Illiquidity: The Central Trade-Off of Private Investing

IlliquidityThe condition of not being able to turn a holding into cash at the holder's choosing. is not one constraint but four, and they arrive together, so they get discussed as one. The holder cannot get out at will. The holder must put more in when asked. The reported value of the holding is a number no buyer has tested. And none of it can be timed or interrupted.

Start with something ordinary. A person puts money into a recurring deposit at a bank and, in the third year, needs it back for a wedding. There is a rule about breaking it early and a penalty for doing so, and the person grumbles about the penalty, but the money comes back. Now change one thing and only one thing. Take away the counter where the request can be made. There is no early-closure form, no penalty schedule, and no officer with the authority to hand anything over, because nobody at the other end has agreed to give money back on request and nobody ever did. The missing counter, and not the penalty, is what a commitment to a private fund actually is. The constraint here is not that getting out is expensive; it is that getting out is not one of the things the arrangement does.

Nilgiri Growth Partners Fund II, an invented closed-end fund, carries every figure below. Nilgiri Alternatives Advisors Private Limited manages it, with Nilgiri Trusteeship Services Private Limited as trustee and Nilgiri Financial Holdings Private Limited as sponsor, all three invented too. The fund carries total commitments of Rs 5,00,00,00,000 across twelve investors plus the manager's own Rs 10,00,00,000. Every figure below is read as at the end of the fund's Year 9 Quarter 2. The record date sits 8.50 years after the fund's final close. The documents and the investors use the vocabulary of general partner and limited partner, so that vocabulary is used here too. In this vehicle the general partner's role is discharged by the manager and the trustee between them, and the contract is a trust deed and a contribution agreement rather than a partnership agreement.

What does illiquidity actually stop a holder doing?

Four things, and a reader who has met only the first will be surprised by the other three. The four are not variations on a single theme. Each is a fact about a different part of the arrangement, written into a different clause, and each would still be true if the other three were somehow removed. The four get discussed as one thing for two reasons: they arrive at the same moment, on the day the money is committed, and English has only one convenient word for the bundle.

How Illiquidity Shapes Private-Market Risk

The four constraints change what risk even means for a holder of this kind of interest, and it is worth being precise rather than atmospheric about how. A person holding a listed share carries the risk that the price moves against them and the separate, smaller risk that they cannot sell in size on a bad afternoon. A person holding an interest in a closed-end private fund carries something structurally different, and it stacks in four layers.

The first layer is that a bad personal moment cannot be answered by selling. If a household needs cash in Year 3, the arrangement has nothing to offer it. The second layer is that the arrangement can make a bad personal moment worse. A capital call may land in exactly that month, and the obligation to meet it does not care what else is happening. The third layer is that the holder cannot even measure the trouble accurately. The number on the statement is an estimate, and the only honest thing to say about it is that nobody has bought at it. The fourth layer is that the holder cannot shorten the exposure by deciding to. Risk here is not only the chance that the value falls; it is the certainty that the holder's own timing has been handed to somebody else for the length of the contract.

The items missing from that list matter as much. Nothing above says the holder does worse or better than they would have done elsewhere. Nothing above prices any of it. The four constraints are facts about the shape of the arrangement. Whether they are worth anything is a separate question.

Constraint one: what happens when an investor wants out and the fund has no way out?

Nothing happens, and that is the whole of it. There is no counter. A closed-end private fund of this kind never agreed to offer exit on demandThe ability to sell whenever the holder decides to, at a price somebody will pay., and so does not. Its documents do not contain a clause under which an investor asks for its money and the fund arranges to give it back, and the reason is structural rather than mean-spirited: the money is not sitting anywhere. By the middle of the life of Nilgiri Growth Partners Fund II the cash that had been called had already been paid to the founders and companies who sold shares to the fund, and the fund holds stakes in nine private companies instead. To pay one investor out, somebody would have to sell part of a company, at a price, to a buyer, on a timetable. Such a sale is not a redemption but a transaction, and transactions in unlisted companies take months and often do not happen at all.

Look at the record of this one invented fund and the point stops being theoretical. Across Years 1 to 5 the fund made no distributionA payment of cash from the fund back to its investors. whatsoever. Not a small one, not a partial one. Its first distribution was Rs 63,00,00,000 at Year 6 Q4, being the proceeds of one holding sold in the quarter before. An investor of that fund who wanted its money in Year 3 had no route the fund could offer, no route the manager could create, and no penalty it could pay to force one. The absence here is not a delay in a process; there is no process.

There is one route the fund does not provide and the reader will reach for anyway, so it should be named plainly and then handed on. An existing investor can, in principle, find somebody willing to step into its position and sell that position to them, in what is called a secondary marketA privately negotiated sale of an existing fund interest from one holder to another. transaction, at a price the two of them negotiate privately and usually with the manager's agreement to the transfer. How such a stake is found, priced and documented is a subject in its own right and is covered separately; no price is established here, and whether such a buyer exists for any given position is never assured. Two things about it are worth carrying away. A secondary sale moves the interest from one holder to another and is not a payment by the fund, so the fund's own cash position is unchanged by it. And the price is negotiated, so the seller is the one under time pressure and the buyer is not. Tapping sell on a screen is a different situation altogether.

Even the right to transfer at all is narrower than most readers assume. In this invented fund, six side letters exist and only one of them touches transfers: investor 4, the treasury of a domestic bank, negotiated a right to transfer its interest to any entity under the same control. The side letter gives a right to move the position inside its own group of companies, not a right to sell it to the world. Investor 3, an overseas sovereign-linked investor, negotiated an excuse right allowing it to be left out of investments in sectors its own mandate excludes, and that right does nothing at all about getting money back. A large investor might be assumed to have negotiated a way out. The large investors here negotiated information, consent and a seat, and never an exit.

Constraint two: why does a commitment work in the opposite direction from what a reader expects?

Here is the part that surprises people, and it surprises them because the word illiquidity points the wrong way. The word suggests money that is stuck inside and cannot come out. In a closed-end private fund the more demanding half of the arrangement is that money the investor still holds, in its own bank account, earning whatever it earns, is not really the investor's to spend. The uncalled money is a live commitmentAn obligation to send more money when the manager calls it., and it can be summoned.

The mechanism is the capital callThe notice requiring an investor to send committed money to the fund.. The investor promises a total figure at the start, called its commitment, and the manager draws that promise down in instalments, over years, with a notice period the fund's own documents fix. Farida Contractor, the chief operating officer of the manager, is the person who signs those notices. The mechanism itself, and how a notice is answered, is covered separately. Only its direction matters here. The investor does not choose when. The investor does not choose how much. The investor cannot decline without consequences its own contract sets out. For as long as the commitment is live, the investor is carrying an obligation of unknown timing and known size.

Now put the two constraints in the same five years and read them together. Nilgiri Growth Partners Fund II called capital seventeen times over its life. Thirteen of those seventeen calls fell inside its five-year investment period, and they add to Rs 4,55,50,00,000. The thirteen calls inside the investment period are 94.9 per cent of the Rs 4,80,00,00,000 the fund would ever draw. In the same five years it distributed nothing at all. In the period when almost every rupee was demanded, not one rupee came back, and that is not a fault in this fund's record but the ordinary shape of the arrangement.

The two named points of this fund's value path make the same reading in a different currency, and the year-by-year path of a private fund's value is covered separately. At the end of Year 5, cumulative capital paid in was Rs 4,55,50,00,000, cumulative distributions were nil, and the fund's total value to paid in stood at 0.95 times. At the record date at Year 9 Q2, cumulative paid in was Rs 4,80,00,00,000, cumulative distributions were Rs 4,38,00,00,000, and total value to paid in stood at 1.50 times. The two figures are readings of one fund at two stated dates. The readings are not a trend, they are not a projection, and no line joins them.

One number here is the single most commonly mangled figure in this whole subject, and it is worth slowing down over. At the record date the fund has an unfunded commitment of Rs 20,00,00,000, being Rs 5,00,00,00,000 promised less Rs 4,80,00,00,000 drawn. The Rs 20,00,00,000 is the fund's figure. Every investor here is drawn strictly pro rata, and every one of the twelve is 96.0 per cent paid in. Investor 1, a domestic life insurance company, committed Rs 1,00,00,00,000 and has paid in Rs 96,00,00,000. So investor 1's own unfunded commitment is Rs 1,00,00,00,000 less Rs 96,00,00,000, or Rs 4,00,00,000. One figure belongs to the fund and one belongs to one investor, so Rs 20,00,00,000 and Rs 4,00,00,000 are five times apart, both correct, and neither may ever be written in the other's place. Whenever a figure like this is quoted, the denominator has to be said out loud.

Constraint three: why does a reported value belong on a list about illiquidity at all?

Because the two are the same fact seen twice. A holding that cannot be sold on demand has, by that very fact, no price from a buyer. The investor is shown an estimateA reported value that no buyer has tested. instead: somebody's careful, documented, professionally produced opinion of what the holding would fetch if it were sold, made without the one event that would settle the question. In this arrangement the opinion is produced by Palani Valuation Advisors LLP, a limited liability partnership acting as the independent valuation agent. The valuer prices every unrealised holding once a year, the manager marks quarterly in between, and Kolar Fund Services Private Limited, the administrator, strikes the fund's total. Rohit Vaz signs at the valuer and Ashwin Baliga strikes the total at the administrator. None of that machinery is careless and none of it is a substitute for a buyer.

Look at the numbers that machinery produces for this fund at the record date. Total value stands at Rs 7,20,00,00,000. Of that, Rs 4,38,00,00,000 is cash that actually arrived, from four holdings that left the portfolio and one that was partly sold. The remaining Rs 2,82,00,00,000 sits in five holdings, and not one of the five has been sold to anybody. The Rs 2,82,00,00,000 is not a forecast and not a guess but a valuation, made on a documented basis, and no buyer has ever paid it. An estimate cannot be turned into cash by agreeing with it.

The five are not uniform, and the spread inside them is the part worth studying. Holding 4, Bhavani Speciality Chemicals Private Limited, is carried at Rs 1,08,00,00,000 against Rs 60,00,00,000 of cost, being 1.80 times. Holding 6, Vaigai Edutech Private Limited, is carried at Rs 21,00,00,000 against Rs 30,00,00,000 of cost, being 0.70 times. The holding has been written down and is now reported as worth less than the fund paid. Holding 7, Manjira Industrial Services Private Limited, is at Rs 39,00,00,000 on Rs 30,00,00,000, being 1.30 times. Holding 8, Kaveri Renewables Private Limited, is at Rs 81,00,00,000 on Rs 45,00,00,000, being 1.80 times. And the remaining 60 per cent of holding 9, Indravati Packaging Private Limited, is at Rs 33,00,00,000 on Rs 15,00,00,000 of remaining cost, being 2.20 times. Every one of those five multiples is an estimate divided by a fact.

One more thing belongs here. Readers most often misread this part as a criticism. The reported value moves late relative to what is happening inside a business, and that is not negligence. In this fund, holding 3, Tungabhadra Logistics Private Limited, was carried at 2.00 times cost at the end of Fund II's Year 7 and was realised at 2.50 times in Year 8 Q1. The reported figure moved by half a turn on the day it was sold, and nothing about the business changed on that day. The lag works the other way too: holding 6 was carried at cost through Year 5 and written to 0.70 times in Year 6, by which point the deterioration had been visible for three quarters. A private mark is made from information arriving on a quarterly and annual timetable. A public price is made from information arriving continuously. The difference in timetables is the whole of the reason, and a difference in timetables is neither an advantage nor a defect.

Constraint four: who decides when any of this happens, if it is not the investor?

The manager does, inside a contracted termThe agreed life of the fund, after which it must wind up. that was fixed before the first rupee moved. Nilgiri Growth Partners Fund II has a term of ten years from its final close and a five-year investment period inside that. Two extensions of one year each exist in its documents: the first at the manager's election with the prior written consent of the investor advisory committee, the second requiring the consent of investors holding more than half of commitments by value. Neither has been taken at the record date. Inside that frame, when to buy, what to pay, when to sell and whether to wait one more year are decisions the manager makes, subject to its investment committee of five, of whom four are from the manager and one is external.

An investor of this fund has real rights and none of them is a right of timing. The investor advisory committee, seven members drawn from seven of the twelve investors and chaired by Meera Sathe for investor 1, consents on conflicts, on valuation policy, on the first extension and on any change to the investment policy. The committee does not approve investments and cannot reject one. Constraint four is clearest there: the body that represents the largest investors in this fund can withhold consent on four listed subjects and can do nothing whatsoever about when a holding is sold. Timing was not retained by the investor, was not shared with the investor, and cannot be reclaimed by any vote the documents provide.

Read that against the clock and it becomes concrete. At the record date, 8.50 years of the ten have run. Six quarters remain. Five of the nine holdings have never been sold, and their combined carrying value of Rs 2,82,00,00,000 is 58.75 per cent of everything the fund has ever drawn from its investors. An investor who wanted the position finished could not make it finish faster; an investor who thought it was being finished too fast could not slow it down. The fund is also Rs 42,00,00,000 short of returning the Rs 4,80,00,00,000 it has drawn, having distributed Rs 4,38,00,00,000, and that gap will close, or not, on the manager's timetable.

What does a description that stops at cannot withdraw actually leave out?

Most of it. A reader who has understood only that money cannot be taken out has understood the least surprising of the four constraints. Everybody prices that one in before signing, so it does the least damage in practice. The three that follow are the ones that catch people. The second is the sharpest of them: it points the other way.

Two smaller facts sharpen the same point and are worth stating because they are easy to miss. The four calls after the investment period ended are small and are fees and expenses and nothing else: Rs 8,80,00,000 at Year 6 Q1, Rs 7,20,00,000 at Year 7 Q1, Rs 6,30,00,000 at Year 8 Q1 and Rs 2,20,00,000 at Year 9 Q1, adding to Rs 24,50,00,000. So the obligation to send money does not stop when the buying stops; it merely gets smaller. And the fund's remaining Rs 20,00,00,000 of unfunded commitment exists precisely to pay the fee and expenses of the six quarters that are left. An investor whose commitment period feels finished still has a live obligation running to the end.

How long is long, on one invented fund's actual record?

The answer comes with its boundary attached. The holding periodHow long one investment was held before it was realised. of the four positions this fund realised in full were 5.75 years for holding 1, 4.75 for holding 2, 5.75 for holding 3 and 3.75 for holding 5, a position written off rather than sold. The four holding periods average exactly 5.00 years. The average is a fact about four holdings of one invented fund over one period, and it says nothing whatsoever about how long anything is held anywhere else. It is stated here because a reader who has never seen a real schedule has no anchor at all, and one honest anchor with its limits printed on it is better than a vague sense that these things take a while.

The five positions still held are the more instructive half. Their clocks are still running. Holding 4 has been held 6.50 years at the record date, holding 6 for 5.50, holding 7 for 5.00, holding 8 for 4.50 and holding 9 for 3.50, with 40 per cent of it already sold at Year 8 Q3, 2.75 years after it was entered. Two of those five have already been held longer than the 5.00 year average of the four that finished, a third sits exactly on it, and every one of them may run longer still. An average computed from the positions that have ended is computed from a group chosen by the very thing being measured, and the positions that are taking longest are exactly the ones missing from it.

What is the compensation supposed to be, and who says so?

There is a named idea, and a reader will meet it sooner or later, so it is better met here with its edges drawn. The idea is that illiquidity is not merely a nuisance to be endured but something an investor is compensated for, and the phrase usually attached to it is the illiquidity premium. Ang, in Asset Management, published in 2014, treats it seriously and at length, and the framing is his. Whether such a premium exists, and at what size, is not something a single worked case can establish: no figure, no range and no direction follows from it. A claim of that kind cannot be supported by one fund's figures, and a plausible number produced to satisfy a reader would be worse than no number at all.

The reasoning is the useful part, so it is worth being concrete about why one worked fund cannot answer the question. To test whether investors are compensated for accepting the four constraints, a study would need a large number of funds and not one; it would need every fund that ever existed and not only those that survived to report, because the ones that stopped reporting are exactly the ones that would move the answer; it would need to decide what to do about the large part of reported value that is a mark rather than a transaction, since Rs 2,82,00,00,000 of this one fund's Rs 7,20,00,00,000 has never been sold; it would need a stated thing to compare against and a stated period, both fixed in advance rather than chosen afterwards; and it would need to settle how to treat money that sat uncalled in an investor's own account for years while it was committed. Change any one of those five decisions and the answer changes. None of them is a detail.

The safe statement is about mechanism and not about reward. Because this vehicle does not offer exit on demand, its manager is never required to sell a holding in a quarter simply because a number of investors have asked for cash at once. A vehicle that deals daily has to be able to meet the requests it has agreed to meet, and that is a different operating constraint. Being able to hold through a period is a property of the structure; being paid for doing so is a claim about outcomes, and only the first of those two is a fact about this arrangement. Comparing what private and listed holdings have returned, in either direction, is a different subject and is covered separately, and a measurement of one fund against one index over one period says nothing beyond that fund.

What do the four constraints look like on one investor's own statement?

Take investor 1, a domestic life insurance company. Investor 1 committed Rs 1,00,00,00,000 and is therefore the largest of the twelve, holding 20.0 per cent of the fund. Every investor here is drawn strictly pro rata and no side letter moves investor 1's economics, so its position is the fund's own position scaled by one fifth. Investor 1 receives a capital account statement each quarter, an unaudited quarterly report, a manager's letter beside the numbers, a notice for every call and every distribution, an audited annual report and an annual valuation report. Here is what those documents show at the record date.

Read that table as an investor rather than as an accountant and the four constraints are all visible in it. The Rs 87,60,00,000 arrived in Years 6, 7 and 8, when four holdings were sold, and not one rupee of it arrived in the five years before that: constraint 1. The Rs 4,00,00,000 is still callable, and the fund's remaining Rs 20,00,00,000 of unfunded commitment exists to pay the fee and expenses of the six quarters that are left: constraint 2, still running. The Rs 56,40,00,000 is the single largest number on the statement that nobody has paid. That figure is one fifth of Rs 2,82,00,00,000 carried across five holdings, one of which is written down to 0.70 times its cost: constraint 3. And nothing in the table was scheduled by investor 1 or can be rescheduled by it: constraint 4.

The statement omits one line that arguably belongs on it: the five companies the Rs 56,40,00,000 is an estimate of. A statement gives a figure and a basis. A statement does not usually say, in words, five companies, one written down, none sold to anybody. A reader who supplies that sentence for themselves has understood the document better than one who reads the total and stops.

With an hour to spend on a private fund's documents, what is actually worth checking?

The practical end of the subject reaches more people than the signing does: far more people read these documents than ever sign one. A treasury analyst modelling what a group can spend next year, a bank looking at an interest pledged as security, a person on an investment committee reading a paper somebody else wrote, and a household deciding whether an offer being described to them is what they think it is, are all doing versions of the same reading. Five checks do most of the work, and every one of them is about the four constraints rather than about performance.

First, find the unfunded commitment and say out loud whose it is. The unfunded commitment is the single most reliably mangled figure in the subject. In this invented fund the number Rs 20,00,00,000 belongs to the fund and the number Rs 4,00,00,000 belongs to investor 1, and quoting either without its denominator is how a five-times error travels. If the document says unfunded commitment and does not say whose, that is the first question, not a detail to sort out later.

Second, ask how much of the stated value has actually been sold to somebody. For Nilgiri Growth Partners Fund II at the record date, distributions to paid in stand at 0.91 times and residual value to paid in at 0.59 times. The two add to 1.50 times exactly, the total value to paid in on the same denominator. The first of those two numbers is cash that arrived. The second is a valuation. The two figures are printed side by side and are not the same kind of thing. Running that addition is a fast way to see how much of a headline figure is still an opinion.

Third, read the transfer clause rather than assuming it. In this arrangement one investor of twelve negotiated a transfer right and it runs only to entities under its own control. Fourth, read the term and the extension mechanics together: ten years here, with a first extension at the manager's election given the investor advisory committee's prior written consent and a second requiring investors holding more than half of commitments by value, neither taken at the record date. The extension mechanics show who can lengthen the wait and who cannot shorten it. Fifth, ask when the remaining obligations fall due. A commitment that feels finished usually is not: here Rs 20,00,00,000 is still to be called across six quarters, and the four calls after the investment period ended were fees and expenses and nothing else.

Notice what is not on that list. No check there asks for an expected return. No document answers that question, and any figure offered in reply would be somebody's estimate wearing a number's clothes. The five checks are all about what is owed, what is settled, what is still opinion and who holds the clock, the only four things the documents can actually establish.