Investment Period and Harvest Period: Capital Out, Then Back

The investment period is a contracted stretch, fixed in the fund's own documents, during which capital may be called to buy something new. The harvest period follows, and it is a description rather than a clause. Nilgiri Growth Partners Fund II, invented, ran a five year investment period, and thirteen of its seventeen calls fall inside it, accounting for Rs 4,55,50,00,000 of the Rs 4,80,00,00,000 it has ever drawn.

Almost everything a reader notices about the middle years of a closed-end private fund comes from one restriction, and the restriction is a sentence in a document rather than a habit of the people running it. For a defined stretch the fund may spend on new things. After that stretch it may not. The two halves of a fund's working life are contractual periods with different permissions attached, not descriptions of what the manager happens to be busy with. With that one idea in place, everything that follows is consequences: the shape of the calls, the size of the fee, what can be done for a company that suddenly needs more money, and why an investor who thinks the calls have stopped is about to be surprised by a notice.

Before any arithmetic, one convention, stated once so that every elapsed figure below can be checked. Time here is counted from each fund's own final close, never in calendar years. Year 1 Q1 sits 0.25 years after the close, Year 4 Q2 sits 3.50 years after it, and in general Year n Q m sits at n minus 1 plus 0.25 times m years. Every date below is written out as Year 4 Q2 and never in any shorter form, inside the figures as well as in the prose. Nilgiri Growth Partners Fund I closed four years before Nilgiri Growth Partners Fund II and the two clocks do not line up, so where two funds are mentioned in the same breath the fund is named with the date.

What is an investment period, and who fixes how long it runs?

Consider a household that has saved for years and finally decides to build a house on a plot it holds. The household hands the job to a builder and writes one thing into the contract: for the next two years the builder may start whatever work it judges necessary, and after those two years it may finish what it has started, pay the watchman and settle the electricity bill, but it may not begin a new wing. Nothing about the builder's skill or intentions changed at the two year mark. The permission did. On the day before, a new foundation was a legitimate use of the household's money. On the day after, it was a breach.

An investment periodThe contracted stretch during which capital may be called to buy something new. in a private fund is that clause, at scale, written by people who expect to be held to it. The clause fixes the window during which the manager may issue a capital callA notice requiring investors to pay part of what they have committed. for the purpose of acquiring something the fund does not already hold. Its length is not a market convention, not a regulatory setting and not the manager's choice made afresh each year. The length is a negotiated term in the fund's own constitutional documents, agreed before a single rupee was called, and a reader who wants to know how long it runs has to read that document rather than assume anything.

Nilgiri Growth Partners Fund II is a closed-end growth and buyout fund managed by Nilgiri Alternatives Advisors Private Limited. The fund is settled as a trust, with Nilgiri Trusteeship Services Private Limited as trustee and Nilgiri Financial Holdings Private Limited as sponsor, and the role a global investor would call the general partner is discharged by the manager and the trustee between them. Its investors committed Rs 4,90,00,00,000 across twelve of them, the manager committed a further Rs 10,00,00,000 of its own money, and total commitments are therefore Rs 5,00,00,00,000. Its contracted term is ten years from final close. Its investment period is five years from final close, ending at the end of its Year 5.

Five and ten are the contracted terms of this one fund, not a market standard and not a regulatory setting. Another fund's documents will say something else entirely, and a reader who carries five and ten away as though they were the shape of the subject has learned the wrong lesson. The relationship is what is worth carrying away: a fund of this kind has a term, and inside that term it has a shorter window for buying, and the two are separate clauses that can be set independently of one another.

What does the end of the investment period actually stop?

The end of an investment period stops one thing precisely, and a reader who widens it by even a little will get the next several years wrong. The permission that ends is the permission to call capital in order to acquire something the fund does not already hold. Nothing else ends.

Notice what does not end. The fund does not end: Nilgiri Growth Partners Fund II has a ten year term and its investment period is five, so half the contracted life is still ahead when the buying stops. The management fee does not end. The expenses do not end. The reporting to investors does not end. The seats on portfolio company boards do not vacate themselves, the veto rights written into shareholders agreements do not lapse, and the manager's obligation to look after nine companies does not soften. The end of an investment period changes what money may be spent on and changes nothing whatever about the work still owed.

And the calls do not stop. The belief that the calls stop is the single most common misreading of the whole subject, and it deserves its own sentence. An investor who hears that the investment period has closed and quietly treats its remaining promise as released has told itself it holds cash it does not hold. The promise a fund holds over an investor and has not yet drawn is that investor's unfunded commitmentThe part of an investor's promise contracted for but not yet called., and nothing about the end of the buying window releases it. Nilgiri Growth Partners Fund II called four more times after its Year 5, and at its record date its investors between them still hold Rs 20,00,00,000 of unfunded commitment, being 4.0 per cent of the Rs 5,00,00,00,000 of total commitments. The uncalled Rs 20,00,00,000 is there for the fee and the expenses of the six quarters that remain.

What may capital still be called for once the buying stops?

Four things, and the fund's own documents number them. The four are not a general rule about private funds and not a list anybody may extend on the day. Nilgiri Growth Partners Fund II wrote the four down before it called a rupee, and every call it has issued since the end of its Year 5 falls inside them.

The first is follow-on investmentMore money into a holding the fund already has, rather than into a new one. in a holding the fund already has, capped in total at 15.0 per cent of commitments. The second is the management fee. The third is fund expenses. The fourth is an obligation already committed under a signed agreement. Without that fourth head a fund would be trapped between a contract it signed on the last day of the period and a permission that lapsed the following morning.

Read the four together and a shape appears. Three of the four are about honouring things the fund has already got itself into, and only one of them lets new money reach a company at all. The fee and the expenses are the cost of continuing to exist. A signed obligation is a promise made while the promise was still permitted. And the follow-on head, the only one that puts fresh money to work, is deliberately fenced: it may go to a company the fund already holds, and it may not exceed a stated share of commitments. The follow-on head is defined by whether the fund is already a holder, not by how the money is spent. A manager therefore cannot start something new by calling the new investment a follow-on.

On this fund the cap works out as follows, and the denominator has to be named or the number means nothing. Fifteen per cent of the Rs 5,00,00,00,000 of total commitments is Rs 75,00,00,000. Which commitments a cap of this sort attaches to, whether the whole Rs 5,00,00,00,000 including the manager's own Rs 10,00,00,000 or only the Rs 4,90,00,00,000 the twelve investors promised, is a question the fund's documents answer and the reader has to look up rather than assume. Fund II's documents fix the denominator as total commitments, and the Rs 5,00,00,00,000 figure is the one used throughout.

Is the harvest period a clause, or just a name for what is left?

Harvest Period

A search of the documents does not turn it up. There is a defined term called the investment period, with a stated length and a stated end, and there are the four heads under which capital may be called afterwards. No clause anywhere establishes a harvest period. The documents do not create one: the harvest period is the name people give to everything left over once the investment period has ended.

The distinction sounds like pedantry until the loose usage is examined for what it smuggles in. Calling those years the harvest implies that harvesting is what happens in them and that money comes back. Nothing in any document promises that. The harvest periodEverything after the investment period, described by what the fund does rather than by a separate clause. is a description of a tendency, and a tendency is not a right. There was nothing to breach, so a fund can sit in its so-called harvest years with nothing selling and nothing coming back and will not have breached anything.

Consider a mango orchard that a household planted eight years ago. There is a stretch of planting and a stretch of picking, and everyone in the house talks about the picking season as though it were fixed. But nobody signed anything about the picking. The planting was what was fixed: the plot only had room for so many trees, and the planting was finished by a certain year. Whether fruit arrives, and how much, is not part of that arrangement at all. One tree may give nothing for a season and nobody is in breach.

Nilgiri Growth Partners Fund II shows this exactly. Across its first five years, the whole of its investment period, it distributed nothing at all to its investors. Not one rupee. Its four distributions all fall in its Years 6, 7 and 8, and in its Year 9 to the record date it distributed nothing again. So the harvest, on this fund, was three years long inside a five year stretch that had already run half its course, and the two quarters immediately before the record date were as quiet on the returning side as any year of the investment period had been. Of the nine holdings, one left the portfolio at Fund II's Year 6 Q4 producing no cash at all. Leaving the portfolio and producing money are two different events.

The full year by year curve of what this fund had drawn, what it had returned and what it was carrying is worked out separately. Two points from it are used here: the end of Fund II's Year 5, where cumulative capital paid in stood at Rs 4,55,50,00,000 against cumulative distributions of nothing at all, and its record date at Year 9 Q2, where cumulative paid in stood at Rs 4,80,00,00,000 against cumulative distributions of Rs 4,38,00,00,000. Two readings are enough to see the two halves of the working life in one glance, and anything further belongs elsewhere.

What shape did this fund's seventeen calls actually take?

A closed-end fund does not take its investors' money on day one. A closed-end fund takes a promise on day one and then draws against that promise over years. The drawing against a promise is why the calls have a shape at all. Nilgiri Growth Partners Fund II has issued seventeen calls to the record date, every one of them to all investors in proportion to what each had promised, and every one met in full. Together they add to Rs 4,80,00,00,000, being 96.0 per cent of the Rs 5,00,00,00,000 of total commitments.

Now split them at the end of Year 5. Thirteen calls fall inside the investment period and four fall after it. The thirteen account for Rs 4,55,50,00,000, being 94.9 per cent of everything this fund has ever drawn, so half the contracted term carried almost all of the drawing. The remaining four, spread across the three and a half years from the start of Year 6 to the record date, come to Rs 24,50,00,000, which is the other 5.1 per cent.

In the picture below the lopsidedness is the first thing to register. The tall bars are all on the left. The last four bars barely lift off the baseline, and they are getting shorter as they go. The taper is not a fund running out of ideas. The taper is the mechanical result of two clauses working at once: nothing new may be bought, and the fee is charged on a base that is itself shrinking. Both of those were written down before the fund drew a rupee.

What were the last four calls, and why are they so small?

Here they are, taken apart. Each of the four is a single call made in the first quarter of a year, and each one is the year's running cost being collected in one go rather than a fund doing anything with the money.

Three details in that table repay a second look. Only half of Fund II's Year 9 had run at the record date, so the Year 9 fee line is Rs 1,80,00,000 rather than a full year's charge and the fund called what had accrued. The Year 9 operating expenses are Rs 40,00,000 for the same reason, being half of the Rs 80,00,000 a year the fund had been drawing. And the Year 8 call carries an extra Rs 50,00,000 of transaction expenses, which arose from a realisation in that year and is the only line in the four that is not simply the cost of continuing to exist.

Not one rupee of the Rs 24,50,00,000 bought anything, and that is the clearest statement of what the end of an investment period does. The calls did not stop. Their purpose changed, completely and on a date fixed years earlier.

One number in the table above is a coincidence, and a reader who does not notice will build a false pattern on it. Fund II's annual management fee during the investment period was Rs 9,80,00,000. Its total fund expenses across the whole nine years to the record date, organisational and operating together, were also Rs 9,80,00,000. The two figures are equal by accident. The fee and the expenses measure completely different things, and there is no relationship between them at all.

Why does the fee fall when the rate never moves?

The fall in the fee is the part of the subject most often taught badly, and it is taught badly because the interesting idea is hidden inside a word most readers slide past. A percentage on its own is not an amount. A percentage only becomes an amount when somebody names what it is a percentage of. The amount it is applied to is the fee basisThe amount the management fee percentage is applied to, which can change on a stated date.. A fund's documents can and often do arrange for the basis to change on a stated date. The percentage sits perfectly still.

Nilgiri Growth Partners Fund II charges a management fee of 2.00 per cent a year. The 2.00 per cent is the same in its Year 1 and in its Year 9 and was never renegotiated. The denominator changes, and it changes once, at the start of Year 6. The start of Year 6 is the first day after the investment period ends.

During the investment period the 2.00 per cent runs on aggregate investor commitments of Rs 4,90,00,00,000. Commitments do not move, so two per cent of that is Rs 9,80,00,000 a year and it is the same figure in each of the five years. Note which number that is and which it is not: it is the Rs 4,90,00,00,000 the twelve investors promised, not the Rs 5,00,00,00,000 of total commitments, because the manager's own Rs 10,00,00,000 bears no fee. The exclusion of the manager's own money is why the annual charge is Rs 9,80,00,000 rather than Rs 10,00,00,000, and a reader who reaches for the wrong denominator will be Rs 20,00,000 a year out and will never find the error.

From Year 6 the 2.00 per cent runs on the acquisition costWhat the fund paid for a holding, before any change in its carrying value. of holdings not yet realised, measured at the start of each year. Acquisition cost is a completely different denominator: it is what the fund paid for the companies it still holds, so it falls every time a holding leaves the portfolio and it never rises again once the buying has stopped. At the start of Year 9 that base stood at Rs 1,80,00,00,000, and 2.00 per cent of it is Rs 3,60,00,000 for the year, of which half had run at the record date.

Set the two endpoints beside each other and the arithmetic does the teaching. Rs 9,80,00,000 in Year 5 and Rs 3,60,00,000 in Year 9. The second is 36.7 per cent of the first. Not one term was renegotiated, no investor pressed for a discount and no manager offered one. The step-downA contracted change in the fee basis that happens on a date the documents already named. happened because the contract had always said it would, on a date everybody signing knew years in advance.

Can the fund still put money into a company it already holds?

Yes, within limits the documents set, and this is the point at which a great many readers quietly go wrong. New investment stops when the investment period ends. Follow-on investment does not necessarily stop, and whether it does is a question about a particular fund's documents rather than about private funds in general.

Think about why a fence is needed at all. A company the fund already holds may need more money for a perfectly good reason: a larger order than it can finance, a competitor stumbling, a supplier demanding cash up front. If the fund could never put another rupee in after Year 5, a holding could be starved for reasons that had nothing to do with the business. But if the fund could put in unlimited amounts, a manager could keep deploying capital indefinitely and simply call it support for what it already has, and the end of the investment period would mean almost nothing.

So the documents keep the permission and put a ceiling on it. Nilgiri Growth Partners Fund II sets that ceiling at 15.0 per cent of commitments, and on its Rs 5,00,00,00,000 of total commitments the ceiling is Rs 75,00,00,000. The 15.0 per cent is this one fund's own contracted figure. The cap is not a market standard and not a regulatory setting, and another fund's documents will set a different one or none at all.

Now the fact that makes this fund a good teaching case, and it is the opposite of what a reader expects. Fund II made exactly two follow-on investments in its whole life: Rs 15,00,00,000 into holding 1, Sahyadri Diagnostics Private Limited, invented, at Fund II's Year 4 Q1, and Rs 10,00,00,000 into holding 4, Bhavani Speciality Chemicals Private Limited, invented, at its Year 5 Q2. Both fell inside the investment period. Together they are Rs 25,00,00,000, being 5.0 per cent of commitments. And because the cap governs follow-on investment made after the period ends, and both of these were made before it ended, this fund has never drawn against that cap at all. Rs 0 of Rs 75,00,00,000.

A cap named on its own implies a constraint that bit at some point. The Rs 75,00,00,000 cap on this fund never bit at all. There is a second detail worth pausing on: both follow-ons fell exactly 2.50 years after their holding's entry, holding 1 having been entered at Fund II's Year 1 Q3 and holding 4 at its Year 2 Q4. The matching gap is a coincidence inside one small record and not a pattern. Two observations cannot establish when a company tends to need more money.

Two things this record does not settle. First, it fixes the cap as 15.0 per cent of commitments but does not spell out whether that means the whole Rs 5,00,00,00,000 or only the Rs 4,90,00,00,000 the twelve investors promised, so a reader looking at any real fund has to find that sentence in that fund's own documents. Second, nothing in this record says whether the investment period could have been brought to an early end, or by whom, or on what vote. Both questions have answers, and the answers live in the documents of each particular fund rather than in any general rule.

Do the buying years and the returning years ever overlap?

On this fund, not once. Every one of the nine holdings was entered between Fund II's Year 1 Q3 and its Year 5 Q3, so every entry sits inside the investment period. Every one of the five realisation events falls between its Year 6 Q3 and its Year 8 Q3, so every one sits outside it. The two sets do not overlap by a single quarter, and that is the cleanest picture of what an investment period and a harvest period actually are that this record has to offer.

The gap is checkable, so count it precisely. The last entry sits at Year 5 Q3, 4.75 years after final close on this fund's clock. The first realisation event sits at Year 6 Q3, 5.75 years after it. Exactly one year apart, with three quarters in between carrying neither an entry nor an exit. The end of the investment period, at 5.00 years, falls inside that quiet stretch.

Two cautions before anybody generalises. First, a fund of this kind is under no obligation to arrange its life this way and many do not: a holding bought early can be realised while the fund is still buying, and there is nothing in the documents preventing it. The documents fix the buying window, not the selling. Second, of the five realisation events on this fund, one produced no money whatever. Holding 5, Palar Foods Private Limited, invented, left the portfolio at Fund II's Year 6 Q4 and the fund received nothing for it. So even a fund whose entries and exits sit in tidy blocks does not have a clean story, and a reader who reads the tidiness as success has read the picture rather than the record.

How long was anything actually held on this fund?

Four of the nine holdings had been fully realised by the record date, and on this fund's own clock they were held 5.75 years, 4.75 years, 5.75 years and 3.75 years. The four average exactly 5.00 years. Five years is a tidy figure and it is tempting to do something with it, so here is the warning that has to travel with it. Five years is the arithmetic mean of four numbers from one invented fund, and it is not a statement about how long anything is usually held. Four observations settle nothing, three of the four sit above or below the mean by a full year, and five holdings on this fund had not been realised at all when the record date arrived, so the four that had are not even a complete picture of this one fund.

The mean of the four is good for a different question, and a more useful one. Notice that a holding entered in the last quarter of the investment period and held for five years will be realised roughly five years into a stretch that has only five years left in it. A five year hold inside a five year remainder is the arithmetic pressure sitting behind the whole back half of a closed-end fund's life, and it is why the returning years feel crowded in a way the buying years never do. Nothing about the crowding is a criticism of anybody. The crowding is the shape a fixed term and a fixed buying window produce between them.

How do the two periods differ once they are set side by side?

Investment Period vs Harvest Period

Set them out in a row and the difference stops being a matter of atmosphere. The two periods are not two moods a manager passes through but two states of a contract, and four things change between them.

The last row carries the asymmetry that all of this establishes. One side of the comparison is written in the documents and the other is not. The investment period is drafted, negotiated and dated. The years after it are simply the residue, given a name by people who needed something to call them. The asymmetry is why a reader can look up exactly when the buying stops and cannot look up when the returning starts.

What would somebody actually do with this clock on a Monday morning?

Three different people read the same two dates for three different reasons. Being concrete about each matters: the mechanism stops being a diagram at this point and starts costing or saving somebody money.

An investor's treasury reads it as a cash planning instrument. A treasury may have promised Rs 1,00,00,00,000, as investor 1 of this fund did, and it needs to know when that money is likely to be wanted. The investment period tells the treasury that the bulk of the calls will land inside a known window. On this fund that window carried 94.9 per cent of everything drawn. The four permitted heads add that the tail after the window is small, predictable and made of running costs rather than sudden purchases. The clock does not tell an investor when money will be called, but it tells it what the money can be called for, and that turns an open-ended obligation into a bounded one. None of the unfunded amount is released, so the treasury still has to hold the whole of it available, but it can hold the tail differently from the way it holds the first five years.

An analyst reads it as a context for anything else the fund reports. A fund three years into its investment period and a fund three years past the end of it are not comparable on any measure of drawing or fee, and an analyst who does not first establish where in the clock a fund is standing will misread everything downstream of it. Two simple questions do most of the work: how many years of the investment period have run, and what is the fee sitting on right now. On this fund at its record date the answers are that the investment period ended three and a half years ago and the fee sits on Rs 1,80,00,00,000 of unrealised acquisition cost.

The manager's own team reads it as a constraint on what it may propose. A company the fund does not already hold cannot be bought after Year 5, however attractive the case for it, and no amount of internal enthusiasm changes that. The team can propose support for what the fund already holds, inside a stated cap, and a constraint and a conflict sit next to each other at exactly that point. A manager whose buying window has closed has an incentive to describe a new opportunity as an extension of an existing holding. The permitted head is defined by whether the fund is already a holder, not by how the money is used, and that definition makes the description harder to sustain.