic, which is the same event that creates the clawback in the first place. The margin of error is not conservatism for its own sake; it is a correction for the correlation between the two failures.

The calendar decides when any of it is enforced. The model agreement tests the general partner clawback at four moments: the first anniversary of the end of the commitment period, the date of the general partner's removal, the liquidation and final distribution of the fund, and any re-advance under the limited partner giveback — with notice and payment within ten business days. Insufficient net asset value coverage is itself one of the events on which a well-defined interim clawback should trigger.

Where the clawback formula is capped after tax, the escrow's apparent adequacy improves without the underlying exposure changing. That interaction is the reason the two terms should be negotiated in the same conversation: an after-tax cap and a thirty per cent escrow together can look like a fully covered position while leaving a material economic shortfall with limited partners, and the arithmetic below shows exactly how much.