How to Document an Ethical Decision in Six Steps
An ethical decision is documented in six steps: the situation stated factually with the date, both duties named along with who each one is owed to, the options actually considered listed, the one chosen named, what it cost said plainly, and then the record dated and filed. The record is written the same day. A record written afterwards records the outcome rather than the decision.
The whole artefact is nine lines long. On the invented assignment worked through below, the writing took nine minutes and the filing took the rest. The whole exercise fits inside ten minutes on a single afternoon. Ten minutes is the number worth holding on to. Nearly every failure below begins with somebody picturing a much larger job and quietly moving it to next week. A record that takes an hour is a record that gets written in September about something decided in March, and that version is close to worthless.
Records of this shape are already kept outside work without being called anything. When Rs 40,000/- is lent to a cousin, the date and the terms go down in writing, not because trouble is expected but because in eighteen months neither party will agree on whether repayment was meant to start after the wedding or after the festival. The instinct is to set the terms down while they are still known. Knowing them later is simply not on offer.
The difference between an ethical decision record and an ordinary one is narrower than it sounds. An ordinary decision record holds the question and the answer, and that is the right artefact for ordinary work. An ethical decision recordA short written account of a choice between two duties. holds a choice between two duties, and the difference shows up in the fields rather than in the tone: two duties with a party named against each, the options actually weighed, and what the choice cost. A record carrying none of those three is a good decision record and a different one.
The Kavery research desk, inside Kavery Capital Services Private Limited, looked at 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 traced the figures, and Latha Menon took the decision on 12 March. The working file held twenty one documents: twelve public, six confidential, and three that had arrived under a written undertakingA signed promise about how a document may be used, given before the document was handed over.. One of those three explained, in a single sentence, why the operating margin had fallen.
Nobody on the Kavery desk behaves badly. The choices that most need recording arrive between colleagues who are all doing their jobs properly, and that is exactly why nobody thinks to write them down. The record is also not a shield. A record does not stop a consequence, and any account that suggests otherwise is selling comfort: what it does is preserve what was known at the moment of choosing, so the choice can be reviewed on the facts that were genuinely available. Preserving what was known is a smaller claim than the comfortable one and a considerably more honest one, and it is the only version that survives contact with somebody rereading the file.
Nothing in the six steps needs standing, permission, a template or a system. Nine lines in a plain file, written by a person in their first month who has met almost nobody, is the entire artefact, and no step below asks for anything more than that.
Why Document an Ethical Decision at All?
Because the outcome, once it arrives, overwrites the thing most worth keeping. On 9 March the person writing the note did not know whether the credit line would be cut, whether anybody would ask about the missing explanation, or whether the year would prove them right. By June, when the line came back for review, all of that was settled and none of it could be unlearned. The state of knowledgeWhat was actually known at the time of choosing, before any of it was settled. at the moment of choosing is what vanished in between, and that state is the only fair basis on which a choice between two duties can ever be judged.
The claim is easy to hear as a platitude, and precision is what stops it being one. A choice made under two live duties can be made carefully, in full sight of both, and still land somewhere nobody wanted. The reverse also happens: a careless choice lands well and looks like judgement. Judging the choice apart from the way it turned out is only possible if something written at the time survives, and the record is that something. The separation of decision quality from outcome quality, and the habit of refusing to score one by the other, belongs to Annie Duke, Thinking in Bets, 2018.
There is a second reason and it is more immediate. Ninety days after a decision, the review is not conducted by a historian. The review is done by somebody with a few minutes per file. A gap in a note is visible to that reader. The reason for the gap is not. A gap that was chosen and a gap that was missed look identical from the outside. The record is the only thing that distinguishes them, and it costs nine minutes.
What does the record preserve that memory cannot?
What Are the Six Steps, in Order?
Six steps, run once, in sequence, and each one finishes before the next begins. Nothing loops back and nothing waits for approval. The figure below sets them out with the number of lines each one adds to the record, and the lines are the point: the whole record is nine of them, and knowing that in advance is what makes the thing get written.
- State the situation, with the dateWhat was in front of the writer, in plain facts. No reasoning, no conclusion, no adjectives about anybody.Complete when a stranger could read the line in six months and see the same starting point the writer had.
- Name both duties, and who each is owed toOne sentence per duty, and a named party against each of them.Complete when two different parties appear on the record.
- List the options actually consideredEvery course weighed, including the ones thrown out in seconds.Complete when the list matches what really went through the writer's head, not what reads well.
- Name the option chosenOne sentence saying which of the listed options was taken.Complete when the chosen line points at a line already on the list above it.
- Say what the choice costPlainly, in the words anybody would use out loud, with nothing rounded off.Complete when a cost is named rather than implied.
- Date it, sign it, and file itThe writer's name, the day it was written, and a location a reviewer can reach.Complete when the record sits somewhere findable and is not sitting inside the deliverable.
Step One: What Is Written Before Any Reasoning?
Write the facts and the date, and stop. Four things go down: what arrived, what it said, what it arrived under, and when. On the Kavery assignment that line read: on 9 March, one of the three restricted documents in the working file states the cause of the fall in operating margin in a single sentence, and it arrived under a written undertaking.
Nothing about what it means, nothing about what the writer intends to do, and nothing about the person who sent it. If the line is running past two sentences, step two has begun without waiting for permission, and the usual sign of that is the appearance of the word because.
Step one is complete when a dated factual line exists that a stranger could read in six months and arrive at the same starting point the writer had, and only then does step two begin. On the assignment this took Sharada Iyer about two minutes.
Why does the situation get stated before any reasoning?
Step Two: Which Two Duties Go Down, and Who Is Each One Owed To?
Two lines, one duty each, and a named party on every line. Not a category of person and not the organisation in the abstract. On this assignment the two lines read: completeness, owed to the person relying on the note to take a decision; and the written undertaking, owed to whoever supplied the restricted document.
The party is not decoration on the end of the sentence. The named party is half the content of the line, and a record where both duties point at the same party is saying something about the record rather than about the situation. Each line stays at one sentence. If a duty needs a paragraph to state, it is being argued rather than named, and argument does not go in this record at all.
Step two is complete when two duties and two different parties are on the record, in two sentences, with nothing weighing them against each other yet. Two minutes on the Kavery assignment, and that includes rereading them once.
Why name who each duty is owed to, rather than only what each one requires?
Step Three: Which Options Belong in the List?
All of them. Every course that went through the writer's head, written one per line, in the order they occurred rather than in order of merit. The options consideredEvery course actually weighed, including the ones rejected quickly. on the Kavery assignment came to three lines: use the restricted sentence in the note; use the point without attributing it to the document, rejected in about two seconds; and reach the point from the public material instead, stating the limit in the note.
The middle line is the one people leave out, and it is the one that carries the most information. The line was rejected almost instantly, it does not read well, and writing it down feels faintly like confessing to something. The line goes in anyway. Rejecting it took no time, so recording it takes no time, and its presence is what shows a later reader where the writer's lines actually sit.
Step three is complete when the list on the record matches the list that went through the writer's head, including the entries the writer would prefer nobody saw. Three lines, about three minutes.
One option was rejected in about two seconds. Does it go in the list?
Step Four: How Is the Chosen Option Named?
One line, pointing at one of the lines above it. Not a fresh description, not a summary of the reasoning, and not a hybrid of two options that never actually existed. On the Kavery assignment the line read: chosen, option three, reach the conclusion from the public material and state the limit in the note.
The reason for keeping it short is mechanical rather than stylistic. When the chosen line points at a listed option, a reader can see instantly which alternatives were live at the moment of choosing. When it is written afresh, the list above stops functioning as a list and becomes background colour, and the connection between what was available and what was taken quietly comes apart.
Step four is complete when one line names one option that already appears on the list, and nothing else has been added. About a minute, and most of that is resisting the pull to justify it.
Step Five: What Goes in the Cost Line?
Whatever the choice gave up, said plainly. The chosen option on this assignment left the note weaker at exactly the point a reader most wants strength, so the cost line read: the note does not establish the cause of the margin fall, and a reader may take that gap as a sign the desk did not look into it. A sentence like that is the costWhat the chosen option gave up, written in the words anybody would use out loud., stated in the words anybody would use out loud.
Resist three temptations here, and they arrive in this order. The first is to soften the wording. The second is to add the reason the cost was acceptable. No reason belongs anywhere in this record. The third is to leave the line out entirely on the grounds that the cost is obvious. The cost is obvious today and will not be in June.
Step five is complete when a cost is named on the record rather than implied around it, and a record that names no cost is describing something that was never a collision between two duties. One minute.
The choice made the note weaker. Does that go down in writing?
Step Six: Where Does It Get Dated, Signed and Filed?
The writer's name, the date of writing, and a location. The location is the part that needs a moment. A record of this kind often names a document that other people may not read, and it therefore has to be findable but not browsableFiled where somebody reviewing the work can retrieve it, and where somebody merely reading the deliverable does not meet it.. On the Kavery assignment it went with the workpaper behind the note, dated 9 March and signed by Sharada Iyer, and not inside the note itself.
Three destinations present themselves and only one of them survives the two tests. Inside the note puts a restricted document title in front of every person who opens the deliverable. A private file on the writer's own machine passes the second test and fails the first. Failing the first is the same as not writing the record at all. With the workpaper passes both. A reviewer already goes there and nobody else has a reason to.
Step six is complete when the record is dated, carries a name, and sits where somebody reviewing the work would reach it without being told it exists. Under a minute. The nine minutes of writing turn into under ten minutes in total once the filing is added.
Where does the record go, given that it names a confidential document?
How Long Should the Whole Thing Take?
Under ten minutes, and that ceiling is doing more work than any other instruction in this guide. Length is the enemy of the whole exercise. Somebody who believes this record takes an hour will not open it on the afternoon of 9 March; they will note it as something to write up properly later, and later has a different set of facts in it.
So the nine lines are a budget rather than a suggestion. Two minutes on the situation, two on the two duties, three on the three options, one on the chosen line, one on the cost, and the filing on top. If a record is running to twenty five minutes, the thing to look for is the argument that has crept in: almost always it is the chosen line defending itself, or a duty being justified instead of named.
Before the worked record: how long should documenting an ethical decision take?
What Did the Record on This Assignment Actually Say?
Here it is in full. Nine lines, written on the afternoon of 9 March, nine minutes of writing plus the filing, and no draft. Read it once as a whole and then look at what is absent: no argument, no justification, and no attempt to persuade anybody that the choice was correct.
| Line | Field | As it was written on 9 March |
|---|---|---|
| 1 | Situation | 9 March. One of the three restricted documents in the working file states the cause of the fall in operating margin in a single sentence. The document arrived under a written undertaking. |
| 2 | Duty one | Completeness, owed to the person relying on the note to take a decision. |
| 3 | Duty two | The written undertaking, owed to whoever supplied the restricted document. |
| 4 | Option one | Use the restricted sentence in the note. |
| 5 | Option two | Use the point without attributing it to the document. Rejected in about two seconds as the same breach wearing a disguise. |
| 6 | Option three | Reach the point from the public material instead, carrying it less crisply, and state the limit in the note. |
| 7 | Chosen | Option three. |
| 8 | Cost | The note does not establish the cause of the margin fall, and a reader may take that gap as a sign the desk did not look into it. |
| 9 | Dated, signed, filed | 9 March, Sharada Iyer. Filed with the workpaper behind the note, not inside the note. |
Two details are worth pausing on. Line five is the one that would go missing from any version written later, and it is also the shortest thing on the record. Line eight names a cost that the desk would rather not have printed anywhere, and it is stated in the flat words somebody would use in a corridor. Neither line argues; together they are what let a reader in June see a chosen gap rather than a careless one.
What Does the Record Do in June That Nothing Else Does?
Ninety days after the committee met on 12 March, the credit line came back for review, exactly as the decision said it would. Somebody opens the note, reaches the paragraph where the cause of the margin fall should be and is not, and has to decide in about a minute what that gap means. Two readings are available: the desk chose not to establish the cause, or the desk never looked. The note alone cannot tell them apart.
The workpaper behind it can. The reviewer opens the nine lines, sees a dated situation, two named duties, three options with one rejected in seconds, a chosen option and a stated cost, and the question closes. The record does not tell the reviewer that the choice was right. The record tells the reviewer that both duties were in view when the choice was made, and nothing else about a decision can actually be demonstrated afterwards.
Now the honest limit. Accounts of this kind usually overclaim at exactly this point. The record does not stop the reviewer disagreeing. The record does not stop the line being cut further, and it does not stop anybody forming a poor opinion of the work. The record stops one specific and very common failure. A considered choice gets read as carelessness when nothing survives to say otherwise. The benefit is narrow and it is real, and the narrowness is worth keeping in view rather than trading up for a comforting version.
Does the record protect the writer if the decision is questioned later?
Before the next block: a record written six months later, honestly, by somebody who knows the outcome. What is missing from it?
What Goes Wrong When the Record Is Written After the Review?
The record written in September
Six months after the decision, during a cold rebuild of the workpaper, somebody notices that the choice about the restricted document was never written down and sets out to fix it properly. The result is thorough, careful and honest. The rebuilt record is also useless, and nobody in the room can tell.
The September version has a situation line, two duties, a chosen option and a cost. In place of three options it has one, possibly two. Option two, use the point without attributing it, does not appear anywhere. By September that option is unthinkable, and the person writing genuinely does not remember entertaining it for two seconds in March. Option one, use the restricted sentence, has softened into something more like a hypothetical. The cost line is present and reads smaller than it felt on the day. The note was in fact accepted, and the gap in fact went unremarked.
No dishonesty occurred at any point. HindsightKnowing how something turned out, and being unable to set that knowledge aside. is not a bias anybody can decide not to have, and a person cannot notice the options their own memory has already tidied away. The cost is precise: the record now describes a decision nobody made, and because it is the only surviving account, that fabricated version is what the next reviewer will read as fact.
Look at the two panels for a moment longer than feels necessary. Five of the nine fields are word for word identical, the ninth simply carries a later date, and the September version reads as the more polished of the two. The trap is finished at that point: three lines out of nine carry all the damage, two of them inside the options list that nobody ever audits, and the overall impression of quality goes up rather than down.
Does Any of This Depend on the Place of Work?
Almost none of it. Stating a situation with a date, naming two duties and their parties, listing what was weighed, saying what was chosen and what it cost, then signing and filing it, would read the same in any country, any firm and any subject. There is no local variant of a dated line. The six steps are not jurisdictional.
The jurisdictional part sits outside the procedure and is narrower than people expect. Whether a conduct duty binds a particular role, who sets it, and what a firm must keep are questions answered by named bodies. The Securities and Exchange Board of India at sebi.gov.in sets conduct duties over registered intermediaries in the securities markets and the people working inside them. The Institute of Chartered Accountants of India at icai.org sets professional conduct standards binding on its members personally. The International Organization of Securities Commissions at iosco.org publishes conduct principles that several national regimes draw on. A remembered requirement is worse than an absent one: it reads like knowledge and it goes out of date silently. The requirement is read at the source, with the date of reading noted.
How Does a Reviewer, a Lender or a Household Actually Use Nine Lines?
Sit on the receiving side. The artefact is finally cashed there. Prakash Nadar, reviewing, has minutes rather than hours for anything that is not his own work. He opens nine lines and can see in one pass what was in front of the writer, the two duties that were live, what was available, what was taken and what it cost. He can accept it, question it, or carry it upward, without opening a single other document. The record is nine lines rather than nine paragraphs for exactly that reason.
A lending committee uses it in the same way and at greater speed. On 12 March, on the Kavery assignment, Latha Menon took a line from Rs 40,00,000/- to Rs 25,00,000/- and set it to be looked at again in 90 days. Somebody holding that decision in June needs to know whether the missing explanation in the note was a hole or a choice, and nine lines answers that in about twenty seconds. Where no record exists, the honest committee answer is that nobody can tell, and the practical answer is that they assume the less flattering of the two.
An analyst reading somebody else's old work uses it as a map of what was not available. A household uses the same instinct at a different scale. When a wedding caterer offers a discount for cash and the household declines it, writing three lines that evening about what was being weighed and what the refusal cost sounds excessive until eight months later, when somebody in the household asks why the bill was higher than the quote. The value in every one of these cases comes from the same place. A short thing written before the outcome is the only artefact that can be trusted after it.
Why a Duty Is Not a Quantity
A relationship between two quantities becomes obvious the moment one can be pushed and the other watched responding. A duty is not a quantity. A duty is a judgement, and a slider labelled how serious is this, or how much detail does this deserve, would turn a judgement into a setting. The setting would then be wrong in precisely the situations that matter. Nobody drew the scale for those.
There is no correct level to discover here and no score attaching to a record, so a control offering one would be teaching arithmetic where the subject is a sequence of six things done in order. The exercise that fits instead is a short one: a choice already made between two obligations, a blank file, and an attempt at the nine lines in under ten minutes. Whatever cannot be recovered is the part the outcome has already eaten, and noticing that is most of the lesson.
Resolving a dilemma is a separate procedure that comes before any of the six steps, and it is set out under the ethical dilemma, where two duties conflict. The decision record kept for ordinary work holds the question and the answer rather than two duties and a cost, and it is set out under the decision record and the workpaper behind it, along with the source log. Classifying documents by sensitivity is covered under handling information by classification and provenance. Judging the quality of a decision apart from the way it turned out is treated under decision quality, judging the decision rather than the outcome, and that idea belongs to Annie Duke, Thinking in Bets, 2018.
References
| Source | Document | Where |
|---|---|---|
| Securities and Exchange Board of India | Conduct duties applying to registered intermediaries and the persons working within them, under which some choices made during work are later examined | sebi.gov.in |
| Institute of Chartered Accountants of India | Professional conduct standards binding on members personally, attaching to a person rather than to an employer | icai.org |
| International Organization of Securities Commissions | Published conduct principles that several national regimes draw on, setting cross-border conduct expectations over market intermediaries | iosco.org |
| Annie Duke | Thinking in Bets, published in 2018. The separation of decision quality from outcome quality, and the refusal to score a choice by the way it turned out, are hers | published book, 2018 |
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.
