How Conflicts of Interest Can Arise in Private Funds

A conflict of interest in a private fund is any decision where the manager's interest and the investors' interest point in different directions. Four of them are structural. Allocating one opportunity between two of its own vehicles, selling a holding to a vehicle it also runs, taking fees at a company the fund has a stake in, and funding a holding that is struggling. Hiring better people removes none of them.

Start away from finance altogether. A wedding caterer takes two bookings for the same evening. He has one kitchen, one senior cook and one van. Whichever household gets the senior cook, the other household does not. Notice when the problem appeared: it appeared the moment he accepted the second booking, before he cooked anything, before he made a single decision, and without anybody behaving badly at all. He is not a bad caterer. He is a caterer standing on both sides of one evening.

Now notice what actually settles it. The fix is not the caterer promising to be even handed, and not him telling both households about the other booking. The fix is the line he wrote into both contracts at the time of signing: the first confirmed booking gets the senior cook. He wrote the line when he did not yet know which household would pay more, complain louder or become the bigger customer. A conflict is answered by a process fixed before the decision arrives, and never by an intention held at the moment the decision arrives.

Every conflict below runs in the same order. The conflict comes first, in plain structural terms, and the process that answers it comes only afterwards. An account that leads with the process leaves a reader reassured. An account that leads with the conflict leaves a reader able to name the purpose of each process.

What is a conflict of interest in a private fund, and what is it not?

A conflict of interest is a decision where the party deciding has an interest of its own that points somewhere other than where the investors' interest points. The definition stops there. A conflict says nothing about anybody's character, nothing about what was decided, and nothing about whether the arrangement is well run.

Why does it arise at all in this setting? Because of three ordinary facts about how a private fund manager is set up. One firm runs several pools of other people's money at the same time. The firm is paid in more than one way, so a rupee reaching it through one route is not the same as a rupee reaching it through another. And the things the pools hold have no traded price, so somebody inside the arrangement has to produce the number that everything else hangs on. Put those three together and the conflicts follow arithmetically. The four conflicts are not a side effect of the arrangement. They are the arrangement.

The distinction a reader most often collapses is the one between a conflict and a failing, so it is worth stating both sides of it clearly. A conflict is a property of an arrangement, and it exists before anybody has done anything. A failing is something somebody actually did, and it exists only afterwards. The one useful question is what process was fixed before the decision arrived, and treating a conflict as evidence of a failing stops that question ever getting asked.

Why is structural the right word for these four?

Structural means the conflict is produced by the shape of the arrangement rather than by anybody's conduct inside it. The consequence is very concrete. A structural conflict cannot be removed by replacing people, by writing a stricter code, by holding more meetings or by being more careful. If the whole investment team of Nilgiri Alternatives Advisors Private Limited, invented, is sacked on a Friday, on Monday morning the new team walks into exactly the same four situations.

The four conflicts each carry a name and a number. One, allocation, where a single opportunity fits two of the manager's own vehicles and only one of them can have it. Two, the continuation vehicleA new fund formed to buy an asset from an existing fund that the same manager runs., where a holding would move from a fund the manager runs into another fund the manager also runs, so the same party stands at both ends of one number. Three, fees taken at a portfolio company, where a business the fund has a stake in pays the manager directly. Four, the follow-on, where a manager decides whether to put more money into a holding that is struggling while also being the party who has to mark that holding.

Each of the four has a different answer, and the four answers are not interchangeable. That matters more than it sounds. A reader who has learnt that a fee reduction answers something reaches for it again at the next conflict and finds it does nothing there. A written policy is useless against a price. An independent price is useless against an allocation. The answer has to be matched to the shape of the particular conflict, and matching it means knowing the shape first.

Conflict 1: what happens when one opportunity fits two of the manager's own vehicles?

Nilgiri Alternatives Advisors Private Limited runs six vehicles. Two of them can plausibly want the same thing on the same day. Nilgiri Growth Partners Fund II, invented, is a closed-end growth and buyout fund with Rs 5,00,00,00,000 of commitments. Nilgiri Venture Fund I, invented, is a venture capital fund with Rs 1,50,00,00,000 of commitments. A fast growing company raising money to expand is not obviously the property of either mandate. The round could be a late venture round or an early growth investment, and honest people inside the same firm will disagree about which.

The overlap is not a rare event. Over its investment period Fund II's manager reviewed 412 opportunities, signed 31 confidentiality undertakings, issued 14 non-binding offers, went to exclusivity on 11 and completed 9. A funnel that wide, feeding vehicles with overlapping edges, produces the overlap case regularly rather than occasionally.

Name the conflict before naming anything else. One opportunity can go to only one vehicle. Whichever way it goes, one set of investors gets it and another set does not. And the party choosing is the same party that is paid by both sets of investors, on terms that are not identical between the two vehicles. There is no version of the decision in which the manager is not a beneficiary of the outcome, so the manager is not a neutral referee between its own two funds. Say that plainly and the reader has understood conflict 1. Nothing about it implies anybody has ever chosen badly.

The answer is a written allocation policyA written rule deciding which fund gets an opportunity that suits more than one of them. that the manager applies before the opportunity is seen. Not disclosed after, not decided case by case with good notes, not settled by a conversation. Written first, and then applied.

Why must an allocation policy exist before the opportunity is seen?

Because after the opportunity is seen, the manager knows something that makes rule writing impossible. The manager knows which of its two vehicles it would rather put the deal into, and it knows why. Maybe one fund is nearly fully invested and the other has capital sitting unused. Maybe one is running behind and a strong entry would help how it reads. Maybe the economics simply favour one side. None of that requires dishonesty. Knowledge on its own is enough. Any rule drafted with that knowledge in the room will end up accommodating it, and the person drafting it will feel entirely honest throughout.

Written first, the rule is drafted behind a wall. Nobody yet knows what will walk through the door, so nobody yet knows which fund the rule will favour. Drafting behind that wall is the whole trick, and it is the caterer's contract line again, and one child cutting the cake while the other chooses the slice. The date on an allocation policy is what makes it a rule rather than a preference, and not one word of its wording.

A date gives somebody something concrete to look at. A policy is a document with a date, a signature and a version. If a fund's investors ever want to know whether an allocation was made by rule or by preference, the answer is not in how the allocation reads. The answer is in whether the policy predates the opportunity. Everything else is commentary.

Conflict 2: what happens when the manager sets both the selling price and the buying price?

Fund II is at its record date, the end of its Year 9 Quarter 2, being 8.50 years after final close. Its contracted term runs ten years, so six quarters remain, and five of its nine holdings are still unsold. Holding 4, Bhavani Speciality Chemicals Private Limited, invented, cost Rs 60,00,00,000 and is carried at Rs 1,08,00,00,000, being 1.80 times its cost. Holding 4 has not been sold to anybody.

A manager in that position could form a continuation vehicle. A continuation vehicle is a new fund, raised from new money and from any existing investors who want to roll, and it buys holding 4 out of Fund II. Fund II has not done it.

Name the conflict. The selling party is Fund II, run by Nilgiri Alternatives Advisors Private Limited. The buying party is the new vehicle, run by Nilgiri Alternatives Advisors Private Limited. There is one number in the middle, and the same firm is standing at both ends of it. A low price is good for the buyer and bad for the seller. A high price is the reverse. Every rupee of movement in that number takes money from one set of the manager's investors and hands it to another set. Conflict 2 is the one conflict where a single party literally sets both sides of a single price, and no amount of process makes the transaction an arm's length one.

What answers conflict 2? Four things, and each of the four is a separate term. An independent valuation produced by somebody who is not the manager. A fairness opinionAn independent view on whether a price sits within a range that can be defended. on the price. A consent rightA defined matter the manager may not proceed with unless a named body agrees to it. held by the investor advisory committee, so the transaction cannot proceed on the manager's say-so alone. And a genuine cash option, so any investor who does not want to roll into the new vehicle can take the money instead and walk away. The cash option is the quiet load-bearing item. A choice that costs nothing to exercise turns a decision made about investors into a decision made by them.

Every one of those four is a term that a fund's own documents carry rather than a rule of law. How a continuation vehicle is actually built, what moves into it, how the valuation is run and what the cash option looks like in practice are covered separately.

Conflict 3: what are fees taken at a portfolio company, and who ends up paying them?

A private fund manager can be paid by the fund. A private fund manager can also be paid by the companies the fund has bought into, and that second route is the one fewer people expect. Two of those payments have names. A monitoring feeA fee the manager charges a company the fund holds, for continuing oversight work. is charged by the manager to a portfolio company for ongoing oversight: attending its board, reviewing its reporting, sitting in on its planning. A transaction feeA fee the manager charges for arranging a transaction at a company the fund holds. is charged for arranging something specific, such as an acquisition the company makes or a refinancing it puts in place.

Follow the rupee and the structure appears on its own. The payer is Sahyadri Diagnostics Private Limited, or Konark Polymers Private Limited, or any of the nine companies Fund II bought into. The receiver is Nilgiri Alternatives Advisors Private Limited. Fund II has a stake in the payer, so Fund II carries the economics of the payer. So a fee travelling from a portfolio company to the manager is, in substance, money moving out of something the investors carry and into the manager.

The conflict is that the manager sets the size of a payment it receives from a business whose value belongs to its own investors, and it is on both sides of that arrangement with nobody in between. Stated in that order, structurally, the conflict comes before any answer arrives. Nothing in that statement suggests the work is not done, that the fee is not earned, or that anybody has charged anything improperly. The point is only where the two sides of the payment sit.

The idea that a manager's economics split into fee income and carried interest, and that the two behave quite differently from each other, is Metrick and Yasuda's, in The Economics of Private Equity Funds, Review of Financial Studies, 2010. Fees taken at a portfolio company sit on the fee income side of that split, so the answer to conflict 3 is an arithmetic one.

What does a one hundred per cent offset actually do to the fee?

The answer to conflict 3 is arithmetic rather than procedural, and this is where the four answers stop looking alike. An offsetA reduction of the management fee by fees the manager received from somewhere else. reduces the management fee the fund pays by fees the manager received at a portfolio company. Set the offset at one hundred per cent and the manager keeps none of them: every rupee taken at a portfolio company reduces the fee by the same rupee, so having charged it makes no difference to what the manager ends up with in total.

Work it on Fund II's own numbers. The fund's management fee is 2.00 per cent a year, and the basis it is charged on steps down after the investment period. The step-down is covered separately and taken as settled here. To Fund II's record date at the end of its Year 9 Quarter 2, the management fee drawn totals Rs 70,20,00,000. Over the same period the manager charged Rs 1,20,00,000 of monitoring and transaction fees at Fund II's portfolio companies, and one hundred per cent of that is set against the management fee.

Two readings of that table, and both are true at once. The offset works exactly as designed. The manager's total is unchanged by having charged a single rupee at a portfolio company, and that is what a full offset is for. And the sum involved is small against the fee: Rs 1,20,00,000 is 1.7 per cent of Rs 70,20,00,000. The offset is not a discount on the fee, it is a rule about routing, and its size shows how much fee travelled by the other route rather than how generous anything is.

Now the boundary that matters most. Fund II offsets one hundred per cent. An offset of less than one hundred per cent is also a term that exists, and funds are contracted on many different terms. Exactly two facts hold: this invented fund's documents carry a full offset, and other offsets exist. Neither of those makes any offset usual, common, typical or expected. Anybody wanting to know what a particular fund's offset is reads that fund's own documents.

Conflict 4: why is a follow-on into a holding that is struggling a conflict?

Holding 6 of Fund II is Vaigai Edutech Private Limited, invented. Holding 6 cost Rs 30,00,00,000 and at the record date it is carried at Rs 21,00,00,000, being 0.70 times its cost. The holding is unrealised and it has been written down. Suppose that company comes back asking for more money.

Two entirely reasonable answers exist. One is that the business needs one more push to reach the point where a buyer will pay for it, and the extra money is the cheapest way to protect what has already gone in. The other is that the money is being sent after a result that is not coming. From the outside, at the moment of decision, those two look identical.

Now add the manager's second interest, and this is the bit a reader has to hold. The manager does not only decide whether to fund the company. The manager is also the party whose valuation process produces the carrying value. A holding that is refused more money and then fails has to be written down, and a write-downA reduction in the value a fund carries a holding at, recorded in its own accounts. is a loss that has to be explained to investors in the next report. A holding that is funded can, for another year, be presented as a business that is being backed. The manager has an interest in the mark as well as in the outcome, and a follow-on is one of the very few decisions that moves both at once.

The interest in the mark is a structural fact and not an accusation. Fund II's manager has never funded anything to avoid a mark, and holding 6 has not asked for money. The point is that the incentive sits in the arrangement whether or not anybody ever acts on it.

The answer is a size test. Above a size that Fund II's own documents fix, a follow-on cannot proceed on the manager's decision alone: the investor advisory committee has to consent to it. Below that size, the manager proceeds within its ordinary investment approval. The size threshold is a term of one invented fund's documents rather than anything a regulator sets.

Why does disclosure on its own answer nothing?

Disclosure is where most people stop, and stopping there is understandable. Disclosure feels like the answer. The investor was told. The investor was not kept in the dark. Surely that is the thing.

Allocation is where the gap shows most plainly, so run it there. Nilgiri Alternatives Advisors Private Limited runs Fund II and it runs Nilgiri Venture Fund I. The manager can disclose the overlap completely, in the fund's own offering document, in every quarterly letter, at every committee meeting, in a standing paragraph nobody could miss. And after all of that, the next opportunity that fits both vehicles is still allocated by the same party that benefits from where it lands. Disclosure changes who knows; it does not change who decides.

The same sentence does even more damage on a priced transaction. A continuation vehicle can be disclosed to the last comma, and the manager is still standing at both ends of the Rs 1,08,00,00,000. Telling everybody about a number the manager is setting twice does not stop the manager setting it twice. Only two things touch that: a price produced by somebody who is not the manager, and a real alternative for anybody who does not want to go along.

Who consents, and what can a consent actually do?

Two of the four answers turn on a consent, so a reader has to know which body gives it. Fund II has two committees and confusing them is the single most common mistake in this subject. The two committees do genuinely different things.

The investment committee has five members. Four of them come from the manager, being 80.0 per cent of the seats, and one is external, being 20.0 per cent. The investment committee approves every investment Fund II makes and every realisation it takes. The investment committee decides what the fund does.

The investor advisory committee has seven members, drawn from investors 1, 2, 3, 4, 5, 6 and 8 of Fund II's twelve investors, so seven of the twelve are represented by count and five of the twelve are not. It consents. Conflicts are among the matters the investor advisory committee consents on, and conflict 2 and conflict 4 both turn on that consent. The investor advisory committee does not approve investments and it cannot reject one, and anybody who thinks it can has mixed it up with the investment committee. Who sits on it, how those seats came to be allocated and what else it consents on are covered separately.

A consent has a limit, and overselling the consent would undo everything above. A consent adds a decision-maker who is not the manager. The addition is real and it is not nothing. A transaction the manager cannot complete alone is a genuinely different transaction from one it can. But the manager is still on both sides afterwards, the committee is made up of investors in the same fund rather than of independent outsiders, and it is acting on information the manager assembled. A consent narrows who can decide; it does not convert a conflicted transaction into one between unrelated parties, and no honest description of it claims otherwise.

How does somebody working near this actually read it?