Knowledge of Loss Aversion Divides by Which Direction It Is Used In: Making Visible, or Manufacturing

Write “as things stand, you are losing out” and that sentence is comparing against something.

Only the counterpart of the comparison goes unwritten; the comparison itself always takes place. You can say something has been lost only where the state before losing it is settled. Where it is not settled, you cannot say whether there is more or less, whether it rose or fell.

So the single sentence “you are losing out” contains the work of placing a counterpart for comparison inside the reader, except where the reader already holds the same baseline. You set up what they supposedly would have had, and present the gap from there as a loss. The setting up completes behind the conclusion, so the reader does not see it. Nor, in most cases, does the writer. They took themselves to be selecting an expression.

Three points get resolved here. First, under what condition the word loss can be defined at all. Second, where the line falls between making visible a loss the reader is already suffering and manufacturing one that did not exist. Third, what happens when the counterpart of the comparison keeps being moved.

Some things will not be settled. How far you may write from the loss side is not decided; the allocation varies with the form of the trade and the length of the relationship. Nor can what the reader actually compared against be observed. The counterpart sits inside them and stays invisible from outside. Methods for making the tendency itself larger or smaller are not covered either, since that side does not move with how you write.

Where this gap becomes damage is on the side of the cautions usually attached. “Don’t overdo the urgency” and “watch the frequency of your limited offers” are both counsels of degree, and neither carries a criterion for where stoking begins. A counsel of degree cannot be carried out without a scale for measuring degree.

And a judgement settled without a scale moves in one direction. In a month when the response fell, towards the stronger side; in a month with room to spare, towards the measured side. The months that fell are the ones that stay in memory, so the average drifts towards strength. The damage of having no scale is not that a wrong judgement comes out, but that the judgement gets settled by circumstance.

Below, the presented account of loss aversion is set down accurately, and from it comes the counterpart of the comparison — the premise of a reference point. On that basis it handles the boundary between making visible a loss already occurring and manufacturing one that was not there.

How the Law of Loss Aversion Is Explained in Circulation

Loss aversion is the tendency by which people choose to avoid a loss more strongly than they choose to obtain a gain of the same size. It is described as a property contained in prospect theory, presented by the psychologists Daniel Kahneman and Amos Tversky (Kahneman & Tversky, 1979, Econometrica, 47(2), 263–291). Terms such as loss aversion bias point to the same thing.

The magnitude gets stated too. For the same amount of money, the pain of losing is felt at roughly twice the pleasure of gaining.

Applications are listed. First, deadlines and quantity limits. Writing “only now” or “three days left” creates a state in which failing to act means losing an opportunity.

Second, restatement. “You will make three hundred thousand yen a year” works less strongly than “you will eliminate three hundred thousand yen of waste a year”, runs the example.

Third, offering a guarantee. Attach a refund guarantee and nothing is lost by purchasing, so the barrier to judgement falls.

Cautions get attached as well: warnings against over-stimulating anxiety, and against running limited offers so often that the baseline drops. English-language treatments add the same conditions — do not create false scarcity, make sure the limitation is genuine.

All of it is accurate as description. The tendency has been confirmed in experiments, the effect of restatement is observed, and the cautions point in the right direction.

Draw out what the set takes for granted. Loss aversion is organised as a choice of expression. The same content, written from the gaining side or from the losing side. A matter of selecting an expression.

Sometimes it really is a matter of expression. Where an already-occurring fact is described from one side or the other, this is purely expression.

The premise fails where the reader’s state differs before and after the writing. Once a counterpart for comparison is newly set up, the reader holds something they did not hold before. Rather than changing the expression, you have added a piece of material for judgement. And the addition leaves no record. To the writer it was a choice of expression, so on later review it gets treated as one.

The three applications listed above also differ in kind on this point. Restatement is an operation on the expression; setting a deadline is an operation on the conditions. The first states the same fact from another face; the second creates the fact anew. The guarantee goes the other way entirely, since it reduces risk on the reader’s side. Lined up as three “applications”, the difference is invisible. The sections that follow open it out one at a time.

Loss Can Only Be Defined as a Deviation From a Reference Point

At the centre of prospect theory is the property that value is evaluated not as an absolute quantity but as a change from some baseline. That baseline is called a reference point.

An income of six million yen is neither good nor bad in itself. For someone who was at seven million it is a loss; for someone who was at five million it is a gain. The same figure carries opposite signs depending on the baseline.

To use the word loss, therefore, a baseline must be settled. Where the baseline is not settled, what counts as a loss is not settled either.

This is where practice drops it. Writing “you are losing out”, the writer holds a baseline. There is no guarantee the reader holds the same one.

Where they do not, that sentence conveys no information. What gets across is a sensation — something is apparently being lost — and not what is being lost against which baseline.

The reader supplies the baseline themselves, so it does not matter — that view is available. Supplying does occur. What is supplied, though, differs by reader. And most supplied baselines are dragged by the figures and states presented in that context. Put out a figure and that figure becomes the baseline. It looks as though the reader supplied it; the material was handed to them by the writer.

Sometimes the baseline is widely shared. A contracted price plan, an amount already paid — baselines settled as fact. Here loss can be defined objectively, and the figure can be checked afterwards.

Beyond that, the baseline does not settle to one. For the same person, the same state becomes a loss or a gain depending on what it is compared against. Against last month, against the assumption held at the point of going independent, against the average in the trade.

Which of these gets brought out is settled, in most cases, by the writer. It looks as though the reader selected it, and the options were presented by you. Selecting the baseline settles the outcome with roughly the same force as selecting the conclusion.

State the baseline explicitly and the size of the loss becomes verifiable. Once verifiable, it can turn out smaller than what was presented. A loss left vague is estimated at its maximum inside the reader.

The shrinkage looks unfavourable to the writer. In practice, the figure that survives verification is the one the next piece of writing rests on. A shrunken figure is also a figure you never have to withdraw later.

Manufacturing a Loss Therefore Means Placing a Reference Point First

To make the sentence “buy now or lose out” hold, the state of having bought has to be set up as the baseline. Having set it up, you present the state of not buying as the gap from there.

The operation replaces the baseline the reader originally held. For the reader, the state of not having bought that product was itself the baseline. What was the baseline turns into a state with a gap in it.

At the instant it turns, the loss occurs. Nothing has happened in reality. What moved is the position of the baseline.

And the operation is invisible to the reader. What is visible is the conclusion — that they are losing out — while the placing of the baseline never rises into awareness.

Announcing the product at all moves the baseline — that objection stands. It does move. Learning anything always changes the set of options. The question is not whether it moves, but whether the movement gets presented as a loss.

“This exists” adds an option. “You lose out by not using this” reseats the added option as the baseline. Two presentations issue from the same information. That there are two is itself evidence that one of them was selected.

Sometimes a deadline genuinely exists. A course with a fixed number of seats, an application with a closing date. Here the deadline is a fact, not a baseline that was set up. A factual deadline has to be written, or the reader misjudges.

Setting up admits of degree. “You lose out by not using this” is a strong setting; “with this, that task is no longer needed” is a weak one. The second contains a comparison too, but stops short of replacing the baseline.

That there is a range means there is a choice. For something you can choose, there is no need to adopt the strongest form unreflectively.

And the weaker setting carries more information. “That task is no longer needed” states what changes and how. “You lose out” hands over an evaluation without stating the content of the change. The difference in information appears as a difference in how much the reader can judge for themselves.

A factual deadline and a set-up deadline are indistinguishable from the reader’s side, since both are presented in the same form. Where the state persists, readers come to discount both alike. Factual deadlines stop working too, and by then the announcement for the course whose seats really do fill is not getting through either.

Separate Making Visible From Manufacturing — Was the Reader’s Loss Occurring Before Reading

The question that can be decided is one.

“Before reading this piece, was the reader suffering that loss?”

If they were, the piece is making an existing state visible. If they were not, the piece is manufacturing the loss.

“Keep publishing, and if the contact details do not remain in your own hands, you have no means of reaching people when you need to.” What this sentence points at holds before it is read. Read or unread, the means does not exist. The writing has only made it visible.

Another. “Apply by the end of today or you will never get this price again.” The loss this sentence points at did not exist before the sentence was issued. The condition on the price was set by the writer.

Both sentences engage loss aversion. As reader responses, the same kind of thing occurs. What differs is where the loss came from.

If it comes to the same thing for the reader, does the distinction matter? Extend the time horizon and it does. A loss made visible can be checked by the reader. Check it, find the state was as described, and the writer’s account has proved accurate.

A manufactured loss cannot be checked. Try to check it and you discover the reference point was placed by the writer. From that moment the writer’s accounts in general get discounted. It is not a problem confined to one announcement: a coefficient is applied to everything after it.

Manufacturing is not always dishonest. Offering something under genuinely limited conditions is legitimate as a business. What parts is the case of feigning a limit that is not there, and the case of designing the limit as the principal reason for the decision. The second, even without feigning, has moved the centre of the reader’s judgement onto the conditions.

The question has borderline cases. The reader is suffering a loss but did not know its size. Is showing the size making visible, or manufacturing?

The determination can be made by whether the size shown is verifiable. If the reader can confirm it against their own records, it is making visible. If you are presenting an estimate there is no way to confirm, it approaches manufacturing. Write “you are losing two hundred hours a year” and the parting line is whether the reader can trace where the two hundred came from.

Making visible is the harder of the two to write. Describing an existing loss accurately requires understanding its structure. Manufacturing can be written without understanding any structure. The difference in difficulty also bears on which gets adopted.

Framing Is Not a Neutral Restatement

Tversky and Kahneman showed that logically identical options produce reversed preferences depending on the frame of the description (Tversky & Kahneman, 1981, Science, 211(4481), 453–458). Describe the same programme by the number of people saved and by the number who die, and the majority choice reverses.

This experiment measured preferences over hypothetical policy choices, not responses to product advertising. What it indicates is a direction: the frame of a description enters into the judgement.

The understanding that restatement is “merely rewording” is therefore inaccurate. Restatement changes the material for judgement.

There are two practical consequences. First, at the moment you select which frame to write in, you have intervened in the reader’s judgement, and that calls for awareness.

Second, where both frames are writable, both can be shown. “Three hundred thousand yen of profit” and “three hundred thousand yen of waste removed” are the same thing. Write both and the frame effect cancels.

Writing both makes the text longer — that is a real drawback. In practice most cases do not require both. What is required is being able to explain to yourself why you selected the frame you did. If you can explain it, that is a judgement rather than an operation aimed at an effect.

Writing without selecting a frame is impossible, since no neutral description exists that uses none. There is no option of not selecting; there is only whether you are aware of having selected.

The selection also has a range. Frame restatement holds only for logically identical content. “Profit accrues” and “waste is removed” are the same content only where the amount is the same.

Line up things that are not the same in the form of a restatement and this is no longer a selection of frame but the presentation of a different claim. The two mix easily, so checking the amounts and conditions against each other after writing brings it out. Even in the shape of a restatement, if the content has moved, it is not a restatement.

Keep selecting the frame with the larger effect and the writer’s own grasp of the content thins. Being able to write both frames means holding the content from both sides. Write only one side and the grasp becomes one-sided. Some years of that and a request to explain it from the profit side leaves you stuck.

Keep Moving the Reference Point and the Reference Point Itself Falls

Run limited offers frequently and the reader’s baseline moves to match. From a state where the standard price was the baseline, to one where the discounted price is.

After the move, the standard price feels like a loss. Buying when there is no discount becomes an act of losing out.

The movement is hard to reverse. Raising a baseline once lowered requires holding the lowered state for a long time, and the response falls throughout.

The same happens with the strength of stimulus. For a reader accustomed to strong terms, the same strength stops arriving. The next one gets written stronger, and the baseline moves again.

Each stage of the movement is small. So each looks, at the time, like a sound judgement. Arriving somewhere you did not want to be as the result of accumulated sound judgements is the property of this structure.

If the readers turn over, the baseline returns — that thought is available. For new readers it does return. Published writing stays up, though. A new reader who sees the past announcements builds a baseline from them.

And the baseline of the list as a whole is settled by the relation between the rate at which existing readers leave and the rate at which new ones enter. Premise the work on turnover and you commit to collecting new readers permanently. The cost of collecting them permanently is a cost that would not have arisen had the baseline not been moved.

Where a transaction completes in one pass, the consequence does not bite. You do not repeat with the same person. It bites where the relationship is premised on continuing.

The speed of the movement depends on frequency. A limited offer once a year and one every month move it by different amounts. Once a year, and it is read with the previous occasion faded.

And whether the movement is happening can be measured. Line up the responses to announcements issued under the same conditions in time order. If the response is falling under the same conditions, the baseline has moved.

The measurement requires the conditions to match. Issuing an announcement identical in content, price and period twice with an interval is rare in practice. What can be used instead is lining up responses to announcements of the same format and reading the trend.

The movement of the baseline happens on the writer’s side too. Once writing more strongly than last time becomes routine, not writing strongly feels like slacking. The standard of judgement moves just as it does on the reader’s side.

The writer’s movement is harder to detect than the reader’s, because the reader’s response leaves figures behind and your own sensation does not. What remains is only the writing, so the only way to check is to line the writing up. Put the announcement from a year ago next to the most recent one, without matching conditions. If the strength of the language has risen, the movement has happened.

A Loss Already Present Is Large Enough Without Stoking

In most cases the loss the reader is already suffering is larger than the loss you could manufacture. There is a structural reason.

A manufacturable loss is bounded by the scope of the product. What is lost by not buying is only what that product would have delivered. The ceiling is set by the value of the product, so it cannot be made larger.

A loss already occurring comes from structure. A state where income is proportional to hours worked, where contact details are not held by you, where the baseline for price is settled by someone else. The gap these generate is larger than the utility of any one product.

And since it is already occurring, the evidence is in the reader’s hands. Count how many unanticipated requests you took on last month and it can be confirmed. An account that can be confirmed does not need stoking.

Structural explanation is difficult and does not get read — the worry is fair. Written badly, it does not get read. The structure itself is not difficult; the difficulty is in the writing. And a structural explanation, once understood, gets used in every judgement afterwards. It has a longer service life than a single response.

There are stages where the reader does not recognise their own state as a problem. Here structural explanation does not arrive. What is needed first is putting the symptom into words. Only once the symptom is recognised as their own does the structural account become material. Reverse the order and a correct explanation passes straight through.

In practical terms, the writing has to be put into countable form. Not “you are losing time” but “how many hours in total did you spend on quotations last month”. Countable form lets the reader confirm it on their side.

A loss that can be confirmed gets across without emphasis from you. Emphasis is required where the loss presented cannot be confirmed.

The quantity of emphasis stands in inverse relation to the certainty of the account. What can be confirmed only has to be stated. What cannot be confirmed is not enough when merely stated, so emphasis gets added. The relation also serves as a gauge when reading other people’s writing.

Explain the structure and the reader sometimes solves it themselves. They do not buy, and the state improves on the explanation alone. This looks like a loss, and that reader has confirmed that your account was accurate. A confirmed account is trusted in the next judgement.

One constraint belongs here honestly. Confirmation takes time. A manufactured loss produces a response the same day; a loss made visible does nothing until the reader looks at their own records. Whether they look is not yours to settle either. Put the two side by side and the manufactured side tends to look like the better result. At the moment of comparison, one of them has not produced its result yet.

A Guarantee Is the Inverse of Loss Aversion, and Differs in Property

The refund guarantee is one form that uses loss aversion without moving the reference point.

A guarantee is also designed on loss aversion, but points the other way. Rather than presenting a loss, it removes the possibility of a loss arising from the purchase.

Where the guarantee parts from the rest is in how the reference point is treated. A guarantee does not move the reader’s reference point. With “not buying” left as the baseline, it erases the downside of having bought.

Since nothing was moved, nothing gets discounted afterwards either. If the guarantee functions, the account has proved accurate.

There is a condition: the guarantee has to be exercisable. Where the procedure is cumbersome and the design ensures it is never exercised in practice, this is not a guarantee but the appearance of one.

A guarantee in appearance does move the reference point. A reader who bought believing it could be exercised suffers a loss, measured against their original expectation, at the moment they find it cannot.

Attach a guarantee and more people ask for refunds — that is so. They do. And the increase consists of people who did not need the product in the first place. The state in which someone who did not need it buys costs the business more. A refund procedure ends in one pass; a relationship with an unsuited party continues.

There are products for which a guarantee does not hold. Things that cannot be returned once delivered, things that supply time itself. Here what can be placed instead is material for judgement in advance. Writing the conditions under which it does not suit performs substantially the same work.

A guarantee has a second working. It is a declaration that the supplier is staking a cost on the accuracy of their account. Where the account diverges from reality, the cost arises in the form of refunds.

Because of that structure, a guarantee disciplines the account. Write with exaggeration and refunds rise in proportion. The moment a guarantee is set, the range of what can be written narrows automatically.

That the discipline is applied from outside is what separates this from the other items. Rather than taking care yourself, the design settles the range. Discipline that depends on taking care does not operate on a tired day.

Attaching a guarantee can also lower the burden of judgement too far. A state in which something can be bought without much thought falls outside the objective of handing over the judgement. Set the purpose of the guarantee not as lowering the barrier but as underwriting the accuracy of your account, and this side effect shrinks. How the purpose is set goes on to settle the term of the guarantee and how its conditions are written.

Apply It Sentence by Sentence: Which Is This One Doing

Any sentence touching on loss can be judged one at a time. There is no need to view the piece as a whole.

Ask whether that loss held before the piece was read. If it held, this is making visible. If it did not, that sentence is manufacturing it. The determination lands at the level of a single sentence.

Next, try rewriting it with the reference point stated. Put “compared with what” into the sentence. If it cannot be put in, you were presenting a loss with the reference point left vague. If it goes in and the size of the loss does not shrink, the account is accurate.

Finally, back the deadline or quantity with your own records. If it can be backed, this is the presentation of a fact. If it cannot, it is a condition you set. Issuing a condition you set is legitimate in itself, but whether you know it to be one changes the next judgement.

None of these three looks at the size of the response. The manufactured loss usually draws the larger one.

How the reader actually received it will not emerge from these three. The reference point sits inside them and cannot be observed from outside.

One indirect route exists: follow how the people who responded moved afterwards. People moved by a manufactured loss do not sustain engagement with what was delivered. In figures it appears as the gap between the initial response and the engagement that follows. Where announcements keep showing a large gap, the reference-point setting is doing too much work.

The permissible frequency of deadlines, the term of a guarantee, the proportion that may be written from the loss side — no figures are offered for these. They vary with the form of the trade and the length of the relationship. A rule of thumb would decide it for you, with those conditions left unexamined.

Left in hand instead is one question. What is that sentence comparing against? If the counterpart of the comparison is in the reader’s hands, the sentence is showing something that exists. If you placed the counterpart, the sentence is manufacturing it. That single question settles which side many of your sentences fall on. What remains is only the case where the reader did not know the size of the loss they were already carrying.

Placing a reference point is not peculiar to loss aversion. For the other mechanisms in which the same shape operates, see psychological triggers before the list; for the form that places the counterpart on the side of other people’s behaviour, see where social proof applies. Writing that hands over a reason to move without presenting a loss is taken up by the principles of sales copy. Where to place a reference point is work that designs what happens inside the reader. How the pieces handling that side are arranged is shown in the design of verbalisation.