Communication Conduct: Fair Presentation, and What Crosses the Line

Fair presentation means a reader who takes the statement at face value is not left with a false impression. A statement can be entirely accurate and still mislead by what it omits, by what it makes prominent, or by the frame it uses, so fair presentation is a stronger test than truthfulness. All three routes are available without a single false word.

What is Fair Presentation, and fair to whom?

Start with something that happens in every street in the country. A vegetable seller shouts that the tomatoes are from the hills. They are. He does not shout that they were picked nine days ago, and nobody asks. The sentence he chose sounded like the whole answer to the question a buyer was already asking. Nothing he said was false. The buyer went home with a false impression anyway, and the seller did not have to lie once to produce it.

Fair presentationPresenting so that a reader taking it at face value is not left with a false impression. is the duty that closes that gap. The duty asks a question about the reader rather than about the sentence. Truthfulness is a property of the words. The impression is a property of the reader, and only the impression is what a communication is for. A communication that is checked line by line and cleared line by line can still fail this test everywhere, because the test was never being run on the lines.

The invented Palash decision log records 240 decisions taken by 60 investors over eight quarters. The two tests come apart on a single sentence from that log. Somebody writes that 61 of the 84 sales in the log were positions in gain, so 72.6 per cent of sales were winners. The accuracy check passes. 61 divided by 84 is 72.619, or 72.6 to one place. Every figure survives. The other check, the one that asks what a reader ends up believing, finds that the sentence has already done its damage.

So the duty is a test on the impression. The harder half of the question is the second one: fair to whom. A sentence that misleads nobody in the room is not the same as a sentence that misleads nobody at all, and the rooms are different sizes. Three readers might meet the same line. The first already knows to ask what a share is measured against and goes looking for the missing figure. The second is reading at speed, takes the headline and stops. The third assumes, quite reasonably, that the number quoted was chosen because it is the one that answers the question.

The standard is set by the third reader, not the first. The reader who already knows was never going to be misled, so their impression is not evidence about anything. Writing for that reader is not really writing at all. The person the duty exists for is the one who has no reason to suspect the sentence and no time to audit it, and there is a plain reason why a duty of this shape would be aimed at them: they are the only reader whose impression the writer actually controls.

What is a Misleading Communication, and what are the three routes to one?

A misleading communicationOne whose effect is a false impression, whatever was intended. is one whose effect is a false impression. The definition leaves two things out: whether anything false was said, and what the writer was trying to do. Both of those are facts about the writer, and the definition is built out of facts about the reader instead. Once that shape is accepted, the interesting question stops being how somebody lies and becomes how somebody produces a false impression while saying nothing untrue.

There are three ways, and between them they cover most of what goes wrong in financial communication. The first is omissionLeaving out something that changes how the rest reads., where a fact that changes how everything else reads is simply not there. The second is prominenceHow likely a passage is to actually be read., where the qualifying fact is present, and set somewhere it will not be reached. The third is framingThe choice of wording for a fact, which changes decisions without changing the fact.. The same fact is stated in whichever of two true wordings suits the writer.

None of the three requires a false word, and care is therefore not the test. A writer can check every figure twice, find every one correct, and ship all three routes intact. The ordinary quality check is an accuracy check, and accuracy is not what failed. Shipping all three routes is therefore not a hypothetical about careless people but the ordinary result of a careful review. The three routes are the reason a conduct duty about communication has to be written in terms of effect. A duty written in terms of truthfulness would clear all three of them on the first reading.

The lever structure also explains why these three and not others. Each one is a separate lever, and closing one leaves the other two open. Supplying the missing figure fixes omission while leaving the frame exactly as it was. Moving the warning to the top fixes prominence while leaving the denominator missing. The three routes come one at a time below, the first worked on the log's own numbers, followed by a procedure that shuts all three at once.

Route one is omission: what happens when the denominator goes missing?

Every word of the cleanest example in the subject is true, so it is worth working slowly. Across eight quarters the Palash log recorded 84 sells. Of those, 61 were positions standing in gain and 23 were positions standing in loss. So 72.6 per cent of the sales were winners. Somebody could put that in a client note tomorrow, and an auditor checking figures would pass it. 61 divided by 84 really is 0.72619, and it really does round to 72.6.

The question the sentence never invites is the one that matters. Seventy two point six per cent of what? Of the sales. And who chose which positions became sales? The person whose deciding is being described. The decisions are the denominator, so a share measured over the decisions somebody took cannot say anything about the deciding. The figure is not a measurement of behaviour at all. The figure is a restatement of the behaviour with a percentage sign attached.

The measurement needs the other number: what was available to be sold. Across those eight quarters 240 positions were open at some point, 108 of them standing in gain and 132 standing in loss. Terrance Odean, writing in the Journal of Finance in 1998, is where this way of measuring comes from, and the whole contribution of that paper is the denominatorWhat a share is measured against, and the commonest thing left out.: count what could have been sold, not only what was.

Do the two divisions. Gains were realised at 61 out of 108, or 56.5 per cent. Losses were realised at 23 out of 132, or 17.4 per cent. Divide one realisation rateThe share of the positions of a given kind that were actually sold, out of all the positions of that kind that could have been sold. by the other and the ratio is 3.2 to 1: a position standing in gain was a little over three times as likely to be sold as a position standing in loss. The ratio is a statement about deciding. The 72.6 was not.

One line in that log makes it concrete. On 12 October, Meera Sundaram, an invented investor of 41 living on a salary, sold Suvarna Chemicals Limited whole at Rs 4,60,000/- against a cost of Rs 4,00,000/-, booking Rs 60,000/-, which is 15.0 per cent. The same day she kept Kesari Logistics Limited, then standing at Rs 1,95,000/- against a cost of Rs 3,00,000/-, saying she would sell it when it got back to what she paid. Her other two positions, the Vindhya index scheme and the Nilgiri mid-cap scheme, she left alone. The twelfth of October contributes an entry to the 61 and an entry to the 132, and a sentence built on the 61 alone will never let a reader see the second half of it.

Both statements are true and only one of them is a measurement. Route one is exactly that difference. Nothing was invented, nothing was rounded wrongly, and no reader was told anything that was not the case. A single number was left out, and it happened to be the number that turns a description of what somebody did into a description of how they decide.

How far can one missing denominator move a true sentence?

A reader who has followed the last section usually grants that the 72.6 is incomplete and then quietly files it as a small problem. The gap is not a small problem, and the way to feel its size is to hold the deciding perfectly still and change only the thing the sentence never mentions. Keep the two realisation rates exactly where the log put them, at 61 of 108 and 23 of 132. Change only the mix of positions that happened to be open.

Suppose the eight quarters had thrown up 50 positions in gain and 190 in loss instead. Apply the identical rates to that mix and the raw share of sales that were winners comes out at 46.0 per cent. Suppose instead they had thrown up 200 in gain and 40 in loss. The identical rates give a raw share of 94.2 per cent. The same conduct, unchanged in every respect, reads as anything from 46.0 to 94.2 per cent depending on a fact the sentence never mentions. The 3.2 ratio sits at 3.2 the whole way across.

Watch which number the picture rewards. The travelling figure is the one a reader finds easiest to repeat, and it is the one carrying no information about the deciding at all. The two figures that do carry the information sit flat and look boring. The asymmetry is not an accident of this log but a general feature of shares taken over what somebody chose to do, and the missing denominator earns a section of its own rather than a footnote.

And notice where the mix itself comes from. The mix is not a fact about the person at all. Across the same eight quarters the invented and illustrative Palash 100 index opened at 100.0, reached its peak of 131.0 at the end of the second quarter and its low of 104.0 at the end of the fourth. Peak to trough, the fall is 20.6 per cent. A run of that shape moves the count of positions standing in gain all by itself, with nobody deciding anything. So a sentence quoting the raw share reports the market as though it were the investor, and stacks a second false impression on the first.

Route two is prominence: is a warning that is present the same as a warning that is read?

Route one removed a fact. Route two leaves every fact in place and moves one of them. A rental agreement, a hospital consent form, the terms on a loan: everything was there. Almost nobody read it, and whoever laid it out knew roughly how much of it would be read. The question this route asks is whether a fact set where nobody reaches it has been communicated at all.

The Palash practice ran a reading test that puts a number on it. Its scheme document runs to 46 printed sides, with the risk statement on side 31 in eight point type. Thirty readers were given the document and afterwards asked to state the main risk. Seven could, or 23.3 per cent. The same risk, written as 90 words and placed at the top, was given to another thirty. Twenty four could state it afterwards, or 80.0 per cent. Nothing about the risk changed between the two rounds, and nothing about the readers did either.

The communication is what gets read; what is merely present is a fact about the file. The distinction sounds like word play until the two numbers are set beside it. The gap between 23.3 and 80.0 was produced by placement and type size alone. Marianne Bertrand and Adair Morse, in the Journal of Finance in 2011, tested exactly this on borrowing disclosures and found that where and how a fact is set changes what people do with it, not merely what they say they saw. John Beshears and his co-authors reported the same shape for simplified scheme disclosure in 2011.

The way gains and losses are weighed against a reference point stops being background here and starts doing work. If placement changed only how pleasant a document was to read, it would be a matter of courtesy. Because placement changes what the reader ends up believing, it is a matter of conduct, and burying a qualification becomes a choice with the same effect as leaving it out. The argument that the information was disclosed therefore cannot settle anything on its own: it answers a question about the document when the question was about the reader.

Route three is framing: can a communicator avoid choosing a frame at all?

The third route is the uncomfortable one. Nobody can opt out of it. The log's own gains figure shows why. Gains were realised at 56.5 per cent. So 43.5 per cent of the gains available were left alone. Losses were realised at 17.4 per cent. The other 82.6 per cent of the losses available were still being held at the end of the eight quarters. Every one of those four sentences is arithmetically correct, and they do not read the same way at all.

Daniel Kahneman and Amos Tversky, in Econometrica in 1979, set out the structure that explains why the wording moves people: outcomes are judged as gains or losses from a reference point rather than as final positions, and the same outcome described on either side of that point is not evaluated the same way. The Palash cohort shows the effect on itself. Asked what gain would make a fifty-fifty gamble against a Rs 10,000/- loss worth taking, the median answer was Rs 22,000/-. Offered a certain Rs 5,000/- against a half chance of Rs 11,000/-, 42 of the 60 took the certain amount, or 70.0 per cent. Offered a certain Rs 5,000/- loss against a half chance of losing Rs 11,000/-, 39 of the 60 took the chance, or 65.0 per cent.

Same 60 people, same afternoon, cautious on one side and chance-taking on the other. There is no unframed way to state a fact, so a writer cannot claim to have been neutral about a frame they had to choose. Every sentence about a proportion picks a side; every sentence about a change picks a direction. The honest response is not to hunt for the neutral wording. There is not one. Give both wordings instead, at a cost of one clause, and the lever is gone. The four-step procedure below does exactly that in its last step.

Why is a communication judged by its effect and not by what was intended?

Here is where most people go wrong, and it is not a failure of ethics. The mistake is a failure of definition. The instinct is to treat intent as the test: somebody who meant well has not misled anybody, and somebody who meant to deceive has. The instinct works fine for lying, a category defined by the speaker. The instinct fails for misleading, a category defined by the listener.

What is Informed Consent, and why is a signature not it?

Everything so far has been about what a writer produces. The remaining question is what a reader is taken to have agreed to, and it is where the three routes do their most expensive work. Informed consentAgreement given with understanding of what is being agreed to. is agreement given with understanding of the thing agreed to. A signature is a record that agreement was given. Consent and the signature are two different events, and only one of them leaves a mark on a file.

Draw the two as separate axes and the shape becomes obvious. A person can sign without understanding, producing a record of consent with no consent inside it. A person can understand without signing, producing consent that nothing can evidence. A person can do both, the case everybody assumes is happening. The signature axis is the one a file can show, and the understanding axis is the one consent actually lives on. A stack of signatures is therefore evidence about a process rather than evidence about consent.

Put the reading test from route two next to it and the size of the gap stops being theoretical. Thirty readers were handed the document of 46 printed sides. Seven could state the main risk afterwards. If all thirty had then signed, the file would hold thirty records and seven of them would sit in the top row. Nobody forged anything and nobody was rushed. The document simply set the risk statement where 23 of the 30 did not reach it, and a signature does not distinguish between a reader who reached it and a reader who did not.

There is a further turn here, covered separately. George Loewenstein, Daylian Cain and Sunita Sah, in the American Economic Review in 2011, and Cain, Loewenstein and Don Moore in the Journal of Legal Studies in 2005, found that disclosing a conflict can make the person receiving the disclosure worse off rather than better, partly because the disclosure licenses the person making it. The point to carry away is narrow: telling somebody a thing is not the same event as their understanding it, and the two can move in opposite directions. The narrow point is enough to show why consent has to be defined on the understanding axis, and the full argument is covered separately.

How to Communicate Financial Uncertainty Fairly, and which step gets skipped?

Everything above is diagnosis. The working method is four steps in a fixed order. The three routes are separate levers, so a habit that closes one of them leaves the others exactly where they were. Run these four in sequence and all three shut together.

Step one: give the range before the point estimateA single number given as the answer, with no spread around it.. Step two: name the denominator in the same sentence as the share. Step three: state what would change the answer. Step four: give both frames. People skip step one. A single number reads as competence and a range reads as hedging, and the reading is exactly backwards. A single number claims knowledge the writer does not have, and a reader who takes it at face value is being handed a false impression of precision by somebody who never said anything false.

The order does its own work. A note that begins with the point estimate and adds the range afterwards leaves the reader anchored on the number, and the range reads as a qualification, in a second run of route two. Naming the denominator only in a footnote leaves route one open. Giving one frame and stopping leaves route three open. The order is not decoration: each step is placed where the reader meets it before the thing it qualifies.

Applied to the log's own sentence, the rewrite is not a retraction. Nothing false went in, so nothing false comes out. The rewrite adds the denominator, a statement of what the eight quarters could and could not support, and the second frame. The result is longer and it is the only version of the two that a reader taking it at face value is not misled by.

One caution belongs inside the procedure rather than after it. The Palash log covers 60 people over eight quarters, far too small to support a claim that any of this changes what somebody earns. Twenty of the 60 adopted a written checklist on 4 November; across quarters five to eight they recorded a written reason on 34 of their 41 decisions, being 82.9 per cent, against 19 of 63 for the other 40, being 30.2 per cent, and their realisation ratio fell from 3.2 to 1.6. The checklist figures are about process and about nothing else. A note that let a reader draw a return conclusion out of them would be running route one on its own reader.

Where exactly does the line fall, stated concretely?

Abstract tests are easy to agree with and hard to apply, so here is the line drawn on four ordinary things every communication has to do. Each row holds two true versions of the same task. Truth is constant across the whole table, so the difference between the columns is never truth. The difference is what a reader is left holding.

Read the table twice, once down each column. Down the left, every cell gives the reader the thing they would need to ask for. Down the right, every cell gives the reader a sentence that answers the question they asked while leaving them worse informed than a silence would have. The line falls where a reader taking the sentence at face value would form a different picture from a reader who had the working, and it falls there whether or not anybody meant it to.

Richard Thaler and Cass Sunstein, in Nudge in 2008, made the same structural point about layouts rather than about sentences: there is no neutral way to present a set of options, so whoever lays them out has chosen something whether they wanted to or not. The line between arranging a choice honestly and arranging it to produce a particular outcome falls in the same place as the line in the table above, and for the same reason. Truth is available on both sides of both lines, so neither line runs along truth.

The test has one property worth naming, and it is the property that makes the test usable. The test can be run without knowing anything about the writer. No access is needed to the writer's motives, their file, or their review process. The test needs the sentence, the reader it is aimed at, and the working behind it, and two people running it on the same communication will usually agree. A test built on intent has none of that. A duty of this shape would be written the way it is for exactly that reason.

How does somebody actually use this on a Tuesday morning?

A conduct idea has to survive contact with an ordinary week, so here it is for four people who are not thinking about conduct at all. A household reading a scheme summary before putting Rs 25,000/- a month into it by standing instruction has one job: find what the headline share is measured against. If the sentence does not say, they have not been given a number yet, they have been given a mood.

A lender reading a borrower's own account of their repayment record runs route two instead. Not what does the file say, but how many sheets had to be turned to reach the qualification, and whether somebody without a lender's training would have turned them. An analyst reading a scheme's own commentary runs route three: which frame is this in, and what does the identical fact sound like stated the other way. A person deciding alone, with no adviser and no committee, runs step three of the procedure on themselves, asking what would change this answer and whether anybody volunteered it or they had to go looking.

Devika Rao, the invented adviser at Palash Advisory Services Private Limited, uses the four questions as a drafting check rather than a reading check, the same instrument pointed the other way. Before a note goes out she reads it as the third reader from the opening section: somebody with no reason to suspect it and no time to audit it. The cheapest check catches the most: read the drafted sentence as somebody who does not already know the answer. None of the four questions asks whether anything is true. All four ask what impression is being built, and the impression is the only thing the duty is about.

What can behavioural reasoning never settle about a requirement?

Everything above is an argument about why a duty shaped like fair presentation would exist, not an account of what any rule requires. The division matters for a practical reason: behavioural reasoning and legal requirement move at different speeds, and an account that mixed them would go stale in one place and mislead in the other.

The four-step procedure is a working method rather than a specification: it fixes no threshold, no period, no rate, no minimum, no format and no prescribed wording. The behavioural case for a duty of this shape is not the duty itself.

Meera Sundaram, Devika Rao, Palash Advisory Services Private Limited, the Palash decision log, the Palash 100 index, the Vindhya index scheme, the Nilgiri mid-cap scheme, Suvarna Chemicals Limited and Kesari Logistics Limited are invented.
Educational material. Not advice on any investment, tax, budget or market position.