Whole-of-fund vs deal-by-deal carry: consequences for LP and GP

The argument that dominates these negotiations — how much carry each structure pays — has a short answer that ends it. On a fund in which every investment returns more than its cost, the two structures pay the general partner the same carried interest in total; the deal-by-deal structure pays it earlier. The difference in quantum only appears on funds containing losses, and on those funds the difference is not paid to the general partner, it is owed back by it.

That reframes the negotiation. The question is not what the structure pays, it is what happens on the path where it pays too much. Under a whole-fund waterfall the answer is nothing, because the general partner receives no carried interest until all capital contributions across the whole fund plus the preferred return have been distributed, so the error cannot occur. Under a deal-by-deal waterfall the answer is a clawback — an obligation running against an entity that has ordinarily already redistributed the money to individuals.

Limited partners therefore trade cash-flow position for credit exposure, and the exchange rate is set by three terms rather than by the waterfall label: the escrow percentage, the frequency and reality of interim clawback testing, and whether the clawback is gross of tax. A deal-by-deal structure with a serious escrow, annual interim testing and a gross-of-tax formula can be a better position for a limited partner than a whole-fund structure with none of those, which is the sort of conclusion the label-level debate never reaches.

The negotiation record supports treating those terms as the live ones. Limited partners win interim clawbacks more than half the time in deal-by-deal structures and more than seventy-five per cent of the time in whole-fund structures — protection is easier to obtain in the structure that needs it least, which tells you where the resistance actually is.

The whole-fund structure defers the general partner's cash for years and creates a specific problem the market solves with a specific tool. Because a partnership must allocate its income currently, a fund deferring carry until capital is repaid still allocates carried interest to the general partner in the year it arises — creating a tax liability on cash that has not been received. Funds address this with tax distributions to the general partner, and the ILPA model's carried interest definition expressly captures distributions received under its tax distributions section.

So a whole-fund waterfall is not always a zero-carry-before-liquidation structure in practice. It is a structure whose interim payments are sized to a tax liability rather than to an economic entitlement, which is a materially different thing and should be modelled as one.

The deal-by-deal structure gives the general partner cash earlier and a set of retained liabilities: escrowed carry, an interim clawback exposure that is re-tested at defined intervals, and — where the security package includes joint and several liability — an exposure to other members' shares as well as its own. The escrow itself is the term where the market and the standard-setter are furthest apart: ILPA's model requires thirty per cent of otherwise-distributable carried interest, ILPA's principles suggest thirty per cent or more, one United Kingdom 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 it defers carry until late in the fund's life.

Accrued carry and earned carry are separate lines for a reason that matters here. Carry accrued against unrealised appreciation is not money; carry earned on realisations, inclusive of the amount held in escrow, is partly money and partly a reserve. A general partner reading a single blended carry figure will misjudge both its cash position and its exposure.

A subscription line is a first-order carry driver in either structure. ILPA warns that such a facility should not be used chiefly to enhance reported internal rate of return in order to accelerate the accrual and distribution of carried interest, and suggests keeping it within one hundred and eighty days and twenty per cent of commitments. Where the preferred return runs from the capital call rather than from the facility draw, the facility suppresses preferred return accrual and brings carry forward — an effect that operates identically under both waterfall structures and is usually negotiated in a different meeting.

Recycling does something similar from the other direction. Reinvesting exit proceeds rather than distributing them creates preferred return accrual gaps and interacts with the giveback limits; ILPA's position is that recycling provisions should expire at the end of the investment period and that recycling should be tied to the limitations on the limited partner giveback. In a whole-fund structure recycling delays the moment the fund-wide test is met. In a deal-by-deal structure it does not, which is a genuine asymmetry between the two.

Four ILPA rules on how carry is calculated cut across both structures and no competitor glossary carries all four. Carried interest should be calculated on net profits rather than gross profits, factoring in fund-level expenses. It should be calculated on an after-tax basis, so foreign or other taxes imposed on the fund are not treated as distributions to partners. No carry should be taken on current income distributions. And accrued carry should be held in escrow with significant reserves plus additional reserves for potential clawback liabilities.

Each of those is worth more on some funds than the waterfall structure itself. A deal-by-deal fund computing carry on net profits with no carry on current income and a thirty per cent escrow is a tighter instrument than a whole-fund fund with gross-profit carry and no escrow, and the term sheet headline says the opposite.

Whole-fund is the dominant structure globally at seventy-seven per cent in the rest of the world, seventy-three per cent in Europe and fifty-eight per cent in North America, with the combined deal-by-deal and hybrid category at twenty-three to forty-two per cent by region. The North American figure is the one that matters in most negotiations, and it says deal-by-deal is a real market position rather than an aggressive ask.

Most of the actual market sits between the two poles. A hybrid is deal-by-deal carry gated by fund-level tests — return of all fees and expenses to date, cumulative loss make-up, a net asset value coverage floor, a capped early-carry percentage, or a limited partner election between waterfalls — and each of those gates moves the structure a measurable distance towards the whole-fund answer while preserving some of the cash timing. Pricing the gate is more useful than arguing the label.

One boundary is worth marking because it is routinely crossed in these conversations. The limited partner giveback is a separate obligation from the general partner clawback: it funds indemnification and other fund obligations, it is capped in the ILPA model at the lesser of thirty per cent of distributions and twenty-five per cent of commitment, and it has nothing to do with carry being over-distributed. Conflating the two overstates limited partner protection.