Vulnerable Investors: Recognising Heightened Duty of Care

Vulnerability is a situation, not a type of person. Bereavement, illness, job loss, caring responsibilities and financial pressure all produce it, most are temporary, and almost everybody passes through one. Vulnerability usually cannot be seen. A duty framed as recognising the vulnerable and treating them differently therefore does not work.

A question with no finance in it at all opens the subject. In the last ten years, has there been a stretch of weeks in which a person would rather not have signed anything important? Most people, asked quietly, can name one. A death in the household. A diagnosis, or somebody else's diagnosis and all the driving and phone calls that came with it. A job that ended on a Friday with no warning on the Thursday. A month when the money going out stopped matching the money coming in and every evening was spent on arithmetic. Anybody who can name such a stretch has already been a vulnerable investor, and was not a different kind of person while it lasted. That one observation is the whole argument, and everything below is worked out from it.

Heightened care is owed to somebody in difficulty, and it consists of particular things done in practice. Who qualifies for it is settled by rule rather than by argument, and the rule is set by the regulator. Vulnerability in this sense is not a clinical or diagnostic condition, and no individual can be assessed for it from the outside at all.

What makes somebody vulnerable, and is it a category of person?

The word is quietly doing two different jobs in ordinary speech, and separating them is most of the work. VulnerabilityA situation that reduces somebody's capacity to act in their own interest. in the sense used here is a situation that reduces a person's capacityThe room somebody has to absorb a bad outcome, in money, in time and in attention. to act in their own interest at a particular moment. The situation reading is a statement about circumstances. The other job the word does in everyday speech is to sort people into a bin marked fragile, and that is a statement about persons. The first is useful and the second is false, and they sound almost identical when spoken aloud.

Think about a street vendor who has run a tea stall outside one office building for eleven years. He knows his margins better than most salaried people know their own outgoings. Then the building is closed for four months of repairs. For those four months he is negotiating with a supplier while sleeping badly and counting a shrinking reserve, and any decision he takes in that period is taken under a pressure that was not there in month one and will not be there in month eighteen. Nothing about him has changed. His arithmetic is as good as it was. The room around the decision changed. The correct object of the word vulnerable is the situation, not the person standing in it, and that single grammatical move decides everything that follows.

The situation reading buys something the category reading does not. Reading vulnerability as a situation gives care a natural end. If the reason for taking more time with somebody is that they are three weeks into settling an estate, then the reason expires when the estate is settled, and nobody has to be told they have been reclassified. There is no list, so the awkward question of who is on it never arises. And the same person falls inside the description on some days and outside it on others, exactly as people actually live.

Which circumstances produce it, and how long do they usually last?

The circumstances are ordinary and they are nameable. Bereavement. Illness, whether a person's own or that of somebody they are looking after. The end of a job, or the end of a business. Caring responsibilities that arrive without notice and take the hours once used for thinking. A separation. Financial pressure severe enough that the next payment is the only thing in view. Isolation, where there is nobody to test a decision against before taking it. None of these is exotic and none of them is rare. Their ordinariness is precisely the point.

These circumstances have one thing in common, and it is not weakness. Each one consumes a scarce input. Deciding well takes attention, time and the ability to imagine an outcome that is more than a fortnight away, and every circumstance in that list eats at least one of those. A household absorbing a funeral has no spare evenings. A person waiting on a test result has no spare attention. Somebody two payments behind has a horizon of eleven days, and a question about an eleven-year education goal simply does not fit inside it.

Most of these situations are temporaryLasting for a period, which most vulnerable circumstances are.. The word matters. Not all: some circumstances are long-running. But the ordinary case is a period with a beginning and an end, and a person who is inside one this quarter is outside it in some other quarter. Because these episodes have ends, any duty attached to them must have an end too, and a process that cannot switch itself off has misunderstood what it was responding to.

Take the invented case this sequence has used throughout. Meera Sundaram is forty one and salaried, investing on her own account through Palash Advisory Services Private Limited, where Devika Rao is the adviser. Rs 25,000/- goes in every month by standing instruction against monthly outgo of Rs 55,000/-. The reserve is Rs 1,10,000/-, exactly two months of that outgo. There is a Rs 2,40,000/- deposit paying 6.5 per cent sitting alongside Rs 1,80,000/- of card borrowing costing 36.0 per cent. Clearing the borrowing out of the deposit would save Rs 64,800/- of interest and forgo Rs 11,700/-, so keeping the two apart costs Rs 53,100/- a year.

Nothing in those figures says Meera Sundaram is vulnerable. The figures say something narrower and more useful. A two month reserve is a two month distance between an ordinary event and a pressured decision, and that distance is a property of the arrangement rather than of the person holding it. Anybody, however capable, standing behind a two month reserve when the income stops is going to be making decisions in month three under conditions they did not choose. The situation reading holds there in one worked instance.

Why can vulnerability usually not be seen?

The argument turns here. Granting everything above, the obvious move is to look out for the situation and respond when it is found. The move fails, and it fails for reasons that have nothing whatever to do with anybody being careless or inattentive.

Start with what a financial conversation is actually made of. A form. Some numbers. A set of questions about goals and a horizon. A signature at the end. Bereavement does not appear on a form. A diagnosis does not appear in a valuation. The pressure of an eleven day horizon does not appear anywhere in a document that is asking about an eleven year one, and the person on the other side of the table is not going to volunteer it in answer to question seven. Most of the circumstances listed above leave no trace at all in the material a financial conversation is made of, so there is nothing there for even the most attentive person to spot.

Now add a second obstacle, worse than the first. Several of those circumstances give people positive reasons to say nothing. ConcealmentReasons somebody has for not disclosing a difficulty. here is not deceit and it is not shame in any simple sense. Concealment is a mixture of privacy, of not wanting to be treated differently, of not wanting a difficulty made real by being said out loud, and of a perfectly sensible worry about what happens to a person once the word gets attached to them. Somebody two payments behind does not lead with it. Somebody whose job ended in March may still be saying the same thing in July about how work is going. A person managing a parent's illness will mention it to a friend and not to a professional they see twice a year.

The two obstacles together give the shape that governs everything else. The reasons for saying nothing get stronger as the difficulty gets worse. A mild inconvenience is mentioned in passing. A serious one is not, precisely because it is serious. The visible signal therefore weakens as the severity rises. The signal is faintest exactly where the need for care is greatest, and no amount of closer attention fixes that.

There is a third obstacle sitting underneath both, and it is purely arithmetical. Even where a sign does exist and nobody is hiding anything, somebody has to be present at the moment it shows. Contact between an adviser and a person advised is occasional by design. Two conversations in a quarter is a normal arrangement and a generous one. Thirteen weeks contain ninety one days, and two days of contact is about 2.2 per cent of the period. Roughly ninety eight per cent of any episode runs its course with nobody in the room. The arithmetic is illustrative, and its direction is not negotiable: an occasional observer sees an occasional slice.

What does the research actually say about capability changing over time?

A body of work exists on how performance on financial tasks varies with age. Because it is the single easiest finding in the subject to misuse, it has to be stated with more care than almost anything else. Sumit Agarwal, John Driscoll, Xavier Gabaix and David Laibson set it out in The Age of Reason in 2009. Measuring performance across a range of ordinary financial tasks, they found that it tends to peak somewhere in midlife and to be lower on average after that.

Read the next sentence twice. The claim is the finding itself, not a qualification tacked onto the end of it. The Age of Reason result is a finding about averages across a large population, and it says nothing whatever about any individual person. The result is not a statement about a reader's grandmother, nor about a person sitting across a table. An average is a property of a group, and a group is not a person. No finding of this kind can be a statement about an individual at all.

The reason this matters is not politeness. A population averageA figure describing a group, which does not transfer to a member of it. is compatible with an enormous spread underneath it. Two groups can have different averages while their spreads sit almost entirely on top of one another, and when that is the case, knowing which group somebody belongs to moves what is known about that particular person by very close to nothing. The average has moved. The person has not been located. Amos Tversky and Daniel Kahneman made the general version of this point in Judgment under Uncertainty, in Science in 1974, when they showed how readily people read a group property onto a single case and how badly that inference performs.

The identical error appears elsewhere wearing different clothes, and the two are worth putting side by side because seeing the shape twice is what makes it stick. The invented Palash decision log sorts sixty investors into five turnover groups of twelve. Net returns across those groups run 10.9, 10.4, 9.6, 8.4 and 6.9 per cent, a spread from the lowest turnover group to the highest of 4.0 percentage points. Gross returns across the same five groups run 11.2, 11.0, 11.1, 10.9 and 11.0 per cent, a spread of only 0.3 points. Now suppose an analyst meets one of the twelve investors in the fifth group. How much is known about that person's net return? The average of the twelve is known, and about the one person in front of the analyst the group average says nothing. Reading it down onto that person is exactly the same error as reading the age finding onto a particular older person.

So what does the research licence? The research licences a claim about a population, and that is a real and useful thing for anybody designing a process many people will pass through. The research does not licence a claim about a person. And it certainly does not licence the move made most often in practice: treating somebody as less able to decide for themselves because of the year they were born in. The move imposes a real cost, in autonomy and in dignity, on the basis of a statistic that was never about them. Nothing in the research supports it.

What has to be true for a duty built on noticing to work?

One worked instance carries the rest of the argument. Doing it slowly and in the open is worth the space. Suppose the trigger for heightened care is that somebody notices a person is in difficulty. The plain engineering question is what would have to be true for that trigger to fire. The answer is three separate things, and they have to hold together rather than one at a time.

First, the difficulty has to produce a sign that is visible from the outside. Second, somebody has to be present at the moment the sign appears. Third, whoever is present has to read the sign as difficulty rather than as an ordinary variation in how a person happens to be that day. Those three conditions are the identification problemThe difficulty that vulnerability is usually not visible. stated as machinery instead of as a complaint, and stating it that way is what makes it possible to see where it breaks.

The three in turn, against what has already been established. On the first, most of the listed circumstances produce no sign at all in a financial setting, and several are actively kept private. On the second, contact is occasional by design, so most of any episode passes with nobody in the room. On the third, distinguishing genuine difficulty from an ordinary tired afternoon means drawing a conclusion from a single observation. A single observation supports very little, and Tversky and Kahneman named that the small sample problem. Concealment grows with severity, so a duty built on noticing fails at all three steps rather than at one, and it fails hardest for the people who need it most.

The arithmetic of three conditions in series is the part people get wrong, and they get it wrong in a specific direction. Give each of the three steps an even chance of holding, a coin toss each. Most readers guess the whole thing works something like a third or a half of the time. The chain works one time in eight. Half multiplied by half multiplied by half is one eighth, or 12.5 per cent, and that is with each individual step being as reliable as a fair coin. Chances in series multiply rather than average, so a chain of three merely adequate steps is not adequate, it is poor.

How fast does a chain of three steps actually decay?

The guess is where the learning is. Commit to one before touching the control below. Seven times in ten for each step sounds fairly solid for any human process. What fraction of the time does the whole chain complete? Write the number down first.

What does heightened care actually consist of?

The obvious response has now been dismantled with nothing yet put in its place, and a certain unease is appropriate. The replacement follows, and the first thing to say about it is that heightened carePractices applied where somebody may be less able to protect themselves. is not an attitude, a sympathy or a state of mind. Heightened care is a short list of concrete things that get done, and every one of them can be described, taught and checked.

The second item is where most of the value sits, and it is the one that gets skipped, so it deserves its own paragraph. Informed consent, set out under communication conduct, is understanding rather than agreement, and the distinction is not academic. Asking whether something is clear tests politeness. Almost everybody says yes. Saying yes is the shortest route out of the room, and a person having a difficult month takes it faster than anybody. Asking instead what would happen to this arrangement if it fell by a fifth next year forces an answer that cannot be produced by nodding, and the quality of that answer is the only honest evidence that anything landed.

Notice something about all five practices that is easy to miss on a first read. Not one of them requires knowing anything about the person. Slowing the pace does not require a diagnosis. Asking what happens if it falls does not require knowing that anybody is bereaved. Offering a second conversation does not require a judgement about capability. Every item on the list is a practice that can be applied without identifying anybody at all, and that property is not a happy accident, it is the whole design.

Why must the response be structural rather than a matter of noticing?

Put the last two sections next to each other and the conclusion writes itself. Noticing fails at three steps in series. The practices that constitute care require no noticing at all. So stop using the first as the trigger for the second. Dropping the first as a trigger is what is meant by a structural responseOne that works without anybody having to notice anything., and it is the closing idea of the whole subject.

Everyday life is full of the same move, and it is accepted everywhere else without comment. A ramp at a doorway does not require anybody to assess who needs it. A handrail on a staircase is not fitted after somebody has been identified as unsteady. Lighting in a stairwell is not switched on for particular people. In each case the alternative design, the one that waits to identify who needs the help and then provides it, would be worse for everybody: slower, more intrusive, and reliably wrong about who needs what. The built environment settled this argument decades ago, and a financial conversation is a built environment too.

The coverage difference is the part worth holding on to. A process that waits to notice reaches only those occasions where the whole three step chain happened to complete, and which occasions those are is decided largely by chance. A process built for the harder case reaches every occasion that passes through it. There is nothing to reach for. Take the illustrative 12.5 per cent default from the control above and draw it out: one occasion in eight against eight in eight. The gap is not a marginal improvement in a rate. One in eight against eight in eight is the difference between a mechanism that works and one that does not.

Two objections are worth answering because both are reasonable. The first is cost: does treating every conversation as the harder case not make everything slower for everybody? Sometimes, a little. But look at the five practices again and ask who is actively harmed by a slower pace, a question that tests understanding, an offer of a second conversation, or a written record of what was said. The answer is nobody. A design that helps some people considerably and costs the rest a small amount of time is an ordinary and well understood trade, and it is the same one every handrail makes. The second objection is that a default is being set for people. Who chose that default is a fair question. Richard Thaler and Cass Sunstein worked through the terrain in Nudge in 2008, and the answer that survives is that the default has to be one a person can decline without difficulty and without explanation.

A change of exactly that shape has already been measured in the invented cohort, and it shows the structural move working rather than merely being argued for. The risk statement in a scheme document of 46 printed sides sat on side 31 in eight point type, and 7 of 30 readers could state the main risk, being 23.3 per cent. The same risk written in 90 words and placed at the top left 24 of 30 able to state it, being 80.0 per cent, a difference of 56.7 percentage points. The document changed, not anybody's view of the reader, so nobody had to be identified, assessed or sorted to produce that difference.

Where is the record thinnest, and why is that where the duty is heaviest?

Complaint behaviour, which asks who complains and who simply leaves without a word, reaches a sentence this section reaches again by a completely different road. Complaining is not free. Complaining takes time, a certain confidence that the complaint will be read by somebody, and a belief that saying something will change anything. Albert Hirschman set out the choice between speaking up and walking away in Exit, Voice, and Loyalty in 1970. The two are not equally easy for everybody.

The three groups described above now line up. The people least likely to mention a difficulty are those whose difficulty is most severe. The people least likely to be observed are everybody, most of the time, since contact is occasional by design. And the people least likely to complain afterwards are those with the least time, the least confidence and the least reason to expect a result. The three descriptions do not pick out three different groups of people; they pick out substantially the same people. The written record is therefore thinnest precisely where the duty is heaviest, and that sentence has now turned out to be the finding twice over.

The invented decision log carries the shape in numbers. Across its eight quarters, 9 of the 60 investors complained in writing and 14 left without saying anything at all. Of the 23 people who were unhappy enough to act, only 9 produced a record, being 39.1 per cent. The other 14 produced none, being 60.9 per cent. Those 60.9 per cent never gave a reason, so the complaints file read on its own shows 39.1 per cent of the dissatisfaction and nothing else.

How does somebody actually put this to work?

Three different people put all this to work in three different ways, so they are taken separately. Devika Rao, the adviser at the invented Palash Advisory Services Private Limited, does not keep a list of people to watch. She writes one conversation that carries the five practices and runs it with everybody. A person having their worst month in a decade then meets the same slower, checked, recorded conversation as a person having an ordinary Tuesday, and nobody has to work out which is which.

Meera Sundaram, deciding on her own account with no committee anywhere, does something adjacent. She writes the rule while things are calm, the only time it can be written honestly. The rule is what meets the difficult month, rather than a fresh decision taken inside it. The invented log records that 20 of the 60 investors adopted a written checklist on 4 November, and across the following four quarters they recorded a written reason on 34 of 41 decisions, being 82.9 per cent, against 19 of 63, being 30.2 per cent, for the other 40. Their realisation ratio fell from 3.2 to 1.6. Sixty people over eight quarters cannot support a claim about returns, and none is made. What the log shows is narrower: a rule written in advance is still there when the month is hard.

The third reader is the one helping somebody else, most often an older relative, and this is where the greatest care is needed about what is being licensed. The useful thing to do is to ask what was explained and what was understood, and to ask for the second conversation. The thing not to do is to take over the deciding. Asking and taking over are easily confused, and they are opposites: asking adds a check to somebody's own decision, and taking over replaces that decision with the helper's.

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.