The Ethical Dilemma: When Two Duties Conflict

A dilemma is two duties genuinely owed, pointing opposite ways, where satisfying one costs the other. A dilemma is not a hard decision, and it is not a rule nobody has looked up. A compliance breach has a right answer available and a dilemma does not. So a dilemma is resolved by choosing openly and recording the reasoning, rather than by finding the correct action.

There is an uncomfortable statistic that no one collects. Most of the situations people describe as ethical dilemmas are not dilemmas at all. Some are questions with a settled answer that nobody has gone and read. Some are decisions that are perfectly clear and simply unpleasant to carry out. Sorting those two out before anything else is not pedantry, and it is not a way of avoiding the hard part. The sorting matters because the procedure for a real dilemma is slow and expensive. Running that procedure on a question already answered wastes it. Treating a real dilemma as a lookup produces a confident answer to a question nobody had settled.

A single assignment supplies every example that follows. The Kavery research desk, inside an invented firm called 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. Twenty one documents sat in the working file: twelve public, six confidential, and three that arrived under a written undertaking.

Nobody in this assignment does anything wrong. There is no fraud, no cover-up and no villain. An account of ethics built on a scandal teaches the reader to watch for bad people. Bad people are the easy case. Everybody already knows what to do about them. The collision described here happened between two entirely reasonable duties on an ordinary working day, and neither of them belonged to anybody who was behaving badly.

Which duty should give way is not something any general ranking can settle. The ranking is precisely the judgement the situation asks the person standing in it to make and to write down, and an ordering supplied from outside would take away the only part of the work that is genuinely theirs. Below there is a test for recognising a dilemma, a comparison that keeps it apart from the breach it is constantly confused with, four steps for working through it, and a way of judging afterwards whether it was done well. All four are dischargeable by somebody in their first month, with no authority at all.

What Makes Something a Dilemma Rather Than a Difficult Decision?

An ethical dilemmaTwo duties genuinely owed, pointing opposite ways, so that satisfying one costs the other. has three conditions, and all three have to hold at once. There must be two duties rather than one. Each must be genuinely owed. Being owed means the party at the other end can be named. And the two must point opposite ways: the action satisfying one is the action that fails the other. With any single condition missing, the situation is something else, something cheaper and usually faster to deal with.

Most people have already lived a version of this outside work. A neighbour says, in confidence over tea, that the builder putting up the block down the road has not paid his suppliers for four months. A week later a colleague from the office mentions she is about to put a deposit on a flat in that exact block and asks what to make of it. Two things are now held at once. The neighbour is owed the confidence. A confidence is exactly that. The colleague is owed candour: she asked directly and she is about to move money. There is no action that does both. The situation has remarkably little to do with character: a completely honest person can still be standing in the middle of it.

The second condition is the one people skip, and it is worth slowing down on. A dutySomething owed to a specific person or party, rather than something merely preferred or believed. has a party at the other end of it. If nobody can be named as the one owed the thing, what is being held is probably a preference, and a preference dressed as a duty makes an ordinary decision feel like a moral one. Wanting the note to read well is not a duty. Wanting to avoid an awkward conversation is not a duty. Both are real feelings and neither is owed to anybody. Nothing that is owed to nobody can collide with anything.

Why Is an Unread Rule Not a Dilemma?

Here is the trap that catches almost everybody. The feeling of being in a dilemma arrives before anybody has checked whether the question is already settled, and once the feeling has arrived it becomes very hard to believe that something as ordinary as a document could settle it. The situation is heavy. Heavy things are not usually answered by a paragraph in a file. So nobody looks, and forty minutes go into weighing two duties against each other while the answer sits in a drawer.

The domestic version is the housing society argument about parking. Two households want the same space, both have reasonable claims, everybody has a view about fairness, and the discussion runs for an hour before somebody thinks to open the society bye-laws. The bye-laws turn out to say exactly how spaces are allotted. Nothing about the hour was dishonest. The hour was simply spent on a question that had already been answered by people who had thought about it in advance, calmly, without any specific pair of households in front of them.

The first step of the resolution procedure is therefore not a weighing at all; it is a lookup, and it exists precisely because the situation feeling hard is not evidence that no rule covers it. In finance work the places to look are short and specific. The rule, if there is one. The firm's policy. The undertakingA written promise about how material received from somebody else will be handled. the material came under. Such a promise very often says in plain language what may and may not be done with it. Any one of those turns the situation from a judgement into a reading task.

Why Is a Decision That Is Merely Unpleasant Not a Dilemma Either?

The second thing that gets mistaken for a dilemma is a decision that is completely clear and that nobody wants to make. An error turns up in a colleague's section two days before delivery. A client has to be told the answer they were hoping against. A word has to be held in place after somebody senior has asked pleasantly for it to be changed. Every one of those is difficult, and none of them has two duties in it. There is one duty, it is obvious, and what stands in the way is the discomfort of discharging it.

The distinction matters more than it sounds. The two situations call for opposite responses. An unpleasant decision asks for nothing but doing it, preferably today and preferably in writing so that the discomfort does not stretch out over a week. A dilemma asks for a procedure and a record, and it is entirely reasonable for it to take an afternoon. Treating an unpleasant decision as a dilemma is one of the most common ways professional courage gets quietly converted into paperwork. A procedure feels like progress and delay feels like diligence.

There is a test that takes four seconds. The test asks what the action would be if the discomfort vanished, if the colleague were a stranger, the client were someone else's, and the senior person were on leave. If an obvious answer appears immediately, there was never a dilemma. If two answers are still standing there with equal claim, there is one.

Ethical Dilemma vs Compliance Breach: Which One Has a Right Answer Available?

Now the comparison that does most of the work. A compliance breachA rule broken, in a situation where a right answer was available to be found. and an ethical dilemma feel almost identical from the inside. Both arrive with weight. Both prompt the urge to talk to somebody. Both leave the person holding them unsure whether they were handled properly. And they differ on exactly one thing: whether a right answer was available at the moment action was required.

A breach is a rule broken. Somewhere there was a correct action, written down by somebody, findable by anybody who went looking. If a different action was taken, the situation had an answer and the answer was not met, and that is not made softer by the difficulty or by nobody having noticed. A dilemma has no such answer. Two courses of action each have a real claim, and no document ranks them. Which one gives way is left to the person in the situation, for want of anybody else it could sensibly be left to.

A great deal falls out of that single difference. The place to look is different: a rulebook in one case, nowhere in the other. The act that resolves the situation is different: reading in one case, choosing in the other. The record does a different job in each: it shows the rule was found and applied in one case, and shows both duties were seen in the other. And the test of whether the situation was handled well is different in a way that surprises people. Two competent people facing the same breach must reach the same answer. Two competent people facing the same dilemma may reach different answers, and both of them may have done the work properly.

Why Does a Breach Have a Right Answer While a Dilemma Does Not?

Where does the right answer in a breach actually come from? The source of it explains the whole asymmetry. A rule is written in advance by people who are looking at a class of situations rather than at any particular one. The writers have the enormous advantage of not knowing who will be standing in it, and not knowing is exactly what lets them decide calmly. When a situation arrives and it fits the class, the thinking has already been done, and the work is to find it and apply it.

A dilemma is what is left when that advance work runs out. Two duties collide in a combination nobody anticipated, or in one somebody anticipated and deliberately declined to rank. Ranking them in advance would produce bad outcomes in the cases nobody could foresee. The absence of a right answer in a dilemma is often a design choice rather than an oversight. Waiting for somebody to publish the answer is therefore not a strategy.

The same design choice is why no general ranking of duties can be handed over. Any ordering that said confidentiality always beats completeness, or the reverse, would be right in most cases and badly wrong in some, and the wrongness would arrive silently in exactly the situations where somebody was relying on it. A procedure survives instead: it makes sure both duties were seen, that a way out was looked for, and that whatever was given up was said out loud rather than quietly dropped.

How Ethical Dilemmas Can Arise in Finance Work: What Are the Three Routes?

Dilemmas in ordinary finance work are not exotic, and they do not need a scandal to appear. Three routes account for most of them, and each one is simply two parties who are both owed something, where the two things cannot both be delivered.

The first route is completeness against a restriction on information. Something in hand would make the work better and arrived under a limit on what may be done with it. The person relying on the work is owed the most useful answer available; whoever supplied the material is owed the promise it came under. The Kavery assignment ran into this route, and it is the most common one on any desk that receives material from more than one source.

The second route is thoroughness against a deadline somebody else depends on. A committee sits at four o'clock. One figure is still untraced at half past three. Sending the work late fails the people who arranged their afternoon around it; sending it unverified fails the people who are about to act on it. Both groups are genuinely owed something, and the fact that this one shows up as a scheduling problem rather than as a moral question is precisely why it gets resolved badly so often.

The third route is one party's interest against another party's, where both are owed. Two teams inside the same firm, two clients with opposite positions, a colleague's reputation against a reader's need to know why a number moved. Nobody has done anything wrong. One person is simply standing between two claims that both carry the same name.

Step One: What Does Naming Both Duties Actually Require?

The procedure has four steps and the first one is the cheapest. Write each duty in a single sentence, and write who it is owed to. The second half is the whole of the step. Everybody who reaches this point can already name the two things pulling at them; almost nobody writes down the party at the far end of each, and the party is what turns a vague sense of pressure into a visible collision between two claims.

Said out loud, it runs like this. Duty one: give the person relying on this work the most complete and useful answer available. Owed to the reader of the note. Duty two: honour the promise the material arrived under. Owed to whoever supplied it. Something now becomes visible that was not visible before. The two parties are different people with no relationship to each other, and neither can release the analyst from the other's claim.

Two duties owed to the same person almost always turn out to be one duty and one preference. A single party rarely wants incompatible things and, when it does, it can simply be asked which. Sorting one from the other is the diagnostic value of the step. If both parties are written down and turn out to be the same person, the rest of the procedure has probably just been saved, and a two minute conversation takes its place.

Step Two: Which Option Gives Up Least of What Matters Most?

The second step is the one everybody jumps to, and it is a trade. The step lists the courses actually open, not the ones that would be preferable, and asks which of them gives up least of what matters most. The shape of that sentence matters. The step does not ask which duty is more important in general. No such question has an answer. The step asks which specific option, in this specific situation, damages least.

Severity is not a fixed property of a duty and it does not travel between situations. The same undertaking can be trivially inconvenient in one case and the entire basis of somebody's willingness to hand over material in another. The same completeness duty can be decorative when the reader has three other sources and decisive when they have none. Severity has to be assessed by the person standing in the situation. The assessment is the judgement being asked for, and a general ranking would replace it with something that quietly fails in the cases nobody foresaw.

Honesty about the options is what the step asks for. Two courses are usually obvious and a third is usually hiding, and the hidden third is what the next step is about. The step also asks for a look at partial versions: less of it, later, at lower resolution, in a form that carries the point without carrying the material. Sometimes there is a version that satisfies eighty per cent of one duty at a cost of five per cent of the other, and it is invisible until the options are written down.

Step Three: Can a Third Option Remove the Collision Entirely?

The third step is the one worth slowing down for, and it is the one that gets skipped. By this point two options are laid out and the trade between them feels like the work. A third optionA course that removes the collision altogether, rather than trading one duty against the other. is not a compromise between the two. A third option is a different action that makes the collision go away: both duties are met, or one of them stops applying.

The shapes a third option takes are worth memorising, and they recur. The party who imposed the restriction can be asked whether they will release it for this narrow purpose, and quite often they will: the restriction was written for a broader case. The same point can be found in material that carries no restriction, an hour of work that settles the matter completely. The claim the work makes can be changed, leaving the point the document was needed for no longer load-bearing. The decision can be moved to somebody who holds both duties differently, and moving it is what escalating a finance concern is for. Or the timing can be changed, so the deadline that was half the collision stops being one.

The household version is the neighbour and the builder again. The third option there is not deciding between the confidence and the candour. The third option is going back to the neighbour and asking whether the colleague may be told, without naming him, that there are payment problems worth checking. Most neighbours say yes. The collision that felt permanent lasted as long as it took to ask, and the reason nobody asks is that by then the situation has been framed as a choice between two things.

Step Four: What Does It Mean to Choose and Say What It Cost?

When the third step comes back empty, a choice has to be made. Choosing is the part nobody else can do, and it is also the part that takes the least time. If the first three steps were done properly the choice is usually clear by the time it arrives. The second half of the step is what takes the time and what people leave out: saying plainly what the choice gave up.

The cost has to appear in two places. The first is the work itself, where the cost appears as a stated limitA sentence in the work saying what the work does not cover, so a reader can adjust.. The person relying on the output then knows what they are not getting. The second place is the record, where somebody rereading the file later can see that the gap was a decision rather than an omission. A choice whose cost is stated is a resolved dilemma; a choice whose cost is quietly absorbed is indistinguishable, from the outside, from work that was simply not done well.

People resist this, and the resistance is understandable. Writing down what has not been established feels like advertising a weakness in a document that would rather look finished. In practice it does the opposite. A reader who has been through several hundred notes has learned that a document with no rough edges is either a very easy question or a very tidy writer, and cannot tell which from the outside. One line saying what is not established buys more credibility than any amount of polish.

What Did the Dilemma on This Assignment Actually Look Like?

Small, real and never dramatic. Nearly all of them are. Of the twenty one documents in the working file, three had arrived under a written undertaking. One of those three explained why the operating margin had fallen, clearly, in a single sentence. The margin itself was not in dispute: it stood at 10.8 per cent for the year just reported against 12.0 per cent the year before, and the internal floor used by this invented desk is 11 per cent. The restricted item supplied the cause, and the cause is exactly what a reader most wants and most often does not get.

Run the three conditions. Two duties, not one. Duty one, to the person relying on the note: give them the most complete and useful answer available. Duty two, to whoever supplied the document: honour the undertaking it arrived under. Both genuinely owed, to two different parties who have no relationship with each other. And pointing opposite ways: the single action of using that sentence satisfies the first duty and breaks the second. All three conditions hold.

Step one was done in writing and took about a minute. Step two produced two visible options and no obvious winner. Step three was checked properly, and it came back empty: the point could not be reached from the twelve public documents at anything like the same strength, and asking for a release was not available in the time the work had. So the choice was made. The conclusion was reached from public material. The public material supported it less crisply, and the note carried one added line saying that the cause of the fall is not established from the material available.

One honest detail that a tidier version of this story would leave out. The restricted sentence had already been read by the person writing the note, and it could not be unread. The sentence shaped nobody's confidence in writing, and it certainly sat somewhere in the head of the person drafting. Sitting there is not a failure of the procedure; it is a limit of it. The procedure delivers three things: the sentence did not reach the reader, the gap was declared, and a person rereading the file can see both.

What Goes in the Record, and Why Is It Most of the Answer?

The record is short. Nine lines will do it, and on this assignment it ran to exactly that. Line one names what was in front of the analyst, carrying the date. Duty one and to whom. Duty two and to whom. Then every option that was actually weighed, one line each, of which there were three here. The choice made. The cost of it. Signed, dated and filed. The artefact is no bigger than that, and writing it takes less time than the second step did.

Now the part that changes how people think about it. The record is not there to prove the choice was right. Its job is to show that both duties were seen before one of them gave way. Being seen can be demonstrated. The correctness of the choice cannot. A record that argues for the choice reads as defensive and, worse, is much less useful. The one question a later reader actually has is whether the collision was noticed at all. A choice made in full sight of both duties is a professional act. The same choice made because only one duty was ever in view is an accident that happened to land somewhere.

How such a record is written as a procedure, where it lives and who may read it are covered separately, under the documentation of ethical decisions. The record is not an administrative afterthought, and here is why. In ninety days, when the credit line comes back for review, nobody will remember the afternoon. The record is the only thing standing between a considered resolution and a gap in a note that looks, from a distance, exactly like carelessness.

How Can a Resolution Be Judged Afterwards?

Ninety days pass. The credit line that was cut comes back for review, and now everybody knows what happened next. The borrower recovered strongly, or the borrower did not. Somebody in the room says that the note should have said more, or that it said too much. The temptation, at this point, is enormous: to read the outcome backwards into the choice and conclude that a good ending means a good decision and a bad ending means a bad one.

The move has a name. ResultingJudging a decision by how it turned out rather than by what was known and weighed at the time. is Annie Duke's term for it, from Thinking in Bets, published in 2018, and the separation between decision quality and outcome quality is hers. The name is worth having: naming the move is most of what stops it. Judging a decision apart from its outcome is treated more fully under behaviour and decision quality, and what matters here is the narrower version: a dilemma resolved carefully can end somewhere nobody wanted, and a dilemma resolved by grabbing the first duty that came to mind can end perfectly well.

So the test is not the ending. Reread the record and ask four things. Were both duties actually named, with their parties? Was a third option genuinely looked for, or was the box ticked? Was the cost stated in the work, where the reader could see it? And, given only what was known that day, does the choice still look defensible? Four yeses is a well resolved dilemma whatever happened afterwards.

The cell to worry about is the one where the resolution was poor and the outcome was good. A poor resolution with a good ending teaches a desk the wrong lesson, and does it quietly. Nobody reviews a happy ending. The forty minutes that should never have been spent, the duty that was never named, the third option nobody looked for: all of it gets ratified by an outcome that had nothing to do with any of it, and next time everybody skips a little more.

That leaves the failure worth describing at length, and it is not the one anybody expects. Nobody in it behaves badly, nothing is concealed, and the whole thing is caused by taking a situation more seriously than it deserved.

Which Bodies Decide Any of This?

The mechanism above is universal. Two duties colliding is not an Indian phenomenon, and the four steps would read the same in any country. The jurisdictional part is narrower: whether a particular duty is set by a regulator or a professional institute over a specific role, and what happens if it is not met. Both questions have answers, set by named bodies, and the answers belong at the source.

The last point is more than a disclaimer, and it feeds straight back into the first step. If a body has set a duty over a role and ranked it against something else, there is no dilemma at all; there is a lookup, and the lookup happens to be at a regulator rather than in a drawer. Knowing that such rankings exist, without knowing their content, is enough to prompt the check.

How Does a Credit Committee, an Analyst or a Household Actually Use This?

Seen from the seat of somebody receiving the work, all of this is finally cashed. A credit committee member has perhaps fifteen minutes with a note, has never met the writer, and cannot audit anything in it. On 12 March, on this assignment, that committee moved a line from Rs 40,00,000 to Rs 25,00,000 and set it for review again in 90 days. The committee did that on the strength of a document that told them, in one line, that the cause of the margin fall was not established from the material available.

Watch what that line does for the reader rather than to the writer. The line tells them the size of the gap, so they can decide whether to wait, ask, or act anyway. The line tells them the gap is known rather than missed. A known gap and a missed one are the difference between an incomplete answer and an unreliable one. And if they ask, the record exists and shows them a collision that was seen and resolved rather than a corner that was cut. The stated limit is what converts a weaker note into a usable one. A reader can work around a gap they have been shown and cannot work around one they have not.

An analyst reading somebody else's research applies the same test in four seconds. Does this document admit anything? A note in which every claim points one way and nothing is qualified has usually been tidied, and the tidying is invisible. And the same reading works at home. When somebody hands over a valuation, an estimate of what a renovation will cost, or a view on a property, the useful question is not whether they are honest. The useful question is what the document says it does not know, and whether anybody could explain, six months from now, why a particular sentence is in it.

Why Is a Duty Not a Quantity?

A relationship between two quantities becomes obvious the moment one can be pushed and the other watched as it responds. A duty is not a quantity, though; it is a judgement. A slider that set the weight of confidentiality against completeness would turn two duties into a dial, and it would teach the exact misreading this whole subject exists to prevent: that a dilemma is solved by finding the right setting rather than by choosing openly and saying what the choice cost.

There is no correct weighting to discover, no score for a resolution and no percentage that certifies it was done well, and a control offering one would be teaching arithmetic in a place where the subject is judgement. In its place stands what stands in real work: a situation actually being faced, run through the three conditions, followed by the lookup and seven written lines.

Kavery Capital Services Private Limited, the Kavery research desk, 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.