Fund Administrator: The Operational Backbone of a Private Fund

A fund administrator keeps a private fund's books and strikes its net asset value. The administrator maintains each investor's capital account, processes the capital calls and distributions the manager authorises, holds the register of investors, runs the checks that admit a new investor, and produces the statements that go out. No administrator decides what an unsold holding is worth. The values come from the fund's valuation process, and the administrator records them.

Something visible from any doorstep makes the point. A housing society collects money from every flat, spends it on a lift contract and a security agency, and once a year somebody has to say what each flat paid, what is sitting in the account and what each flat still owes. Where the person who decides what to spend is not the same person who keeps that ledger, nobody reads the arrangement as an accusation against the secretary. Deciding and recording are simply two different jobs, and combining them in one pair of hands quietly removes the only check anybody had.

A private fund is that arrangement at a much larger size, with one extra difficulty stacked on top. A private fund has no exchange to tell it what it is worth and no market maker to settle its trades, so every single number an investor ever sees has been assembled by somebody. There is no closing price to look up and no screen to glance at when the market shuts. When Nilgiri Growth Partners Fund II, invented, tells its twelve investors that it holds Rs 2,82,00,00,000 at its record date, that figure did not arrive from a market. The figure was assembled line by line by named parties, each doing a separate job.

What does a fund administrator actually do?

A fund administratorThe outside firm that keeps a fund's books and strikes its net asset value. is an outside firm the fund appoints to keep its books and to work out, at each reporting date, the value of the fund and each investor's share of that value. In this worked case that firm is Kolar Fund Services Private Limited, invented, and the person who actually strikes the net asset valueWhat the fund's holdings are recorded as being worth, less what it owes. is Ashwin Baliga, invented, its fund controller. He does not work for Nilgiri Alternatives Advisors Private Limited, invented. Nilgiri Alternatives is the manager, and the manager decides what this fund buys and sells. The separation between the two firms is the whole structural idea, and the separation is mechanical rather than moral: the party that computes the figure is not the party whose performance the figure describes.

The work itself is unglamorous, and almost all of it is record keeping. Seven jobs cover it. Which of them Kolar takes, and on what terms, is fixed by this fund's own documents and by nothing else.

Function 5 settles arguments, so it deserves a name of its own. The register of investorsThe record of who holds what interest in the fund. is what the fund treats as true about who holds what. If an investor transfers its interest, the transfer becomes real when the register says so, and not when a letter says so. Function 6 is the other quiet one: an investor cannot simply send money and become an investor. Checks on identity and on where the money came from are completed first. The conditions attaching to a vehicle of this kind in India are set by the Securities and Exchange Board of India at sebi.gov.in.

Now read the seven again and notice what is missing from all of them. Not one of the seven asks anybody to form a view on what a business is worth. The administrator's list is a list of records, movements and arithmetic, and there is no line on it that requires an opinion about a company. Everything else about the administrator follows from that absence.

How does a net asset value actually get struck?

This fund's record date is the end of its Year 9 Quarter 2. At that date, five of its nine holdings have not been sold. Each one sits in the books at a carrying valueThe value a single unsold holding is recorded at.: holding 4 at Rs 1,08,00,00,000, holding 6 at Rs 21,00,00,000, holding 7 at Rs 39,00,00,000, holding 8 at Rs 81,00,00,000 and holding 9 at Rs 33,00,00,000. The five carrying values add to Rs 2,82,00,00,000, and that total is the net asset value the fund reports on that date.

Where did the five come from? Not from Kolar. Every unsold holding of this fund is valued once a year by Palani Valuation Advisors LLP, an invented limited liability partnership (LLP) and the independent valuation agentThe outside firm that forms a view on what unrealised holdings are worth., and the manager marks each holding in the quarters between those annual valuations. Kolar receives the five numbers. The administrator's contribution to a figure of Rs 2,82,00,00,000 is the addition, and the addition is the only part of it that can be checked by looking at anything.

Three details in that sequence are worth pausing on. The first is that a net asset value is a net figure, so anything the fund still owes comes off before it is published. In this fund's own record at the record date, the net asset value is exactly the Rs 2,82,00,00,000 of carrying values, so on that date nothing came off; that is a fact about this invented fund on this date and not a rule about any other. The second is that the allocation across twelve capital accounts is arithmetic on shares the fund's documents fix. Investor 1 holds 20.0 per cent of total commitments of Rs 5,00,00,00,000, so investor 1 is shown Rs 56,40,00,000 of that Rs 2,82,00,00,000. The third is the one readers skip: the statements leave in the manager's name. The administrator produced the numbers and the manager sends them.

Where does the administrator's work stop?

The administrator records values. The administrator does not set them. Everything it can check is a matching exercise: does this equal that, does this trace to a document, does this add up. Everything it cannot check is a judgement about a business it has never run, never visited and has no mandate to form a view on.

An example makes the split concrete. Between this fund's Year 7 year end and its record date, the five holdings that are still held moved from Rs 2,39,50,00,000 to Rs 2,82,00,00,000. Kolar recorded both totals, and it chose neither of them. A reader who trusts the administrator is trusting the handling of the numbers, and the handling is a genuinely different thing from the numbers. Both trusts are worth something. The two trusts are just not the same, and the failure set out below comes from confusing them.

How is the administrator different from the valuation agent and the auditor?

Four parties touch the numbers an investor eventually reads, and telling them apart is most of what a reader of a private fund's reporting needs. The manager decides and marks. The valuation agent values, once a year and independently. The administrator computes and records. The statutory auditorThe firm that audits the fund's annual financial statements. checks afterwards.

The distinction that gets blurred most often is between the second and the third of those. Palani Valuation Advisors LLP forms a view. The valuation agent looks at a business, applies a method, and arrives at a figure it is willing to put its name to; Rohit Vaz, invented, is the partner who signs. Kolar Fund Services forms no view about any business at all. The administrator takes what it is given and does the arithmetic and the record keeping around it. If those two jobs sat inside one firm, the independence of the valuation and the independence of the books would be the same independence, and a reader would have one check where the structure is designed to give two.

The auditor is different again, and the difference is timing as much as substance. An audit happens after a financial year has closed. The audit looks at the annual financial statements the fund has produced and reports on whether they present the fund fairly. An audit is not a running check on the quarterly arithmetic and it is not a second valuation. Three parties, three questions, and none of the three answers another one's question.

Who can actually move the fund's money?

Ask a room of students who can move a fund's money and most of them say the manager. The manager makes the decisions, so the answer sounds right. The decision and the movement are not the same event. In this fund the path is fixed, and it runs through more than one party by construction.

A capital call or a distribution begins as a decision by the manager. The decision becomes an instruction when Farida Contractor, invented, the chief operating officer of the manager, signs the notice. Kolar then checks that notice against the fund's own records, works out each investor's amount, issues the notice and prepares the movement. The money itself sits in accounts that Nilgiri Trusteeship Services Private Limited, invented, holds as trustee. No single one of those parties completes a movement on its own, and that is a property of how the arrangement is built rather than a statement about anyone's character.

What does independence here buy an investor, and what does it not?

Take the two halves separately. The whole misunderstanding lives in the gap between them.

Independence buys something real. A second, separate set of hands does the arithmetic, keeps the ledger and holds the register, and those hands do not get paid more when the fund looks better. If the manager's own back office had produced the Rs 2,82,00,00,000, an investor asking whether the five carrying values had been added correctly would be asking the manager to check the manager. With Kolar in the middle, the addition, the allocation across twelve accounts, the cash reconciliations and the register are all produced by a party whose position does not improve when the fund's reported figures improve.

Independence does not buy everything, and the limit is equally definite. Independence of the administrator secures the arithmetic and secures nothing whatsoever about the estimates that went into it. Kolar can add Rs 1,08,00,00,000 to four other figures flawlessly and still be adding up a set of numbers nobody has tested against a buyer. None of this is a criticism of Kolar. Adding up what it is handed is the job description. An investor who reads an administrator's name and feels comfortable about the marks has taken comfort from the wrong document.

Which way is a report travelling?

Two separate reporting arrangements sit around any fund, and they point in opposite directions. Each portfolio company reports up to the fund, on a cadence of five items that the company's own agreement with the fund fixes; a monthly pack is one of those five. The fund reports out to its twelve investors, on a separate list of six items that the fund's own documents fix; a capital account statement and a notice for every call and every distribution are two of those six. Each list is set out in full separately. The counts alone matter for the comparison.

The administrator produces items on the second list and touches nothing on the first. When a monthly pack lands from Bhavani Speciality Chemicals Private Limited, invented, it was written inside that company and sent to the fund. Kolar never saw it, and asking Kolar about it is like asking the housing society's accountant about the lift contractor's own internal service log.

What does all of this look like on investor 1's own statement?

Every function listed so far ends up as a line on a sheet of paper somebody actually reads. Investor 1 of this fund is a domestic life insurance company with a commitment of Rs 1,00,00,00,000, and that commitment is 20.0 per cent of the fund's total commitments of Rs 5,00,00,00,000. Below is what the administrator's record produces for that one investor at the record date. How a capital account is built, line by line, is covered separately. Every figure in the column below was assembled by a named outside firm.

Two of those lines carry a trap that has caught readers before, and both traps are traps about levels. Investor 1's unfunded commitment is Rs 4,00,00,000, and that is an investor-level figure that must never be confused with the fund-level Rs 20,00,00,000. The fund has drawn Rs 4,80,00,00,000 of its Rs 5,00,00,00,000, leaving Rs 20,00,00,000 unfunded across everybody. Investor 1's 20.0 per cent share of that is Rs 4,00,00,000, the same figure that comes from subtracting its own Rs 96,00,00,000 from its own Rs 1,00,00,00,000. Writing the fund's figure on this investor's statement overstates what it still owes by five times.

The second trap is the multiple. Investor 1's Rs 1,44,00,00,000 of total value against the Rs 96,00,00,000 it has actually paid in is 1.50 times, and that denominator is what it paid in and nothing else. A multiple on what the holdings cost would use a different denominator on a different level again. And there is one line the statement does not carry that a careful reader supplies themselves: Rs 56,40,00,000 of that Rs 1,44,00,00,000, being 39.2 per cent of it, is a share of estimates about five companies, not one of which has been sold to anybody.

What does an investor test before it commits?

The account so far has been descriptive; what follows is something a working person actually does. An investor deciding whether to commit money to a manager runs two quite separate examinations. One is about the investments. The other, operational diligenceThe examination an investor runs on how a manager actually operates, before committing., is about how the operation is built, and the administrator sits inside it. The phrase gets used for two different exercises, so it is worth saying which one is meant every time: a manager runs operational diligence on a company it is thinking of buying, and an investor runs operational diligence on a manager it is thinking of backing. An investor's examination of a manager is the second of the two.

Investor 5 of this fund is a fund of funds, an investor whose whole business is choosing managers rather than choosing companies. Before it committed to Nilgiri Growth Partners Fund II it tested five things: whether the administrator was independent of the manager, how the valuation process worked, who could move cash and how many people it took, what the business continuityThe arrangement for keeping records and payments running if something fails. arrangement was, and whether the auditor had ever issued anything other than a clean opinion.

Read that list and notice that not one of the five questions asks whether anybody is honest. A construction can be examined by an outsider in an afternoon and a character cannot be examined at all, so each of the five asks how the arrangement is built. The cash question is the sharpest of the five. The cash question does not ask whether the people are trustworthy. The cash question asks how many separate hands sit between an instruction and money leaving an account, and that count stays true whoever is sitting in the chairs.