Conflict of Interest and Related-Party Transactions Compared

A conflict of interest is a situation. Somebody taking or influencing a decision holds an interest of their own that could pull against the interest they are there to serve. A related-party transaction is a defined category with a process attached, and whether an arrangement falls inside it is settled in law rather than by judgement. The two overlap often, and neither one implies the other.

Conflict of interest and related-party transaction turn up in the same meetings, in the same reviews, and often in the same sentence, and the two phrases are quietly treated as two ways of saying one thing. The two phrases do not mean one thing. One of them describes a position that a person is standing in. The other describes a box that an arrangement either falls inside or does not. Confusing them produces a review that is thorough about the wrong question, and thoroughness aimed at the wrong question is a particular kind of expensive.

Start on a street rather than in a transaction. A residents' committee is choosing who repaints the building, and three firms have quoted. One of the people on the committee has a cousin who runs the second firm. Nothing has happened yet. No decision has been taken, no favour has been done, and the cousin's quote may well be the cheapest and the best. But the person on the committee is now in a position where their own interest and the interest of the residents could pull in different directions, and everybody in the room can feel that even before anybody says it out loud.

Now change the facts. The committee decides to buy paint from a shop that the building's own housing society already runs. Nobody on the committee gains a rupee. The price is the ordinary price. But the shop and the society are the same set of people wearing two hats, and that fact alone puts the purchase into a category that gets handled differently, whatever the price turns out to be.

The painting quote and the paint purchase are the whole of this guide. The first is a conflict of interest. The second is what a related-party transaction looks like when it is stripped of the finance vocabulary. The first is found by asking who stands to gain, and the second by asking who the counterparty is, and those two questions have nothing in common except that they tend to get asked in the same room.

Harivansh Packaging Limited, an invented listed maker of rigid and flexible packaging, has bought all of Sundarban Polymers Private Limited, an unlisted maker of flexible packaging films. Devyani Kulkarni is the chief financial officer of Harivansh Packaging Limited and Ashwin Rege leads its transaction team. Every rupee figure below is worked for teaching.

What is a conflict of interest, and when is a person standing in one?

A conflict of interest is a situation in which a person taking a decision, or influencing one, holds an interest of their own that could pull against the interest they are there to serve. The sentence above is the whole of the definition, and every word in it is load bearing. There has to be a decision. There has to be a person with some hold on it. And there has to be an interest of that person's own that points somewhere other than where their duty points.

Notice the requirements the definition leaves out. The definition does not require that the person acted on the interest. The definition does not require that the decision came out badly, or even that it came out differently from how it would have come out otherwise. Dishonesty, carelessness and bad faith are not required either. A conflict of interest describes a position rather than a behaviour, and a person can be squarely inside one while doing nothing at all that anybody could criticise.

The descriptive character of the phrase is the part that most treatments blur, and the blurring is not harmless. If the phrase is heard as an accusation, the natural response to being told one is in it is denial. What has been heard is that somebody thinks the person has behaved badly. If it is heard as a description of where that person is standing, the natural response is to say so and let somebody else take the decision. The second response is the useful one, and it is only available to people who have understood that the words are descriptive.

The word that carries the most weight in the definition is could. The interest has to be capable of pulling against the duty. The interest does not have to be shown to have pulled. A bar set that low is deliberate, and it is low for a practical reason: whether an interest actually swayed a decision is almost never knowable, even to the person who took it. People are not reliable narrators of their own reasoning. The structure of a situation can be seen from outside and a mind cannot, so the test is set on the structure instead of on anybody's account of their own reasoning.

Three ordinary shapes it takes

The first shape is the one everybody recognises. The person deciding gains directly if the decision goes one way. The committee member's cousin has quoted for the painting. Ashwin Rege, who leads the transaction team at Harivansh Packaging Limited, is paid partly on transactions reaching completionThe day the purchase actually takes effect, when the money moves and the shares change hands. Signing and completion are separate events and can be weeks apart.. Nothing dishonest is implied by either sentence; both are simply statements about where the incentives sit.

The second shape is quieter, and it is about standing rather than money. A person who has argued in three meetings that a purchase is a good idea now has to review whether the diligence findings are serious enough to stop it. Their interest is in not being wrong in public. No rupee turns on it. A wish not to be wrong in public is still a conflict, and it is the one that gets missed most often. There is no payment to point at.

The third shape is the one this guide is built around, and it appears after the transaction rather than during it. Somebody has an interest in a number that will be measured later. The people holding that interest cannot decide what the number will be, but they can influence the things the number is made of. An entire section below works that case.

A conflict asks of everybody that it be named. DisclosureTelling the people who are relying on a person's judgement what interest of that person's own touches the decision, before that decision is taken rather than afterwards. is the first move, and only the conflicted person can reliably make it. The conflicted person is usually the only one who knows the whole of their own position. Disclosure by the conflicted person separates a conflict from a related-party transaction, and the difference is worth carrying forward.

What is a related-party transaction, and who decides that it is one?

A related-party transaction is a transaction whose counterparty falls inside a defined class of persons. That is all. A related-party transaction is a classification, and classifications all work one way: the counterparty is held up against the definition, and either the counterparty is inside the class or outside it. There is no scale, no shading and no middle. The classification is a category rather than a judgement, and the answer does not move with anybody's view of whether the arrangement is a good one.

Say that a related partyA person or a business that sits inside the class the law defines for this purpose, usually because of control, holding or position rather than because of anything they have done. is a person or a business sitting inside that defined classThe list, written in law, of who counts for this purpose. A list rather than a principle, and a list can be checked rather than argued about.. Who is inside that class, and on what basis, is written in company law and, where a listed company is involved, in what the Securities and Exchange Board of India (SEBI) publishes. The definition is set in law and is best read at the source.

The everyday version is a concession category on a railway ticket. Either a passenger falls inside the category the rules describe or does not. The clerk at the window is not weighing whether the passenger deserves the concession, whether the passenger seems like a decent person, or whether the fare is fair. The clerk is matching the passenger against a list. A category built like that has a property worth naming: it can be checked by somebody who knows nothing whatever about the merits.

One question trips people up. Does it help that the terms are entirely ordinary? For the classification, no. An arrangement struck at exactly the price anybody else would pay, on exactly the conditions anybody else would get, is still inside the category if the counterparty is inside the class. Terms may well matter enormously to what the attached process then asks for, and to what anybody reading about it later concludes. Terms do not touch the classification itself, and they were never meant to.

There is a reason the category was built on identity rather than on terms, and it is a good one. A rule built on terms is a rule about judgement, and judgement is exactly what a rule of this kind is trying not to depend on. Anybody can dispute whether a price was reasonable. Nobody can dispute who signed on the other side. Building the trigger on the fact that can be checked, and leaving the arguable part to whatever process follows, is what makes the category usable at all.

One more property follows from that, and it matters later. Because the trigger is identity, it fires by itself. Nobody has to notice anything, volunteer anything or feel uncomfortable about anything. The check runs against a list, the answer comes back, and if the answer is yes then the process attaches whether or not a single person in the room has an interest of their own anywhere in sight.

Why is one a situation and the other a category?

Put the two tests next to each other and the difference stops being a matter of vocabulary. To find a conflict of interest, the question is who stands to gain from the outcome. To find a related-party transaction, the question is who is on the other side of the arrangement. One question is about incentives and the other is about identity. The two questions share nothing but the room they are asked in. Answering one of them well says nothing at all about the other.

Look at who is capable of answering each one. The identity question can be answered by somebody with the share register, the list of directors and the definition in front of them. Such a person needs no view about the transaction and no knowledge of anybody in it. The incentive question cannot be answered that way at all. A great deal of what people stand to gain is simply not written anywhere. The gain sits in employment arrangements, in reputations, in what somebody said in a meeting last month and would prefer not to be wrong about.

The two also differ in when each can be answered. Identity is knowable before anything is signed, and it does not change because the transaction went ahead. Incentives move. A person who had no interest in the outcome during the negotiation can acquire a sharp one at completion, simply because of what the agreement turned out to say. Acquiring an interest at completion is not a hypothetical. A conditional payment does precisely that, and the second half of this guide works exactly that case.

And look at what each one produces. A yes on the identity question produces a process, set out in law, that starts running without anybody deciding it should. A yes on the incentive question produces nothing at all by itself. A yes on the incentive question produces something only if somebody says it out loud, and then only if the people who hear it do something about it. One of these tests has an engine attached and the other has a person attached, and people can be busy, junior, new, or simply unaware that what they are standing in has a name.

The practical consequence is the reason this comparison is worth setting out. Neither test substitutes for the other. A clean identity check establishes that no process attaches. A clean identity check does not establish that nobody stands to gain. A room full of conflicts, all properly disclosed and properly handled, establishes nothing about whether the counterparty was inside the defined class. Each question is blind to what the other one is looking for, so both have to be asked, every time, and asked separately.

Can a conflict exist when nobody is a related party at all?

A conflict can and does exist with no related party anywhere in sight, and by a wide margin that is the commoner of the two situations. The commoner situation is also the harder one, for a reason that has nothing to do with how difficult the facts are. No defined category catches it, so nothing whatever happens automatically, and the entire response depends on a person choosing to raise something about themselves.

Work it on the invented purchase. Harivansh Packaging Limited has bought Sundarban Polymers Private Limited. The people who sold Sundarban Polymers are strangers to Harivansh Packaging: they held no shares in it, sat on no board of it, and had no connection to it beyond selling to some of the same customers. Whatever the defined class contains, they are outside it, and a check run against the register and the definition comes back clean in an afternoon.

And yet. Ashwin Rege, who led the transaction team, has spent fourteen months on this purchase and is paid partly on transactions completing. The people who sold the business are staying on to run it, with Rs 60 crore of conditional money turning on how one year goes. Devyani Kulkarni, the chief financial officer, has told the board twice that the price is defensible. Every one of those is an interest of somebody's own that could pull against the interest they are there to serve. Not one of them is touched by a related-party check, and none of the people involved is a related party of anybody.

The street version is the food stall outside a single office building. The stall holder has no relationship with the building, no share in it and no seat on anything. But the stall lives or dies on whether that one office stays open, so when the building's committee asks the stall holder whether the road repairs should be done in June or in December, the answer that comes back is not a disinterested one. Nothing about the two parties makes them related. Everything about the situation makes the answer interested.

Can a related-party transaction carry no conflict at all?

Equally yes, and running the comparison in this direction is what most treatments leave out. An arrangement can fall squarely inside the defined class and be completely ordinary in its terms, its size, its timing and its purpose, with nobody involved standing to gain a rupee they would not otherwise have had.

Take the position after completion. Sundarban Polymers Private Limited now sits inside the group headed by Harivansh Packaging Limited, and the two businesses start doing the ordinary things that businesses inside one group do. Harivansh Packaging runs a shared testing laboratory and lets Sundarban Polymers use it, charging a group chargeAn amount one business inside a group charges another for something it provides, such as shared services, use of a facility or head office costs. for the time used, at the rate its own units are charged. Nobody in either business gains personally. The rate is the rate. Asked who stands to gain from the charge being higher or lower, the honest answer is nobody in particular.

The category attaches anyway, and it attaches for the reason set out earlier: it runs on identity, and identity is not in doubt. The category was never a verdict on the terms in the first place, so a transaction can be entirely unremarkable in its terms and still sit inside it. Whatever process the law attaches to it applies, and the fact that nobody would raise an eyebrow at the rate does not switch that process off.

People find this frustrating the first time they meet it, and the frustration is worth answering. If the terms are ordinary, why put the arrangement through anything? Because the process is not a punishment and the category is not an accusation. The process is a way of making a particular set of arrangements visible, on the view that arrangements between people who are connected are worth seeing, whatever their terms turn out to be. Whether a specific one is arm's lengthOn the terms two unconnected parties negotiating for themselves would have reached. The phrase describes the terms, not who the parties are. in its terms is a separate question from whether it sits inside the class, and the two are answered separately.

How does an earn-out create an interest in a measured number?

Everything so far has been definitional. The earn-out is where the two ideas meet the arithmetic of a transaction, and it is the reason the comparison belongs among transaction figures rather than in a chapter on governance.

The figures are already settled from earlier in this sequence. The earn-outA part of the price that is paid only if the business hits an agreed measure after completion. An earn-out is a conditional payment, not a deferred certainty. pays a further Rs 60 crore if Sundarban Polymers Private Limited reaches earnings before interest, tax, depreciation and amortisation (EBITDA) of Rs 145 crore in the first year after completion. The business earned Rs 132 crore in the year before, so the threshold sits 9.8 per cent above where it started. Rs 1,137 crore has already been paid for the shares, after the completion adjustments settled earlier in this sequence. Rs 60 crore against that Rs 1,137 crore is 5.28 per cent of what the sellers have already received.

Now add one ordinary fact. The people who sold the business stay on to run it. A buyer who wants the business to keep working the way it did usually asks for exactly that. From the moment they stay, they hold an interest in a measured numberA figure struck later, out of a business's own records, so what goes into it can be affected by decisions taken in the meantime. that will not be struck for another year and will then be computed out of the business they are running.

On the other side, Harivansh Packaging Limited now holds Sundarban Polymers and runs it as part of a larger group. Running it inside a group means doing the things a group does. Moving a cost centre. Changing where the films are bought from, now that the group buys more and buys cheaper. Charging the business for a shared laboratory, a shared warehouse, a shared finance team that it never paid for when it stood alone. Every single one of those is an ordinary management decision that a buyer is entitled to take, and every single one of them moves the number the Rs 60 crore is measured on.

Two interests are now pulling on one figure from opposite directions, and both of them are completely legitimate. The answer has to be definitional rather than moral. Nobody is behaving badly. The sellers are not scheming when they want the number measured high; that is the deal they struck. The buyer is not scheming when it charges the business for the laboratory it now uses; that is what running a group means. The conflict is in the structure, and it would be there if every person involved were a saint.

Why the step makes the incentive sharp rather than mild

If the earn-out paid a little more for every extra crore of EBITDA, this would be a mild incentive and a mild problem. It does not. The earn-out is a step. Below Rs 145 crore the payment is nothing at all. At Rs 145 crore it is the whole Rs 60 crore. There is no partial payment on the way up, and nothing between the two states.

Follow one crore of cost around that step and the sharpness becomes obvious. Suppose Harivansh Packaging Limited decides to charge Sundarban Polymers Rs 1 crore a year for the shared laboratory. If the business is heading for Rs 132 crore of EBITDA, that decision costs the sellers nothing. The earn-out was never going to pay at Rs 131 crore any more than at Rs 132 crore. If the business is heading for Rs 145 crore, the same decision costs them the entire Rs 60 crore. The same Rs 1 crore of cost is worth nothing in one place on the scale and worth sixty times itself in another, and that is what turns a mild interest into a sharp one.

The sharpness is also nobody's fault. The step was not designed to create the effect. A threshold is easy to write and easy to test, and that is why the shape was chosen. The effect arrives with the shape whether anybody intended it or not. A payment structure creates incentives by its geometry, and the geometry gets chosen for reasons that often have nothing to do with the incentives it produces.

What can an agreement do about a number measured a year later?

The response to a conflict about a measured number is not a moral one, and it is not a promise from anybody to behave well. The response is definitional. If the trouble is that two sides read the same figure differently, the answer is to write down, in advance, what the figure means.

Three things get defined, and they are always the same three. First, how the EBITDA is computed: on what basis, from which records, with which items in and which out. Second, what may and may not be charged to the business after completion. The charging rule catches the shared laboratory, the group finance team and everything else the buyer might reasonably want to allocate. Third, who resolves a disagreement if one arrives anyway, and on what timetable. A dispute then has somewhere to go other than a court.

All three are drafted before completion for a number that nobody can see until a year afterwards. Allocating a transaction risk in advance always has that shape, and that is why the pattern keeps recurring. The work has to be done at the point when everybody is agreeable and nothing is at stake, because the moment the number matters the two sides have Rs 60 crore of reasons to read it differently.

Now close the loop on the two ideas, on this case. The two stay separate all the way through. If the sellers turn out to sit inside the defined class, then whatever process the law attaches applies as well, in addition to everything above, and it applies because of who they are and not because of the Rs 60 crore. If they sit outside it, nothing at all attaches, the conflict is completely unchanged, and the only thing standing between the transaction and a fight in thirteen months is whether somebody wrote the definitions down. Both questions are routed to their own answer and each is answered separately.

What does each one ask of the people involved?

The practical difference lands here, and it is the difference worth remembering if everything else in this guide fades.

A conflict asks first for disclosure. The person in the position says what their interest is, to the people relying on their judgement, before the decision rather than after it. Then somebody other than that person decides what to do about it. Sometimes the answer is that the interest is too small to matter and the person carries on. Sometimes it is recusalStepping back from a particular decision, so that it is taken by people who have no interest of their own in how it comes out.. The person steps back and the decision is taken without them. Either way the choice belongs to the people who are not conflicted. Handing that choice to them is the part that makes it work at all. And the whole sequence rests on the first move. Only the conflicted person can reliably make it.

A related-party transaction asks for whatever the applicable process asks for. The contents of that process, when it applies and who has to do what are set out in law. The point for the comparison is the trigger rather than the contents: the process attaches because the counterparty is inside the class, and it attaches whether or not anybody in the room feels anything is amiss.

One of these mechanisms depends on a person choosing to speak, and the other one does not. The unglamorous discipline of asking the incentive question out loud is therefore worth more than any amount of enthusiasm about the category question. The category question gets asked because a form has a box for it. The incentive question gets asked because somebody decided it should be on the list, and if they did not, nothing anywhere in the process will notice its absence.

Where are these definitions actually written down?

The published text on all of this is short enough to place in a paragraph. Who counts as a related party, what has to be approved and by whom, and what has to be disclosed are set by company law, published by the Ministry of Corporate Affairs at mca.gov.in. Where a listed company is involved, SEBI publishes what applies to it at sebi.gov.in. A reader who needs the current position reads it at those two places.

The reason is not caution for its own sake. A definition of this kind is live text that gets amended, and a summary of it written on a Tuesday is a snapshot of a Tuesday. A reader who takes the snapshot for the thing has a wrong answer with no way of knowing it has gone wrong. A wrong answer is worse than no answer at all, and a person with no answer goes and looks. Naming the source and stopping is the only version of this that stays correct.

Neither the definition, nor the process, nor the approvals, nor the disclosures can be settled anywhere but at those two addresses. The durable part is the distinction: which question each idea answers, why they do not substitute for one another, and how they behave when they meet a conditional payment. A distinction survives an amendment to a definition, and that is what makes it worth teaching.

How do the people around a transaction use the two questions?

Watch how somebody doing the job actually handles it. The theory turns into two lists and very little else.

Devyani Kulkarni, as chief financial officer of Harivansh Packaging Limited, keeps them apart deliberately. The identity list is short, checkable and finished: every counterparty in the transaction held against the definition, with the answer written down and the working kept. The identity list can be built by anybody with the register and the definition, it does not need her, and it is genuinely closed once it is done. The incentive list is the opposite of all of that. The incentive list names people and what each of them stands to gain from which outcome, it can only be built by asking, and it is never finished. A new incentive can arrive on the day the agreement is signed.

A lender looking at the purchase reads the same facts from a different angle and reaches a short answer. A conditional Rs 60 crore is a possible cash outflow with an uncertain date, and an outflow that is likely to be argued about is worse than one that is not. The argument itself has a cost and the timing becomes unpredictable. So a lender cares less about whether anybody is inside the defined class and more about whether the agreement says exactly how the EBITDA is computed. A well defined earn-out and a vague one are different obligations, whatever the amount printed on them.

An analyst reading the disclosures afterwards has the hardest job of the three, and the reason is worth stating plainly. Related-party arrangements get disclosed. A category with a process attached produces a paper trail. Conflicts of interest mostly do not. A conflict produces a paper trail only when somebody discloses it and the disclosure gets recorded somewhere a reader can reach. An analyst who reads a clean set of related-party notes and concludes that nobody had an interest in anything has drawn a conclusion the disclosure was never capable of supporting.

The household version of all of this is a person who sells their shop and stays on as the manager, with a bonus payable if next year's takings cross a line. The buyer and the seller are strangers, so there is no category anywhere in it. And still the manager has a reason to push a repair into the following year, and the buyer has a reason to load the shop with a share of the warehouse rent, and the two of them are going to spend next April arguing unless somebody wrote down in advance what counts as takings. Same structure, four zeroes fewer.

Which questions are settled somewhere else?