Conflict of Interest vs Confidential Information
A conflict of interest is a pull on a person's judgement from something they stand to gain or protect. A confidentiality duty is a restriction on information handed over for a purpose. Both feel like caution, and they are discharged in opposite directions: a conflict by telling the right person, a confidence by telling nobody at all. Instinct says be careful, and careful is right for only one of them.
Two duties that pointed the same way could safely be handled by one habit. The two duties here point in opposite directions while feeling identical from the inside, so a single habit will discharge one of them perfectly and leave the other completely undone, and the person doing it will feel careful the whole time. No feeling makes the distinction on its own. The distinction has to be made explicitly, in words.
One assignment shows both duties arriving at once. The Kavery research desk, inside Kavery Capital Services Private Limited, was asked whether a borrower called Meenakshi Tubes Private Limited had weakened over the year just reported. Sharada Iyer produced the work, Prakash Nadar reviewed it, and Latha Menon commissioned it and took the decision at the end. The file opened on 3 March, the note went out on 11 March, and the decision followed on 12 March.
Nobody in that assignment behaves badly, and that matters here more than almost anywhere else. A cousin of Sharada Iyer is employed at Meenakshi Tubes. Personal connections of that kind are ordinary in a country where most people work within a few hours of where they grew up, and treating one as a warning sign teaches embarrassment rather than organisation. The connection is not the problem. The whole of the subject is what was done about it, and when.
Classifying a document when it lands, building a wall inside a firm, and deciding what may be typed into a tool are all set out under the control of confidential information. The two duties keep arriving alongside each other, and the boundary between them settles four things: which of the two is being held, what each one asks for, the order to run them in when a single situation produces both, and why the safe looking mistake is the one that cannot be undone.
What Is a Conflict of Interest, and What Is Actually Being Pulled?
A school sports day makes the shape plain. Somebody has been asked to judge the finish of the hundred metres and their niece is running in it. Nobody in the field thinks the judge would cheat, and the judge would not, and the organisers still put a second judge on the line and the first one still mentions it to them beforehand. None of that is an accusation. The second judge is a small adjustment, put there because everybody understands that a person watching their own niece is not looking at the race in quite the same way as a stranger, however hard they try.
A conflict of interestA pull on a person's judgement from something they stand to gain or to protect. is that situation written down so it can be recognised in an office. A conflict has three parts and needs all three at once. There is a judgement being made that affects somebody else. There is an interest held by the person making it, something gained or something protected. And a path runs between the two, so the interest could move the judgement. Take away any one of the three and there is nothing to declare.
Two things in that definition surprise people. The first is that the pullThe direction a conflict would move a judgement if nothing were done about it. does not have to have moved anything at all. A conflict is measured by what an ordinary outside reader would say could shift the view, not by what actually shifted. Insisting that no influence occurred is not a defence. Influence is precisely what nobody, including the person concerned, can check from the inside.
The second is that a conflict is a state rather than an act. Having a conflict is not itself an act, so having one is not a fault and the entire duty lives in what is done next. The judge at the finish line committed no offence by being related to a runner. There was one thing to do about it and she did it before the gun.
What Is a Confidentiality Duty, and What Exactly Is Restricted?
A second everyday scene feels remarkably similar even though almost nothing about it is the same. A friend asks for help filling in a loan application, and doing it means looking at her salary slip. She did not say keep this private. She did not need to. The slip changed hands so that one form could be completed, and the limit of that was understood without anybody stating it.
A confidentiality dutyA restriction on information that was handed over for a purpose, limiting what may be done with it afterwards. is that understanding made checkable. Something moved from one party to another for a stated or an obvious purpose, and passing it further on was not part of that purpose. The restriction is narrow and specific. The restriction does not stop the material being read, thought about, or used to answer the question it was sent to answer. The restriction stops exactly one act: the onward move.
The part that governs everything else is where the duty sits. The duty sits on the information and not on the person holding it. The transfer created the duty, and the duty stays attached to the material afterwards, so it survives a change of desk, a change of employer, and a completely honest failure to remember which document a sentence came from. The holder is not the keeper of a promise; the holder is the current location of something that already carried a condition when it arrived. Three of the twenty one documents on the Kavery file came in under a written undertaking. The undertaking puts that condition on paper and names who may open them, and the other eighteen documents carry the same logic without the paperwork.
A cousin works at the company being analysed. Is that a conflict, a confidence, or both?
Both duties are now on the table, each defined on its own terms and without reference to the other. Everything from here is the contrast, and it runs along four criteria: what each duty attaches to, how each one arises, how each one is discharged, and what the passage of time does to each.
Why Are These Two Constantly Confused?
Not through carelessness. The two duties are confused because the experience of meeting them is the same experience. Somebody mentions something, or a document lands, and a small alarm goes off that says there is something here I must not do. The alarm is genuinely useful, and it is what makes a decent person pause instead of forwarding. The alarm cannot say which of the two duties has just arrived. The alarm is identical either way.
The two are also confused because the vocabulary around them overlaps almost completely. Both live under words like care, propriety, discretion and integrity. Both get raised in the same induction session, usually on the same slide. And they really do arrive together: on this assignment the connection and the restricted documents were live in the same week, on the same file, in the same person's inbox.
Then instinct finishes the job. Pressed by an alarm that cannot be classified, the reflex is to shrink one's own footprint. Say less. Wait. Do not make it somebody else's problem. Do not raise something that might turn out to be nothing. The reflex to say less discharges a confidentiality duty perfectly and leaves a conflict of interest entirely undischarged, and the person following it feels equally responsible in both cases.
Why are these two duties confused so persistently, even by people who understand both perfectly well in the abstract?
Criterion One: Does the Duty Attach to a Person or to an Item of Information?
The other three criteria fall out of this one, so it is worth slowing down on. A conflict attaches to a person and, more precisely, to that person's situation: who they are related to, what they hold, who pays them, what else they do. A confidence attaches to an item of information and to nothing else. Neither of those is a metaphor. The two attachments decide, quite mechanically, what happens next.
A single event tests it. An analyst leaves one firm on a Friday and starts at another on the Monday. The conflicts are re-drawn over that weekend. Each was a property of a situation, and the situation changed. Some vanish: the old desk's fee arrangement stops paying, and the old firm's mandate for the other side of a matter no longer attaches. Some appear that were never there before. And a few walk across untouched. A cousin employed at a borrower is a cousin employed at a borrower wherever the analyst is sitting.
Now the confidences. Nothing at all happens to them over the same weekend. The three restricted documents are exactly as restricted on the Monday as they were on the Friday, the sentence somebody said on a plant visit is exactly as unrepeatable, and a figure whose source has been forgotten is exactly as constrained. Conflicts belong to the person and therefore change when the person's situation changes. The confidences held were never theirs in the first place, so they have nothing to change with.
There is a second, quicker test, useful for classifying something in the middle of a conversation: could somebody else have exactly the same one? Two colleagues can hold an identical confidence: the same document reached both of them under the same condition, and their duties are indistinguishable. Two colleagues cannot hold an identical conflict. A conflict is made of a particular person's circumstances and has to be declared by that person about themselves. Nobody can declare it on their behalf, and a large part of why undeclared conflicts stay undeclared is exactly that.
Criterion Two: Does It Arise From a Relationship or From a Transfer?
A conflict arises from a relationship or a position. A relative works somewhere. Shares are held. Pay rises with one outcome rather than another. A seat on a committee that will be affected by the outcome. Nothing has to be handed over for any of that to be true. A conflict can therefore exist before the assignment opens, and very often does. The cousin was employed at Meenakshi Tubes on 2 March, when nobody at Kavery had heard of the file.
A confidence arises from a transfer, and only from a transfer. Something has to move from one party to another for a purpose. No handover, no duty. A confidence therefore cannot pre-date the moment the material arrives, and a confidentiality duty always has a start time that could in principle be written down. The three restricted documents created their duty on the morning they landed and not a day earlier.
The difference in origin decides where to look. Conflicts are found by examining one's own situation before the work begins, and confidences are found by examining what came through the door while it was going on. Two completely different searches, run at two different times, and another reason a single habit will not cover both.
| Ordinary origins on a working file | Conflict of interest | Confidentiality duty |
|---|---|---|
| Something about a person | A relative employed by the business being assessed | Not applicable |
| Something held | Shares in a listed comparable, or in the subject itself | Not applicable |
| How the analyst is paid | A fee or a pool that rises with the size of the book | Not applicable |
| Something handed over | Not applicable | Management accounts sent so a lending request could be assessed |
| Something said to the analyst | Not applicable | A sentence offered on the walk back from a plant visit |
| Something written down | Not applicable | Three documents arriving under a written undertaking |
A document marked confidential arrives about a company whose shares the analyst holds. What is in play?
Criterion Three: Is It Discharged by Telling or by Not Telling?
To dischargeTo do whatever satisfies a duty, so that nothing further is owed on it. a duty is to do whatever satisfies it. Nothing further is then owed on it. On this criterion the two answers are not merely different. The answers are opposites, and one of them is the opposite of what caution suggests.
A conflict is discharged by disclosingTelling the right person, in writing, before the work rather than after it. it. Disclosure is a single act with a describable shape: in writing, to the person who will decide, before the work starts. Once it is done properly, the duty is discharged, and here is the part that surprises almost everybody the first time: the conflict is still there. The conflict did not have to be removed. Most conflicts in ordinary professional life are managed rather than avoided, and management usually means one small procedural change plus a written record that the connection was known about.
A confidence is discharged by withholdingNot passing something on, continuously and with no completion date, rather than as a one time action.. Withholding is not an act at all; it is a continuous state. There is no moment at which the withholding is finished, no version of it done thoroughly, and nothing done today reduces what is owed tomorrow. The state can only be maintained or broken.
Hold those two side by side and the whole confusion resolves. One duty is completed by adding something to the world, in five minutes, on a specific morning. The other is satisfied by never adding anything, for as long as the information exists. Caution reduces what gets done, so caution finishes the second duty and does not begin the first.
The instinct under both duties is to be careful and say less. On which of the two does that instinct point the wrong way?
Criterion Four: Does Either of Them Ever End?
A conflict has a deadline at the front and an ending at the back. The deadline is the point at which the work begins. A declaration is an input into somebody else's decision, and an input that arrives after the decision is not an input. Declared before any document is opened, it is worth everything. Declared halfway, it is worth something and costs a re-check. Declared after the note has gone out, it is worth very little and looks like a confession. Declared after somebody else found it, it is worth less than nothing. The question has changed from what was known into why nothing was said.
The ending is real too. A conflict can genuinely stop existing. The cousin takes a job elsewhere. The holding is sold. The mandate finishes. Firms therefore re-ask the question at intervals rather than collecting one answer forever.
A confidence has neither. There is no deadline at the front. Nothing can be late, and there is no early form of not passing something on. And there is usually no ending at the back. Closing a file disposes of documents; it does not dispose of sentences already read, and nobody has ever devised a procedure that does. The conflict is urgent and finite, the confidence is unhurried and indefinite, and treating both as one kind of thing produces either a person who panics about everything or a person who misses the only window that mattered.
An analyst changes firms on a Monday. What happens to the conflicts held and to the confidences held?
The contrast is now complete. Four criteria, and on every one of them the two duties answer differently. The four are worth having in one place before a single situation produces both.
| The criterion | Conflict of interest | Confidentiality duty |
|---|---|---|
| What it attaches to | A person, and that person's situation | An item of information, wherever it now sits |
| How it arises | From a relationship, a holding or a position | From a transfer made for a purpose |
| How it is discharged | By disclosing: one act, in writing, before the work | By withholding: a continuous state with no end |
| What time does to it | It has a deadline at the front and can genuinely end | It has no deadline and usually never ends |
| Can two people hold the same one | No. A conflict belongs to the person concerned, and only they can declare it | Yes. Everybody the material reached holds it identically |
| What being late looks like | Damaged, embarrassing, and still repairable | There is no late. There is only kept or broken |
What Happens When One Situation Creates Both at Once?
Constantly, is the answer, and it is worth saying plainly because the two duties are usually taught in separate sessions and that leaves people expecting them to arrive separately. They do not. A colleague forwards a memo marked confidential about a business whose shares the analyst holds, and one email has created two duties. A review is requested of work produced by a former deskmate, and the review pack contains supplied material. A relative mentions, at a wedding, that the place they work at is having a difficult quarter, and that business is due to be assessed next week.
The mistake in every one of those is not failing to notice. The mistake is noticing once. A person who classifies the situation as a confidentiality matter will withhold impeccably and never declare the holding. A person who classifies it as a conflict will declare it and, in the course of declaring it well, will say things that were not theirs to say. The situation has two duties in it, so it needs two separate answers, and the only reliable way to get both is to ask two questions instead of one.
Then the order. The conflict branch runs first, though not because it is more important. The conflict branch carries the only deadline of the two. The confidence gains nothing at all from early attention. There is no early, and nothing can be withheld in advance and banked as credit. The declaration, on the other hand, is worth its full value only in the window before the work starts, and that window closes on its own.
One situation has created both duties. Which is handled first, and on what grounds?
What Did Both Duties Look Like on This One Assignment?
The conflict comes first here, in the order it was actually handled. A cousin of Sharada Iyer is employed at Meenakshi Tubes, in a role with nothing to do with the accounts. The connection is not a shareholding, it is not a payment, and it was never a secret. A connection of that kind turns up on perhaps one assignment in five in a market this size, and treating it as remarkable would be both wrong and unhelpful.
On 4 March, the day after the file opened and before a single document had been read, it was declared in writing to Latha Menon in three sentences. Three sentences, not three printed sides. Length is the enemy of declarations. A person who believes a declaration has to be a document will put it off until they have time to write one, and there is never time before the work starts. The window before the work starts is the only moment that counts.
Two things changed because of it, both small. The review was made independent rather than collegial. In practice Prakash Nadar traced the figures back to their sources instead of reading the draft and commenting on it. And the connection was written into the basis field of the decision record. Anybody picking the file up in ninety days can see that it was known about rather than discovered.
One thing did not change. Sharada Iyer stayed on the work. Most conflicts in professional life are managed rather than avoided, and the difference between a declared connection and an identical undeclared one is not the size of the pull but whether anybody else is in a position to correct for it. A declared conflict is a governed conflict. The same connection, unmentioned, is a hole in the file.
Now the confidence, running in parallel the whole time. Three of the twenty one documents arrived under a written undertaking. One of them explained, clearly and in one sentence, why the operating margin had fallen to 10.8 per cent against an internal floor of 11 per cent. Using it would have made the note better and would have broken the undertaking. The explanation was withheld. The conclusion was reached from public material, less crisply supported, and the note said so in terms: the cause of the fall is not established from the material available.
Same assignment, same person, same week. One duty discharged by writing something down and sending it to somebody. The other discharged by leaving the best sentence in the file unused and saying nothing about why. The pair is the image worth keeping.
A third interest on the file is worth naming, of the kind people wrongly get excited about. The relationship manager who holds the borrower connection is paid in a way that rises with the size of the book they hold. The pay arrangement is an interest. The arrangement is disclosed, it is ordinary, and it is not a wrongdoing; almost everybody in commercial life is paid in a way that rewards some outcomes over others. Visibility to the people reading the work is what makes it manageable.
Sharada Iyer declared the connection and then stayed on the work. Was the conflict removed?
The disclosure that breaks a confidence
One failure combines the two duties, and it happens to people who are trying hard. Somebody sits down to declare a conflict properly. Knowing that a thin declaration is worthless, they explain: which client it was, on what matter, and what they came to know. The conflict is now disclosed thoroughly and correctly. A confidentiality duty has just been broken in the same breath, to the same person, in the act of discharging the other duty well.
Notice what makes it so hard to catch. Every instinct in the room says more detail is better. More detail is what turns a declaration from a formality into something the recipient can act on. The instinct is right about the conflict. The instinct is simply borrowing from an account it does not hold. The detail being added belongs to somebody who is not present and never agreed to any of this.
The fix is precise rather than philosophical: disclose the existence and the shape of the conflict, and never the confidential content behind it. The shape of a conflictThat one exists and roughly what kind it is, without any of the content behind it. means that there is one, roughly what type it is, and what ought to change. A sentence of the form I have a connection to a party in this matter and cannot say more, said to the right person before the work, discharges the conflict completely without spending anything.
The second sentence is the one worth rehearsing, not the first. The person receiving the disclosure will very often ask a follow up question, in complete good faith, trying to work out how serious it is. The follow up is where confidences actually go. Having a flat, unembarrassed answer ready matters more than having a well phrased opening.
Rewrite this so it discloses without breaching: I should mention I worked on the other side of this deal for the buyer last year.
Which is worse to get wrong: a conflict never declared, or a confidence broken?
Which One Is Worse to Get Wrong, and Why Is It Not the Obvious Answer?
The obvious answer is the conflict. Not declaring something sounds like concealment, and concealment sounds worse than an accident. A broken confidence, by contrast, usually happens through momentum: a sentence goes one step further than it should have, in a room where everybody meant well.
The obvious answer is wrong, and the test that settles it is repairability. An undeclared conflict can still be declared. Late is genuinely bad: the work may have to be re-done, the reviewer's confidence takes a knock, and in a serious case somebody else has to take the decision again from the start. Every one of those is a repair. Something can be done, this afternoon, that improves the position. And the people harmed are almost always in the room, so they can be told and can adjust.
A broken confidence has no repair available at any price. There is no version of the next hour in which the person who now knows stops knowing, no apology that unmoves the information, and no procedure that recovers it. Worse, the party harmed is the supplier, who is not in the room, often never learns of it, and therefore cannot object, negotiate or forgive. The duty that has to hold without any supervision at all is the one whose failure nobody will ever report back.
None of that is a licence to be relaxed about conflicts. Repairability is an argument about which mistake to be most careful around when both are in play and attention is short, and it works only where the judgement is made on what was known at the time rather than on how things turned out. Annie Duke makes that separation the centre of Thinking in Bets, published in 2018: a good decision can end badly and a poor one can end well, so the thing worth grading is the process run on the information available.
Which Indian bodies set duties of this kind?
Three bodies are worth knowing by name. The Securities and Exchange Board of India, at sebi.gov.in, sets conduct duties for registered intermediaries and the people working inside them, and a regime exists in India covering the declaration of interests and the keeping of records about them. The Institute of Chartered Accountants of India, at icai.org, sets professional conduct standards that attach to a member personally rather than to whoever employs them at the time, and both objectivity and confidentiality sit among those standards. The International Organization of Securities Commissions, at iosco.org, publishes conduct principles that several national regimes draw on. Who a declaration must go to, in what form, within what period, and what record has to survive are all set by the regulator.
How Does a Lender, an Analyst or a Household Actually Use This Distinction?
A credit committee uses it as a reading test on work it has no way of auditing. Two things are checkable from the outside. Did the note carry its interests at the front, before the analysis rather than after it? And does the note stop where its material stops, or does it explain more than its stated sources could support? A committee that sees a declared connection in the basis field trusts the note more rather than less. A writer who volunteered an awkward fact is probably volunteering the others. The reverse test is subtler and worth knowing: a note that explains a great deal about supplied material is telling the committee that its author cannot tell the two duties apart.
An analyst uses it as two separate housekeeping habits, run at two different times. Once, at the start of every assignment, a short pass over the analyst's own situation: holdings, relatives, previous work, anything that pays differently depending on the answer. Then, continuously, a running sense of what came through the door and under what condition. The first list is about the analyst and gets written down and sent. The second list is about the file and never leaves it.
An investor reading published research uses it as a reason to read the front matter first. Whatever the author stands to gain is disclosed material, put there so the reader can adjust; what the author will not say is usually a confidence rather than an evasion, and the two look identical to anybody who has never separated them.
And a household uses it at a wedding, the place where most of this actually gets practised. Suppose somebody is helping a relative choose between two caterers. One of them is run by their neighbour, and they say so at the start, before anybody looks at a price. The whole duty is that one sentence. Separately, the other caterer showed them a supplier price list to justify a higher quote, and that list does not travel to the neighbour, ever, no matter how the conversation goes. Same evening, same two people, and the two duties are discharged in opposite directions: the first thing is said out loud immediately and the second thing is never said at all.
References
| Source | Document | Where |
|---|---|---|
| Securities and Exchange Board of India | Conduct duties applying to registered intermediaries and the people working within them, together with the regime covering the declaration of interests | sebi.gov.in |
| Institute of Chartered Accountants of India | Professional conduct standards binding on members, including an objectivity duty and a confidentiality duty that attach to a person rather than to an employer | icai.org |
| International Organization of Securities Commissions | Published conduct principles that several national regimes draw on, covering interests and client information held by market intermediaries across borders | iosco.org |
| Annie Duke | Thinking in Bets, 2018, on separating the quality of a decision from the quality of its outcome, the basis on which a late declaration and a broken confidence are compared above | Portfolio, Penguin Random House |
The Kavery research desk, Kavery Capital Services Private Limited, Meenakshi Tubes Private Limited, Sharada Iyer, Prakash Nadar and Latha Menon are invented.
Educational material. Not advice on any investment, tax, budget or market position.
