Look-Through Reporting: Seeing the Underlying Holdings

Look-through reporting shows an investor the businesses sitting underneath its own single line, one by one, instead of one total. Investor 5 of Nilgiri Growth Partners Fund II, invented, committed Rs 50,00,00,000 against total commitments of Rs 5,00,00,00,000, so its 10.0 per cent share of the Rs 2,82,00,00,000 still held at the end of Year 9 Quarter 2 is Rs 28,20,00,000, spread across five businesses.

A small example holds the whole thing. Ten neighbours put money into a pooled kitty each month and one of them keeps the accounts. At the end of the year every neighbour is handed a slip that reads: this neighbour's share of the kitty's value is Rs 4,200. The slip is complete. The slip is also correct. Nobody has been misled by it and nothing has been left out of it. And yet the neighbour reading it has no idea whether the kitty is sitting on a fixed deposit, a stack of gold coins, or a loan to somebody's cousin who has stopped answering the phone. One number can be entirely honest and still say nothing about what it is made of, and closing that gap is what look-through reporting does.

What is look-through reporting, and what does it show?

A private fund reports to each of its investors, and one of the things it reports is that investor's share of what the fund still holds. The share arrives as a figure. A look-through scheduleA statement listing each underlying holding separately, with the investor's own share of each one, rather than a single combined figure. is the same figure taken apart: the same rupees, listed against the individual businesses that produce them, so the reader can see the parts rather than the sum.

Splitting the total adds nothing. The total was already exact and it stays exactly the same. Only the resolution changes. Before the schedule, the reader has one number. After it, the reader has five numbers that add back to that one number, plus the names attached to them. Look-through, then, is not a better report and it is not a worse one. Look-through is an arrangement that lets a reader at one level see past the level immediately below it, and every level it crosses is a level of adding-up being undone.

The invented arrangement worked here has three levels in it. Most explanations of this subject stop at two. Nilgiri Growth Partners Fund II, invented, is managed by Nilgiri Alternatives Advisors Private Limited, with Nilgiri Trusteeship Services Private Limited as trustee and Nilgiri Financial Holdings Private Limited as sponsor. In this vehicle there is no limited partnership and no general partner as a matter of Indian law: the vehicle is settled as a trust, the contract is a trust deed and a contribution agreement, and the role a general partner plays elsewhere is discharged by the manager and the trustee between them. The economic vocabulary the documents actually use, capital account and carried interest and the rest, was designed in the other form and imported wholesale, so both vocabularies sit side by side and neither is a mistake.

Investor 5 of that fund is a fund of fundsAn investor whose own investors have put money in so that it can commit to other funds rather than directly to businesses.. Investor 5 committed Rs 50,00,00,000. A fund of funds has investors of its own, and those investors sit one level further out again. So there are three levels here and each one sees a different amount of detail. The investors of investor 5 see a single line for its stake in Nilgiri Growth Partners Fund II, invented. Investor 5 itself sees a capital account statementThe periodic statement showing one investor's own contributions, distributions and share of value in a fund. What it contains line by line is covered separately. from the fund and, if the arrangement provides for it, the nine holdings behind the fund's numbers. The fund bought the nine businesses, so the fund sees them without asking anybody.

How is an investor's share of each underlying holding computed?

By one division, done once, followed by one multiplication per holding. One division and one multiplication per holding is genuinely the whole of it, and the reason so few readers ever do the arithmetic is not difficulty but habit.

The division first. Investor 5 committed Rs 50,00,00,000. The fund's total commitments are Rs 5,00,00,00,000. Fifty crore over five hundred crore is 10.0 per cent exactly. The denominator is the single thing here most likely to be got wrong. The Rs 5,00,00,00,000 of total commitments includes the manager's own Rs 10,00,00,000. Commitments from the twelve investors alone come to Rs 4,90,00,00,000. Against that smaller denominator investor 5 would be 10.2 per cent instead of 10.0. Both divisions are arithmetically fine, and only one of them is the key the fund actually draws capital on. A share percentage with no denominator named beside it is not a fact yet. Nilgiri Growth Partners Fund II calls capital on total commitments of Rs 5,00,00,00,000, and every figure worked below uses that key.

The multiplication next. Drawing in this fund is strictly pro rataDivided in proportion to what each party committed, with nobody treated differently in the division.. A share of commitments is therefore a share of every single thing the fund holds. Nobody needs to work out which investor's money bought which business, and the vehicle does not divide its holdings that way. One percentage covers everything.

Five businesses remain unrealisedStill held rather than sold, so its stated value is an estimate rather than a price somebody paid. at the end of Year 9 Quarter 2, and the fund's residual valueThe part of a fund's value that is still held rather than already returned to investors in cash. is the sum of their carrying values: Rs 1,08,00,00,000, Rs 21,00,00,000, Rs 39,00,00,000, Rs 81,00,00,000 and Rs 33,00,00,000, adding to Rs 2,82,00,00,000. Applied to each, 10.0 per cent gives investor 5's schedule. Applied to the sum, it gives investor 5's single line. The two must agree, and they do.

A reader who checks once never has to trust again. Check the addition yourself, in crore: 10.80 plus 2.10 plus 3.90 plus 8.10 plus 3.30 is 28.20. The cost column adds to Rs 1,80,00,00,000, and the 1.57 times on the total row is Rs 2,82,00,00,000 divided by that Rs 1,80,00,00,000, derived here rather than reported by anybody, for this invented fund at this one date. Notice also that investor 5's own line reads Rs 48,00,00,000 contributed and Rs 43,80,00,000 of distributions received, so its total value of Rs 72,00,00,000 is 1.50 times what it paid in, for this invented fund to the end of Year 9 Quarter 2. Pro rata drawing hands every investor the same arithmetic, and the 1.50 is the fund's own figure, unchanged.

What does the total conceal that the five parts show?

Shape. Nothing else, and nothing less.

Take the fund's Rs 2,82,00,00,000 of residual value at the end of Year 9 Quarter 2 and ask what proportion of it each holding carries. Holding 4 is Rs 1,08,00,00,000 of it, being 38.3 per cent. Holding 8 is Rs 81,00,00,000, being 28.7 per cent. Holding 7 is 13.8 per cent, holding 9 is 11.7 per cent, and holding 6 is 7.4 per cent. The five rounded figures add to 99.9 and the unrounded ones add to 100.0. Running that check every time costs nothing and catches a slipped decimal. Two holdings of the five carry 67.0 per cent of everything the fund still holds.

Two holdings carrying two thirds of the value is a fact about concentrationHow much of a total sits in how few of its parts., and a fact about concentration is not yet a verdict on it. No percentage on its own makes a portfolio high, acceptable, risky or strong. A verdict on a portfolio needs facts about the businesses, and a percentage cannot supply them. The figure stands with the invented fund and the date attached, and no more. Two holdings carrying two thirds of the value could describe a manager who backed a winner twice, or a manager who has lost everything else, and the schedule says nothing about which. The schedule puts the question in front of the reader instead of hiding it inside a sum. The contribution is a small one and it is worth having.

The household version runs like this. Two people run a shop and state that their stock is worth Rs 8,00,000. Fine. Then they show the shelves, and it turns out Rs 5,36,000 of that stock is one item, ordered in bulk, sitting in the back. The total did not change when the shelves came into view. The questions changed completely, and every one of the new ones is about that one item rather than about the Rs 8,00,000.

What happens when there are two layers rather than one?

The multiplications stack, and the second one is the one almost nobody performs.

Investor 5 is a fund of funds, so its own investors are two steps away from the businesses. Suppose one of them holds 10.0 per cent of investor 5. Its share of Nilgiri Growth Partners Fund II is 10.0 per cent of 10.0 per cent, or 1.0 per cent. Adding the two would give 20 per cent. Forgetting the second layer would give 10 per cent. Neither number is the share. Layers multiply, and each layer crossed shrinks the share by the factor of that layer.

Apply the 1.0 per cent and the whole chain resolves. The outer investor's indirect share of holding 4's Rs 1,08,00,00,000 is Rs 1,08,00,000. Of holding 6's Rs 21,00,00,000 it is Rs 21,00,000. Of holding 7 it is Rs 39,00,000, of holding 8 it is Rs 81,00,000, and of holding 9 it is Rs 33,00,000. The five amounts add to Rs 2,82,00,000, being 1.0 per cent of the fund's Rs 2,82,00,00,000 of residual value at the end of Year 9 Quarter 2, exactly as they must.

Two multiplications, and here is why the second one matters more than its arithmetic suggests. An institution that has committed to eight funds of funds, each of which has committed to a dozen funds, is somewhere behind roughly a hundred pooled vehicles and several hundred businesses. Nothing in that structure is hidden and nothing about it is unusual. But the report that lands on its desk has had the adding-up done twice, and undoing it twice takes a request at each level, an arrangement that permits the request, and somebody willing to key the answer into a spreadsheet. The arithmetic is trivial. The plumbing is not, and the plumbing is why the second multiplication so often goes unperformed.

What does a look-through schedule still not tell?

One limit survives every level of detail.

Look-through does not turn an estimate into a fact. Seeing the five holdings individually shows the parts the estimate is made of. Not one of those parts has been sold to anybody, so the schedule cannot show that any of them is worth what it says. Holding 4 is carried at Rs 1,08,00,00,000. The Rs 1,08,00,00,000 is a considered figure, struck by a valuation process covered separately, and it remains a figure that no buyer has agreed to. Splitting it out from a total makes it more legible. The split does not make it more certain by a single rupee. A reader who feels better about a number because five of them are now visible instead of one has confused resolution with verification, and those are different things.

Three specific absences follow from that, and it is worth naming them one at a time. First, a look-through schedule does not give a price. A schedule gives a carrying value: the fund's own statement of what a holding is worth on a stated day. Second, it does not give a date unless the schedule carries a valuation-date column, and a value with no date attached cannot be reasoned about at all. One point settled earlier bears on this: a mark can move well after the event that justified it, and a schedule with no as-at column silently hides which quarter each figure belongs to. Why marks move late is covered separately. Third, it does not give cost unless a cost column is there. Without a cost column, nothing shows that holding 6 is carried at 0.70 times the Rs 30,00,00,000 paid for it, and the 0.70 is the single most informative line in this fund's schedule.

There is a fourth absence that only shows up on a count. The look-through of this fund shows five holdings. Nine were bought. Holding 5, Palar Foods Private Limited, invented, was written off in full at Year 6 Quarter 4 and returned nothing at all, so it appears nowhere in a schedule of what is still held. The omission is not concealment: a residual schedule is a schedule of the residual, and a business worth nothing has no residual to report. But a reader who takes the five lines as the fund's history has quietly dropped the one holding that went to zero, along with three that were sold. Resolution on what remains is not the same as a record of what happened.

What exposure can a look-through hide completely?

Investor 5 held no co-investment at the end of Year 9 Quarter 2. The exposure described below is one its side letter permits rather than one it has taken.

Why would a manager report a total rather than the holdings?

Because both are legitimate reports, and because the second one costs something the first does not.

Take the reasons in turn, without deciding between them. A fund holds unlisted businesses, and the carrying value of an unlisted business is commercially sensitive to that business, to its competitors and to whoever might buy it later. A schedule naming each holding and its value distributes that information to every investor of the fund and, in practice, to whoever those investors talk to. Second, a look-through schedule is produced by people. Someone has to strike a value for each holding, reconcile the shares, and stand behind the columns, and that is work the fund pays for out of its own expenses. Third, what any given fund reports is fixed by its own documents, agreed before anybody committed a rupee, and a manager reporting a total is doing what was agreed rather than withholding something.

The reader loses shape rather than accuracy, and the loss is complete: from a single figure of Rs 28,20,00,000 there is no way at all to recover that two of the five holdings behind it carry 67.0 per cent of it. The shape cannot be estimated, bounded, or worked backwards to. The information is not compressed, it is absent. The trade is real in both directions, and which side of it a manager or a reader ought to prefer turns on facts no schedule carries.

Nilgiri Growth Partners Fund II, invented, reports nine holdings by name with cost and carrying value. Reporting a total and nothing else would have been the other available choice. How reporting practice differs across managers is covered separately, at the end of this sequence.

What does a reader actually do with a look-through once they have one?

The people who read these schedules for a living work in a fixed order rather than by judgement. The case is the analyst inside a fund of funds like investor 5. The institution has committed to a number of private funds, and every quarter a stack of reporting arrives. The analyst is not trying to decide anything on the strength of a schedule. The aim is to answer four questions about it, and only then does anything else start.

First, they read the column headings, not the numbers. A schedule with five columns, being holding name, cost, carrying value, the share applied and the investor's own share of each, supports arithmetic. A schedule with two columns supports almost none of it. Whether a look-through is useful is settled by which columns are present, and that question is answered before a single figure is read.

Second, they look for the as-at date, the column most often missing and the one that changes the meaning of everything to its left. Third, they add the investor's share column and check that it equals the single line on the capital account statement. If it does not, something is wrong with one of the two, and finding out which is the whole afternoon. Fourth, they look for what is not there: businesses the same institution holds through another fund, or directly, or beside this fund under a co-investment arrangement, none of which the schedule can see.

A household version of the same discipline. A hospital bill is not read starting from the total. The first check is whether the bill is itemised at all, then whether each item carries a date, then whether the items add to the total, and only then what is on the list. Same four moves, in the same order, and for the same reason: the structure of the document sets which questions it can survive.