Carried Interest: The Hurdle, the Catch-Up, the Split and the Clawback

Carried interest is the manager's share of a fund's profit. Carried interest is paid only after the distribution waterfall has run in order: every rupee of capital ever drawn returned to investors first, then the preferred return, then a catch-up to the manager, then a split of what is left. Once the split is reached the manager holds exactly 20.0 per cent of the profit, whatever the preferred return was.

Two people open a food stall. One of them puts up every rupee of the money and never stands behind the counter; the other stands behind the counter every day and puts up nothing. The two of them write down four sentences about how the takings will be divided. First, give the one who put up the money every rupee back. Second, give that person eight rupees a year for each hundred they put in, on top. Third, now give the one who ran the stall a lump big enough to put the two of them back at the ratio they agreed. Fourth, everything after that, eighty to the money and twenty to the work. The four sentences are a distribution waterfall, and carried interest is the third and fourth of them. A private fund of Rs 5,00,00,00,000 uses the same four sentences, written at greater length, and they do exactly the same thing to its cash.

What is carried interest, and who pays it?

Carried interestThe manager's share of a fund's profit, paid only after a defined order of payments. is a share of profit. A share of profit is the whole of the definition, and the temptation to soften it into something more familiar is exactly what gets it wrong. Carried interest is not a charge, it is not a rate applied to a balance, and there is no base it is calculated on. The manager's share is a slice of whatever profit turns up, and if no profit turns up there is no slice.

The reason this matters is that a manager receives two quite different things and only one of them is a fee. Nilgiri Alternatives Advisors Private Limited, invented, manages Nilgiri Growth Partners Fund II, invented. Fund II is settled as a trust. The role written up elsewhere as the general partner's is therefore discharged between that manager and Nilgiri Trusteeship Services Private Limited, invented, as trustee, and the contract is a trust deed and a contribution agreement rather than a partnership agreement. Under that contract the manager has drawn Rs 70,20,00,000 of management fee to the record date at the end of Fund II's Year 9 Quarter 2. Over the same eight and a half years it has been paid no carried interest at all, and the two facts sit in the same set of accounts without contradicting each other in the slightest.

Who pays it? The investors, out of the money the fund distributes, and only out of that money. Nothing is invoiced. Nothing is deducted from a balance. Cash arrives at the fund from a sale, the fund works down the order of payments its own documents fix, and if that order reaches the manager, the manager is paid from the same pile the investors are paid from. If the order does not reach the manager, the manager receives nothing and there is nothing left over to argue about.

How does cash leaving a fund get split, tier by tier?

How a Distribution Waterfall Works

A distribution waterfallThe order in which cash leaving a fund is split between investors and manager. is an order of payments with four numbered tiers. Every rupee the fund distributes enters at the top and falls as far as it needs to. No rupee jumps a tier, no tier is partly filled while a lower one is being paid, and nothing about the fund's quality, its holdings or its reputation enters the decision. The only input is how much has already been distributed.

Here are the four tiers as Nilgiri Growth Partners Fund I and Fund II, both invented, have contracted them. Tier 1, return of capital: everything to the investors until they have received back every rupee of capital ever drawn from them. Tier 2, the preferred return: everything to the investors until they have received, on top of that, a return of 8.0 per cent a year compounded on their contributions from the date each was made. Tier 3, the catch-up: everything to the manager until the manager holds 20.0 per cent of Tier 2 and Tier 3 added together. Tier 4, the split: everything after that, 80.0 per cent to the investors and 20.0 per cent to the manager. The four rates are these two funds' own contracted terms.

The most useful thing to notice is how narrow the question at each rupee actually is. Somebody in the fund's operations team does not weigh anything up. The operations team looks at one running total, compares it with three numbers written in the trust deed, and the answer falls out.

The four sentences are easy to read and easy to misread, so the clearest test of them is a fund whose waterfall has actually finished. Nilgiri Growth Partners Fund I, invented, wound up at its Year 10 Quarter 4. Fund I and Fund II run on separate clocks: Fund I's final close was four years before Fund II's, so Fund I's Year 10 Quarter 4 is the same real moment as Fund II's Year 6 Quarter 4, and every date in this guide names the fund it belongs to.

What did all four tiers look like when one fund actually finished?

Nilgiri Growth Partners Fund I, invented, drew Rs 2,40,00,00,000 from its investors across ten capital calls and distributed Rs 4,80,00,00,000 back across four payments. The profit above return of capital is therefore Rs 2,40,00,00,000, being 480 crore less 240 crore.

Tier 1 took Rs 2,40,00,00,000. Tier 2 took Rs 1,18,00,00,000, being the preferred return accrued on the fund's own convention to the wind-up date. Tier 3 took Rs 29,50,00,000, being Rs 1,18,00,00,000 multiplied by 20 and divided by 80. Tier 4 then took the remaining Rs 92,50,00,000 and split it Rs 74,00,00,000 to the investors and Rs 18,50,00,000 to the manager. The four tiers add to 240 plus 118 plus 29.50 plus 92.50, and that comes to 480 crore exactly. Every rupee the fund ever distributed has a tier attached to it.

Why does Tier 1 include the rupees that paid the manager's own fee?

Return of capitalThe first tier, paying investors back every rupee ever drawn from them. means every rupee drawn, not every rupee invested. The distinction between drawn and invested is small on paper and large in cash. Nilgiri Growth Partners Fund I, invented, drew Rs 2,40,00,00,000 in total, being 96.0 per cent of the Rs 2,50,00,00,000 its nine investors and its manager had committed. Of that, Rs 2,00,00,00,000 went into seven companies and Rs 40,00,00,000 went on the management fee and the fund's own expenses. Tier 1 is not Rs 2,00,00,00,000. Tier 1 is Rs 2,40,00,00,000.

The consequence is that the manager's fee is drawn from investors early and has to be earned back through the waterfall before the manager sees a rupee of profit share. Think of a household that borrows to open a small shop and pays a broker's commission out of the loan. The lender does not care that part of what was borrowed went to the broker; the whole loan has to come back. Tier 1 works the same way. Every rupee that left the investors, whatever it was spent on, has to come back before Tier 2 begins.

What is the preferred return actually doing?

The preferred returnA compounding rate the investor class receives before the manager shares in profit. is Tier 2. The preferred return is not a promise and it is not a coupon. Nothing pays it if the cash does not arrive; it simply accrues, and it stands in the queue ahead of the manager. On Nilgiri Growth Partners Fund I, invented, it accrued at 8.0 per cent a year compounded on each contribution from the date that contribution was made, and by the wind-up at Year 10 Quarter 4 it had reached Rs 1,18,00,00,000. The 8.0 per cent is the fund's own contracted convention, applied to its own ten drawdowns, and Rs 1,18,00,00,000 is the amount Tier 2 paid.

Tier 2 sets the point at which the manager's share begins, and that point is a cumulative rupee figure rather than a rate anybody has to beat. Rs 2,40,00,00,000 of capital plus Rs 1,18,00,00,000 of preferred return comes to Rs 3,58,00,00,000. Until cumulative distributions cross that line, the manager receives not one rupee of profit share, however good the holdings have looked on paper. The preferred return is sometimes called the hurdle for that reason, and the name is fair enough. A hurdle is a height the cumulative cash has to clear.

What is Tier 3 catching up to?

The Catch-Up, and the Ratio It Restores

Here is the question the catch-upThe tier paying the manager until it holds its full percentage of everything above return of capital. answers. Tier 2 has just paid the investors Rs 1,18,00,00,000 and paid the manager nothing. The contract says the manager takes 20 per cent of profit. At this moment the manager holds 0 per cent of it. Tier 3 pays the manager, and only the manager, until the ratio is back where the contract says it should be.

How much is that? The manager needs to hold 20 per cent of Tier 2 and Tier 3 added together. Call the catch-up amount X. Then X must equal 0.20 multiplied by the sum of Rs 1,18,00,00,000 and X. Rearranged, X equals 118 multiplied by 20 and divided by 80, being Rs 29,50,00,000. Check it the other way: Rs 1,18,00,00,000 plus Rs 29,50,00,000 is Rs 1,47,50,00,000, and Rs 29,50,00,000 divided by Rs 1,47,50,00,000 is exactly 0.20. The catch-up is always 0.25 times whatever Tier 2 paid, because 20 divided by 80 is one quarter.

A caterer at a wedding is a fair everyday picture of it. The agreement is that the hall's costs come out of the takings first, then the household that booked the hall takes a fixed slice, then the caterer is paid a lump so that the caterer's share of that whole surplus is back at the fifth they agreed, and only then is anything left divided. Nobody thinks the caterer has been given a bonus. The caterer has been brought back to the ratio that was written down before anybody started cooking.

Once the split is reached, what does the manager actually hold?

Exactly 20.0 per cent of the whole profit above return of capital. No more and no less, whatever the preferred return happened to be. The identity that lands the manager on exactly one fifth is what makes the catch-up understandable, and it is worth checking rather than accepting. On Nilgiri Growth Partners Fund I, invented, the profit above return of capital was Rs 2,40,00,00,000. The manager received Rs 29,50,00,000 in Tier 3 and Rs 18,50,00,000 in Tier 4, a total of Rs 48,00,00,000. Dividing Rs 48,00,00,000 by Rs 2,40,00,00,000 gives 0.200000.

The algebra is one line and it holds for every preferred return. Write the preferred return as P and the profit above capital as G. Tier 3 pays 0.25P. Tier 4 pays the manager 0.20 of whatever is left, and what is left is G less P less 0.25P. So the manager holds 0.25P plus 0.20 multiplied by G less 1.25P. Multiplied out, that comes to 0.25P plus 0.20G less 0.25P, and the whole of it is 0.20G. The P cancels, and that is the entire point. A larger preferred return produces a larger Tier 2 and a proportionally larger catch-up, and the two move together and leave the manager on the same fifth.

So the gap between what a fund distributes and what its investors keep is the carried interest and nothing else. Fund I distributed Rs 4,80,00,00,000 on Rs 2,40,00,00,000 paid in, a gross multiple of 2.00 times. Its investors received Rs 4,32,00,00,000 on the same Rs 2,40,00,00,000, a net multiple of 1.80 times. With the denominator named, both numbers are true at once; the 0.20 times between them is Rs 48,00,00,000, and there is nothing else in it. The fees and expenses were drawn as capital and therefore sit inside the Rs 2,40,00,00,000 denominator of both figures, so none of that gap is fee.

If the preferred return does not change what the manager finally gets, what does it change? Two things, and both of them are large. The preferred return changes when. The manager waits until cumulative distributions clear Rs 3,58,00,00,000, and on this fund that took until the last of four payments in the tenth year. The second change is whether, and it matters far more. If a fund never reaches Tier 4 at all, the identity never engages and the manager holds whatever the tiers actually reached, perhaps nothing at all.

When did the manager of that fund actually get paid?

Once, at the end. Timing is the part of the mechanism no static total shows, so take the four distributions of Nilgiri Growth Partners Fund I, invented, in order. All four dates are on Fund I's own clock.

Read the second distribution again. Rs 80,00,00,000 and then Rs 1,00,00,00,000 is Rs 1,80,00,00,000 of cash returned to investors, and the fund was still Rs 60,00,00,000 short of finishing Tier 1. Read the third. Rs 1,50,00,00,000 arrives in the ninth year, Rs 60,00,00,000 of it completes Tier 1 and Rs 90,00,00,000 of it begins Tier 2, and the manager receives nothing at all on the third distribution of a ten-year fund. The whole of the manager's Rs 48,00,00,000 fell in the fourth and final payment, at Fund I's Year 10 Quarter 4, the day the fund wound up. Check the last row: 28 plus 29.50 plus 74 plus 18.50 is 150 crore exactly.

Why is twenty per cent of the profit above the preferred return the wrong answer?

How do a manager's two incomes actually behave?

How Private-Fund Fees and Carried Interest Work

Metrick and Yasuda, in The Economics of Private Equity Funds, Review of Financial Studies, 2010, showed that a fund manager's pay is not one thing but two, and that the two must be valued separately. One is a stream that arrives whatever happens. The other is contingent on an outcome. The separation between a stream and a contingent share shows about as starkly as it can be shown in the two incomes of this invented manager.

Nilgiri Alternatives Advisors Private Limited, invented, charges Nilgiri Growth Partners Fund II, invented, a management fee of 2.00 per cent a year. During the five-year investment period the basis was the Rs 4,90,00,00,000 of investor commitments, giving Rs 9,80,00,000 a year. From Year 6 the basis became the acquisition cost of holdings not yet realised, measured at the start of each year, so the charge fell to Rs 8,00,00,000, then Rs 6,40,00,000, then Rs 5,00,00,000, and Rs 1,80,00,000 for the half of Year 9 elapsed at the record date. Five years at Rs 9,80,00,000 is Rs 49,00,00,000, and the post-period charges add to Rs 21,20,00,000, giving Rs 70,20,00,000 drawn in total. Fund II has not completed Tier 1, so over the same eight and a half years the carried interest was nil in every single year.

The difference is structural rather than a matter of size, so set the two side by side in words. The fee is charged on a base, arrives on a schedule, and does not ask whether the fund has made money. The carried interest is charged on nothing at all, arrives only if a cumulative cash total is crossed, and can be nil for a decade and then land in a single quarter, as Fund I's did. A reader who treats them as two lines of the same income will mis-read every fund document they ever open, and the separation Metrick and Yasuda drew is the reason to keep them apart. How the fee itself is calculated, and what it pays for, is covered separately.

Would any of this change if the carry were taken deal by deal?

Fund II's waterfall is whole-of-fundA waterfall in which all capital drawn must come back before any carried interest is paid.. Every rupee of investor capital ever drawn comes back before the manager sees anything. The fund has distributed Rs 4,38,00,00,000 against Rs 4,80,00,00,000 drawn, so it is Rs 42,00,00,000 short of finishing Tier 1 after eight and a half years, and its manager has received no carried interest, has no clawback exposure and holds an empty escrow.

The alternative ordering is deal-by-dealA waterfall taking carried interest on each realisation as it happens., under which the manager takes its share of each sale's profit as that sale happens rather than waiting for the fund's whole capital to return. Fund II's documents do not say that, so what the alternative ordering would have done to this fund can only be worked out as a counterfactual. Had Fund II run deal by deal on the same nine holdings, the same three sales and the same dates, holding 2 would have paid the manager Rs 3,60,00,000 at Fund II's Year 6 Q4 on Rs 18,00,00,000 of profit, holding 1 would have paid Rs 26,60,00,000 at Year 7 Q3 on Rs 1,33,00,00,000, and holding 3 would have paid Rs 18,00,00,000 at Year 8 Q2 on Rs 90,00,00,000. Under that counterfactual the manager would hold Rs 48,20,00,000 by Fund II's Year 8 Quarter 2, on a fund that has still not returned its investors' capital.

A builder is the everyday version. Paid per flat as each flat sells, the builder has money in hand long before anybody knows whether the whole project covered its cost. Paid only once the project's total cost is recovered, the builder waits, and may wait for ever. Same flats, same buyers, same prices. The ordering does not change the rate and does not change the holdings; it changes only when the money moves, and therefore who is holding it if the arithmetic later turns out badly.

What is the clawback, and how far does the escrow actually reach?

A clawbackAn obligation on the manager to repay carried interest it turns out not to have earned. is an obligation on the manager to hand back carried interest that the finished arithmetic says it never earned. The clawback exists because of exactly the gap the counterfactual above opens up. Under deal-by-deal the manager is paid on each sale as it happens, but whether the manager was entitled to any of it can only be settled once the fund is finished and every rupee drawn has been counted against every rupee returned. Between those two moments the manager is holding money on a provisional basis.

The same arrangement turns up outside finance. A salesperson takes an advance against commission on a sale that has not completed; the sale falls through; the advance goes back. Nobody thinks the advance was a gift and nobody thinks the repayment is a penalty. The advance was a payment made before the fact that decided it was known. A clawback is that arrangement written into a fund's documents. Fund II's is settled at the end of term and calculated net of the taxes the manager has actually borne, so the amount recoverable is never simply the amount paid.

The escrowMoney held back from each carried interest payment against a possible clawback. is the security behind it. Fund II's documents hold back 30.0 per cent of every carried interest payment against a possible clawback. Run that against the counterfactual and the honest size of the protection appears: 30.0 per cent of Rs 48,20,00,000 is Rs 14,46,00,000 held, and Rs 33,74,00,000 already released to the manager. The fund is Rs 42,00,00,000 short of returning its investors' capital. A fully funded escrow would therefore cover Rs 14,46,00,000 of a Rs 42,00,00,000 shortfall, being 34.4 per cent of it, and no more.

Covering a third of the shortfall is the honest size of the protection, and it is worth stating plainly rather than letting the word escrow do work it cannot do. An escrow is a partial security against an obligation, not a fund of money sitting ready to make investors whole. The rest of the clawback is a promise by a company to pay, and what it is worth depends on whether that company can pay it years later. Nilgiri Growth Partners Fund II, invented, is whole-of-fund, so none of this has happened: no carried interest has been paid, the clawback exposure at the record date is nil and the escrow is empty.

What does somebody actually do with this when a document lands on the desk?

More people read these clauses than write them: somebody on an investment committee weighing a commitment, a monitoring team inside an institution that has made several, an analyst covering such an institution, an auditor checking a distribution notice, and a student who will do one of those jobs in a few years. Four questions, and every one is answerable from a document rather than from any judgement about the manager.

First, find whether the waterfall is whole-of-fund or deal-by-deal. One word there decides whether a clawback can ever be needed. Second, find the catch-up percentage. A full catch-up at 100 per cent to the manager, as both these invented funds have, means the identity engages and the manager lands on its full share; a partial catch-up does not, and the arithmetic is different. Third, read Tier 1's definition and check whether it says capital drawn or capital invested. On Fund I those two are Rs 2,40,00,00,000 and Rs 2,00,00,00,000, so the difference is Rs 40,00,00,000 of when. Fourth, ask where the fund's cumulative distributions sit against Tier 1 today. One comparison settles whether the manager can currently be paid anything at all, and on Fund II the answer is a plain no by Rs 42,00,00,000.

None of those four needs an opinion about anybody. All four are readable, and three of them in a single afternoon, from the fund's own documents and its latest distribution notice. At Fund II that notice is signed by Farida Contractor, invented, as chief operating officer of the manager.