The Disposition Effect: Selling Winners and Holding Losers

The disposition effect is the habit of realising gains more readily than losses. Across the Palash decision log, 61 of the 108 positions standing in gain were sold, or 56.5 per cent. Only 23 of the 132 standing in loss were sold, or 17.4 per cent. The ratio is 3.2 to 1. The raw share of sales that were winners, 72.6 per cent, is a much weaker number.

The whole thing rests on a division almost nobody makes when looking at their own record. How many of an investor's sales were winners is not, on its own, a fact about the investor. A winner that is not held cannot be sold, so the share is partly a fact about what happened to be held. Picture a fruit seller at the end of a long day being asked what share of what he sold was mangoes. The answer describes the cart before it describes the seller. Learning something about the seller means asking how much of the mango went and how much of the banana went, and comparing the two. The comparison of those two rates is the whole of the disposition effect. Each count of sales is divided by the positionsA position is one holding of one thing, tracked on its own, with its own cost and its own current value. of that kind that were there to be sold, and what survives the division is a fact about the deciding rather than about the cart. Sales divided by opportunity is the only form of the number that carries information about behaviour.

Both rows at one scale. Every position is 2.5 pixels of width. 240 POSITIONS OPEN ACROSS EIGHT QUARTERS 108 IN GAIN 45.0 per cent of the 240 132 IN LOSS 55.0 per cent of the 240 HOW MUCH OF EACH BLOCK WENT 61 SOLD 47 held 23 109 held 56.5 per cent of the gains went 17.4 per cent of the losses went The top row is what was there to sell. The bottom row fills each block from its own left edge, so the coloured fraction of each block is the realisation rate itself. Invented log, illustrative.
The 240 open positions split 108 in gain and 132 in loss, and the coloured part of each lower block is literally the rate at which that kind was sold.

What is the disposition effect, once it is stated as a measurement?

Shefrin and Statman named the pattern in the Journal of Finance in 1985: a disposition to sell winners too early and to ride losers too long. Naming it was the easier half. The harder half was turning it into something a record could confirm or refuse. Count the sales, see how many were profitable, quote the share. The share of sales that were winners moves for reasons that have nothing to do with the person who made the sales.

Odean supplied the repair in Are Investors Reluctant to Realize Their Losses, in the Journal of Finance in 1998, and the repair is a denominator. Take every position standing in gain and ask what fraction of them was sold. Take every position standing in loss and ask the same. Two realisation ratesSales of one kind divided by the number of positions of that kind that were available to be sold. come out, one for gains and one for losses, and the disposition effect is the ratio between them. Stated as a measurement, the disposition effect is the rate at which gains are realised divided by the rate at which losses are realised, and a number above one is the effect. In the Palash decision log that ratio is 3.2 to 1. The ratio is not a claim that anybody sold too early. Each kind was equally available to sell, and 3.2 counts how differently the two were treated.

Four steps. The raw share of winning sales uses only the first. 1 2 3 4 Count the sales of one kind. 61 gains went, and 23 losses. Count what existed to be sold. 108 stood in gain, 132 in loss. Divide the first by the second. 56.5 per cent and 17.4. Compare the two. The ratio is 3.2 to one. RAW SHARE STOPS HERE THE MEASUREMENT NEEDS ALL FOUR
A realisation rate is built in four steps, and the raw share of winning sales is that same sequence abandoned after the first one.

Why is the raw share of winning sales the wrong number?

Because it answers a question about the menu while sounding like a question about the diner. In an examination where nine of every ten questions are arithmetic, nine of every ten questions a candidate gets right will be arithmetic, whether arithmetic is that candidate's strength or that candidate's weakness. The share of the correct answers that were arithmetic is a fact about the paper. To learn anything about the candidate, somebody has to ask what share of the arithmetic was answered correctly and what share of the rest, and set the two side by side.

The raw shareThe proportion of the sales that were winners. The raw share rises and falls with what was available to sell, and describes the holding as much as the person. of winning sales in the Palash decision log is 61 of 84, which is 72.6 per cent, and it has exactly that defect. The 72.6 per cent was computed from a document that records sales and nothing else. Nowhere in a list of sales is there any record of what was sitting there unsold, and without that count there is no way to separate a person who reaches for gains from a person who simply held a great many gains. A share of sales is a fact about the holding and a fact about the deciding jammed together, and no amount of care with the sales list will pull them apart. The repair is not a better sales list. The repair is a second document.

The document that produces 72.6 per cent, and the count it does not hold. SALES EXTRACT, EIGHT QUARTERS sale 001, closed in gain sale 002, closed in gain sale 003, closed in loss sale 004, closed in gain sale 005, closed in gain and onward, to sale 084 61 in gain, 23 in loss so 72.6 per cent winners nothing here counts the unsold 1 2 3 What the sales list cannot show How many positions stood in gain that day, and how many stood in loss. Not in this document. Why the missing count matters 61 of 84 moves with the mix being held, not only with the deciding. Two facts in one number. What repairs it The open positions register: 108 in gain and 132 in loss. Two documents, one measurement. Invented extract, drawn for teaching.
A sales list can produce the 72.6 per cent on its own, but the count of what stood unsold lives in a different document and without it the share means nothing.
Try it out

What is a realisation rate?

Financial Literacy Bootcamp — Fin Maverick

What are the two realisation rates in the Palash decision log?

Now bring in the second document and the arithmetic falls out in one line each. Across the eight quarters the Palash decision log carries 240 open positions. Of those, 108 stood in gain and 132 stood in loss, and 108 plus 132 returns the 240. Of the 84 sales, 61 closed a position that stood in gain and 23 closed one that stood in loss, and 61 plus 23 returns the 84. Every position is therefore in exactly one of four states, and the four states account for all 240.

What kind of positionAvailable to sellSoldNever soldRealisation rate
Standing in gain108614756.5 per cent
Standing in loss1322310917.4 per cent
All positions24084156ratio 3.2 to 1

The two rate cells carry the whole argument. A gain standing in the log had a 56.5 per cent chance of being closed over the eight quarters; a loss had a 17.4 per cent chance. The first divided by the second is 3.2, and that ratio is the disposition effect stated as a measurement. Gains in the Palash decision log were realised at 3.2 times the rate at which losses were realised, and that sentence survives every change in what the 60 investors happened to be holding. Notice what the table also shows. Only 47 gains were carried on unsold against 109 losses, so the holding fills up with losses, not because anything was chosen but because the gains kept leaving and the losses kept staying.

Bars to scale: two pixels of height for every percentage point. 100 75 50 25 0 72.6 56.5 17.4 moves with what was available THE RATIO 3.2 to 1 a ratio, not a percentage the raw share 61 of 84 sales gains realised 61 of 108 open losses realised 23 of 132 open
The raw share of 72.6 per cent and the pair of rates behind it are computed from the same 84 sales, and only the pair holds still when the mix changes.
Try it out

61 of the 84 sales in the Palash decision log were winners. What does 72.6 per cent establish about behaviour?

What does the ratio survive that the raw share does not?

Here is the test that settles it, and it is worth doing slowly. Freeze the behaviour. Say that whoever is deciding will always close 56.5 per cent of the gains standing in front of them and 17.4 per cent of the losses, no matter what is in front of them. Now change nothing about the person and change only the opportunityHow many positions of each kind were there to be sold. The count is the denominator, and it is set by what happened to be held rather than by any decision. they face.

With 50 of the 240 standing in gain and 190 standing in loss, that unchanged behaviour produces 28.2 gain sales against 33.1 loss sales, and the raw share of winning sales comes out at 46.0 per cent. With 200 in gain and 40 in loss, the same unchanged behaviour produces 113.0 against 7.0, and the raw share comes out at 94.2 per cent. The same two rates produce a raw winners share anywhere from 46.0 per cent to 94.2 per cent, and the person has not changed by so much as a decision. The two rates divided instead give 3.2 to 1 at the first setting, 3.2 to 1 at the second and 3.2 to 1 at the log's own split. A measurement is the number that stops moving when the things it is not about are moved.

Three panels, identical in every respect but the mix. Bars at 0.75 pixels per position. 50 IN GAIN, 190 IN LOSS 108 IN GAIN, 132 IN LOSS 200 IN GAIN, 40 IN LOSS the mix available to sell the mix available to sell the mix available to sell RAW WINNERS SHARE RAW WINNERS SHARE RAW WINNERS SHARE 46.0 72.6 94.2 THE RATIO THE RATIO THE RATIO 3.2 to 1 3.2 to 1 3.2 to 1 had the mix been this the log's own split had the mix been this Both realisation rates are held at 61 of 108 and 23 of 132 in all three panels. Only the mix moves. Invented log.
Holding the two realisation rates fixed and changing only the mix available swings the raw share from 46.0 per cent to 94.2 while the ratio sits still at 3.2.
Try it out

Before the control below is moved: can one unchanged pair of realisation rates produce a raw winners share of 46 per cent and also one of 94 per cent?

Play with it

Move the mix and watch one number swing while the other refuses

One variable moves: how many of the 240 open positions stand in gain, from 20 to 220, with the rest standing in loss. Two things are held fixed by stipulation at every setting: gains are realised at 61 of 108, which is 56.5 per cent, and losses at 23 of 132, which is 17.4 per cent. At the log's own split of 108 and 132 the raw share of winning sales is 72.6 per cent. At 50 and 190 it is 46.0 per cent, and at 200 and 40 it is 94.2 per cent. The ratio of the two rates is 3.2 to 1 wherever the slider is set.

20 in gain108 in gain, 132 in loss220 in gain
WHAT IS AVAILABLE TO SELL 108 in gain 132 in loss WHAT GOES, AT THE TWO FIXED RATES 61 gains sold, 56.5 per cent of them 23 losses sold, 17.4 per cent of them RAW SHARE, PER CENT, LEFT SCALE RATIO OF THE TWO RATES, RIGHT SCALE 0 25 50 75 100 0 2 3 1 4 5 RATIO: 3.2 TO 1, ALWAYS 46.0 at 50 94.2 at 200 72.6 20 60 100 140 180 220 positions standing in gain, out of 240
Standing in gain, what moves
108
Raw winners share
72.6
Ratio, held by stipulation
3.2 to 1

With 108 of the 240 positions standing in gain and 132 standing in loss, the two fixed rates close 61 gains and 23 losses, so 72.6 per cent of the sales are winners and the ratio is 3.2 to 1.

Educational illustration. The two realisation rates are held constant by stipulation at every setting of the control, at 56.5 per cent on gains and 17.4 per cent on losses, so nothing about the deciding changes as the control moves. Only the mix available changes, and only the raw share responds. The two hollow rings on the plot mark the settings quoted above, at 50 in gain and at 200. Sale counts are shown to the nearest whole position.
Try it out

The raw share has just travelled most of the scale. What stayed fixed while it moved?

What is Realisation Utility, and what does it add to the account?

Everything so far has been counting. One idea is not enough to explain why the pattern happens. A second idea is needed. Ask why a loser gets kept and the answer is reasonably well settled: the position is being scored against what it cost, and closing it turns a paper lossA loss on a holding that is still held. The money is just as gone as in a realised loss, but it has not been written into the record yet. into a settled one. The reference point handles that much on its own. Now ask the other question. The winner is still doing exactly what it was bought to do, so why is it sold, and often early? A position above its cost is comfortable to hold, so the reference point has nothing much to say about the selling.

Realisation utilitySatisfaction taken in the act of closing a position at a gain, counted separately from the money the closing produces. is the second idea, set out by Barberis and Xiong in the Journal of Financial Economics in 2012. The idea is that a person takes satisfaction from the act of booking a gain itself, separately from what the money then does. Think of a household that has spent three weekends repainting a room. The pleasure is concentrated in putting the brush down and standing back, not in the paint. Realisation utility puts a second term into the account: one for the position and one for the act of closing it, and only the second explains why a gain gets taken early. Both terms are needed, and the log cannot show how much of the 56.5 per cent belongs to which.

Two things are being scored when a position is closed, not one. + THE ACT OF CLOSING A POSITION A burst of satisfaction at the moment a gain is booked, whatever the money then does. THE POSITION ITSELF What the holding is worth, and what the whole of what is held is worth after the sale. The ordinary account, and not enough on its own. WHAT THIS TERM EXPLAINS Why the gain is closed early, while it is still rising. WHAT THIS TERM EXPLAINS How much money is left, which is the same whether the closing felt good or not. Barberis and Xiong set the upper term out in the Journal of Financial Economics in 2012. Illustration, not a measurement.
Realisation utility adds a second term above the ordinary one, and it is the upper term rather than the lower that explains a gain being closed early.
Try it out

What does realisation utility add that the reference point does not supply on its own?

Why does the reference point alone leave half the pattern unexplained?

Because the pattern has two halves and the reference point is built to explain one of them. A holding standing below its cost sits in the region where taking the chance is preferred to settling, and that is a complete account of a loser being kept. Run the same reasoning upward and it says a holding standing above its cost will be closed rather than risked. The direction is right and the timing is missing. The reference point does not explain why the closing happens now, at a modest gain, in a position nothing has gone wrong with. A satisfaction attached to the act of closing is collected the moment the act happens, and waiting postpones it, so realisation utility supplies the missing timing. They answer different questions, and a reader who collapses them into one loses the ability to say which part of a record is being explained by which.

Same two questions put to each account. Identical panels, one difference. THE REFERENCE POINT ALONE WITH REALISATION UTILITY ADDED Why is the loss left open? Why is the loss left open? ANSWERED ANSWERED Why is the gain closed early? Why is the gain closed early? NOT ON ITS OWN ANSWERED One question left standing. Both questions accounted for. Two ideas doing two jobs, kept apart deliberately. Neither is measured by the Palash decision log. Invented illustration.
The reference point on its own answers the question about the loss and leaves the question about the gain standing, which is the gap the second term fills.

How does the pattern differ from booking a loss on purpose?

Disposition Effect vs Tax-Loss Harvesting

The disposition effect and tax-loss harvesting look like opposites, and are usually confused in the other direction. Tax-loss harvestingClosing a position that stands at a loss on purpose, so the loss is recorded and can be set against something for tax. The reason is the tax treatment, not the position. is the deliberate closing of a position standing at a loss so that the loss is recorded. Whatever a reader thinks of it, it is a decision taken for a stated reason, and it produces a record that is the mirror image of the disposition effect: losses realised readily, gains left standing. Measured against opportunity in the same way, its ratio would sit below one rather than above.

The difference matters for reading any log, including a reader's own. The disposition effect and tax-loss harvesting produce opposite records from opposite intentions, and neither of them is visible in a return figure. A person doing one and a person doing the other can finish a period with the same number at the bottom of the statement. Only the two realisation rates, counted separately against what was available, tell them apart. And a ratio below one is not proof of careful tax practice either. The measurement identifies the pattern. The measurement never identifies the reason.

Same four rows put to each. Opposite records, opposite reasons. THE DISPOSITION EFFECT TAX-LOSS HARVESTING THE INTENTION THE INTENTION none stated; the gain itself is the reason deliberate; the tax treatment is the reason WHAT THE RECORD SHOWS WHAT THE RECORD SHOWS gains go readily, losses stay open losses go readily, gains stay standing THE TWO RATES THE TWO RATES 56.5 against 17.4, a ratio above one the ratio would sit below one instead SHOWS UP IN A RETURN FIGURE SHOWS UP IN A RETURN FIGURE no no Two people with opposite records can finish a period with the same figure at the foot of the statement. Only the two realisation rates, counted against what was available, separate them. Invented illustration.
The disposition effect and tax-loss harvesting produce opposite records from opposite intentions, and a return figure alone shows neither of them.
India

Where the tax treatment of a realised loss is settled

How a realised loss may be treated, and what conditions attach, is a matter of the tax law in force and of the disclosure and conduct requirements that apply to a registered intermediary. The Securities and Exchange Board of India at sebi.gov.in is the place to confirm the conduct side, and the tax treatment itself must be confirmed against the law as it stands. Whether realising a loss is worth doing at all turns on a tax position that differs from one person to the next.

Try it out

Somebody deliberately closes a position standing at a loss, wanting the loss recorded. Which pattern is that?

Portfolio Management Bootcamp — Fin Maverick Ratio Analysis That Says Something — free micro-course from Fin Maverick

What did the pattern cost in the case, and what does that figure not prove?

On 12 October Meera Sundaram sold Suvarna Chemicals Limited whole at Rs 4,60,000/-, booking Rs 60,000/- on a cost of Rs 4,00,000/-. The same evening she kept Kesari Logistics Limited, then standing at Rs 1,95,000/- against the Rs 3,00,000/- it had cost, and said she would sell it when it got back to Rs 3,00,000/-. The winner went and the loser stayed.

One evening, two decisions, and the log read back at the far end. SUVARNA CHEMICALS SOLD at Rs 4,60,000/-, booking Rs 60,000/- ROSE 8.0 PER CENT forgone Rs 36,800/- the period between the two readings 12 OCTOBER 31 MARCH FOLLOWING KESARI LOGISTICS KEPT at Rs 1,95,000/-, already down Rs 1,05,000/- FELL 20.0 PER CENT a further Rs 39,000/- gone One pair of decisions in an invented case, and it establishes nothing at all about whether any rule works.
Between 12 October and the following 31 March the sold holding rose 8.0 per cent and the kept holding fell 20.0 per cent, and both movements were read off the same log.

Reading the pair as a verdict on the sale, and what that reading costs

By the following 31 March, Suvarna Chemicals Limited had risen a further 8.0 per cent, so the Rs 4,60,000/- realised would have been Rs 4,96,800/-, and Rs 36,800/- was forgone. Kesari Logistics Limited had fallen a further 20.0 per cent, from Rs 1,95,000/- to Rs 1,56,000/-, losing another Rs 39,000/-. The pair cost Rs 75,800/-. And in the same breath: one pair of decisions establishes nothing whatever about whether a rule works.

The error is treating Rs 75,800/- as proof that selling the winner was wrong and keeping the loser was wrong. The figure is proof of neither. Reverse the two movements, no less possible on 12 October, and the identical pair of decisions looks shrewd. The error costs the ability to tell a finding from an anecdote. A reader who accepts Rs 75,800/- as evidence has accepted a sample of one, and will accept the next one just as readily, including one pointing the other way.

The Palash decision log does support the rate difference: across 240 open positions, gains were realised at 56.5 per cent and losses at 17.4 per cent. A rate difference is a claim about a pattern across many decisions, and it is the only kind of claim 240 positions can carry. Whether any single sale was sound is not measurable from a log at all.

The artefact on the left, the consequence on the right, and the refusal with it. WHAT WAS WRITTEN AGAINST THE DECISION Sold Suvarna Chemicals whole. Kept Kesari Logistics. Will sell it when it gets back to Rs 3,00,000/-. Rs 3,00,000/- is what it cost. Nothing about the position said it. Invented log entry. WHAT THE PAIR COST BY 31 MARCH Rs 36,800/- forgone Rs 39,000/- lost Rs 75,800/- and one pair of decisions proves nothing about whether a rule works. The refusal is part of the finding. What the log supports is the rate difference across 240 positions.
The reason recorded on 12 October points at a purchase cost rather than at the position, and the pair that followed cost Rs 75,800/- without proving anything general.
Try it out

The 12 October pair cost Rs 75,800/- by the following 31 March. What does that figure establish?

Three numbers off one log, and what each is entitled to claim. 72.6 3.2 to 1 Rs 75,800/- THE RAW SHARE OF WINNING SALES Entitled to say nothing at all about behaviour on its own. THE RATIO OF THE TWO REALISATION RATES Entitled to say gains went more readily, across 240 positions. THE COST OF ONE PAIR OF DECISIONS Entitled to say what that pair cost, and nothing general.
Only the middle number is entitled to say anything about behaviour, and the other two are entitled to say nothing general however striking they look.
Selling the winner and keeping the loser costs money. See what the figure proves.

When is selling a winner first exactly the right move?

Often. A rate difference of 3.2 to 1 makes no particular sale wrong. A position that has grown until it takes up far more of a holding than it was ever meant to is being sold because of the position. A holding sold because the reason for buying it has gone is being sold because of the position. Money needed in March is a reason about the money. A sale is the disposition effect when the gain itself is the reason, and not when a winner merely happens to be the thing sold. The test is a question asked before the sale rather than after it, and the honest answer is usually available in about ten seconds: if the position were standing exactly where it was bought, would the sale still be happening today? If yes, the reason is about the position. If the only thing that changed is that it is up, the gain is the reason.

One question, asked before the sale rather than after it. WHY IS THIS POSITION BEING SOLD? answer it before the sale, not afterwards BECAUSE IT IS UP the gain itself is the answer BECAUSE OF THE POSITION it outgrew its intended share THIS IS THE EFFECT the gain is doing the deciding and nothing more follows from it THIS IS MANAGING the position is doing the deciding and a winner gets sold either way The effect is about the gain being the reason. It is not a claim that a winner should never be sold. Invented illustration.
Selling a winner first is correct whenever the reason concerns the position, and it is the disposition effect only when the gain itself is doing the deciding.
Try it out

When is selling a winner first the right move?

Column widths are the counts. Shaded bands are what went. 108 STANDING IN GAIN 132 STANDING IN LOSS 61 SOLD 23 SOLD 47 STILL HELD carried on unsold 109 STILL HELD the holding fills up with these 56.5 per cent of this column went 17.4 per cent of this column went Each column is drawn to its own full height, so the shaded band of a column is that column's realisation rate. Invented log.
The gains kept leaving and the losses kept staying, so what remained unsold at the end was 47 gains against 109 losses without anybody choosing that.

How does somebody reading a decision log actually use two rates?

What an adviser, an analyst or a person deciding alone does with this on a Tuesday

Devika Rao, the adviser at Palash Advisory Services Private Limited, needs no study to run this. She needs two counts she already has. At the end of a quarter she lists every position that stood in gain and counts how many were closed, then does the same for every position that stood in loss. Two divisions, then one more, and the ratio is written down. The count takes an afternoon and needs no record that does not already exist.

The ratio hands her a question rather than an answer. A ratio near one says gains and losses were closed at similar rates. A ratio well above one says something is selecting for gains, and the next step is to read the written reasons, not to conclude anything. A ratio below one is the opposite pattern, and it might be deliberate tax practice or might not.

A person deciding alone, with no adviser and no committee, runs the identical count on the identical two documents. The measurement is cheap, it needs no new paperwork, and it answers something a turnoverHow much was bought and sold over a period, as a share of what is held. Turnover counts volume and says nothing about which positions moved. figure cannot even ask: not how much went, but which ones. What it never hands anybody is a verdict on a particular sale.

The same 84 sales twice. Bars at 5.95 pixels per sale. WHAT AN ACTIVITY COUNT REPORTS 84 SALES WHAT THE TWO RATES READ FROM THE SAME 84 61 CLOSED A GAIN 23 CLOSED A LOSS Both bars are the same 84 sales. Only the lower one can be divided by what was available to sell. Invented log, illustrative throughout.
An activity count reports how much went and stops there, while the same 84 sales split by kind is the only version that can be divided by opportunity.
Bars to scale: forty pixels of height for each whole point of ratio. 0 1 2 3 4 3.2 1.6 the 40 without a written checklist the 20 who adopted one on 4 November WHAT IS NOT CLAIMED HERE No return difference is claimed, measured or implied. Eight quarters and 60 people cannot carry one. What moved is the ratio, and the ratio is the only thing that moved. Written reasons were recorded on 34 of 41 decisions by the 20, against 19 of 63 by the other 40. Invented log.
Twenty of the sixty took up a written checklist on 4 November, and their realisation ratio came out at 1.6 rather than 3.2, with no claim about returns attached.
Three readers, one procedure, and the same limit on all three. DECIDING ALONE DECIDING FOR OTHERS READING A LOG AFTER Both counts are held already. Count what stood open, not only what was sold. The register is kept anyway, so the two rates can be counted per person, per quarter. A ratio can be worked out years afterwards from records nobody kept for the purpose. The limit, for all three: a pattern, never a verdict on any one sale, by anybody. The procedure asks for no new record. Both counts come out of documents that exist already.
Two rates can be counted by a person deciding alone, by an adviser deciding for others, or years afterwards, and the same limit binds all three readings.
Loss aversion and the reference point supply the mechanism sitting underneath the pattern, they were set out by Kahneman and Tversky in Econometrica in 1979, and they are treated under loss aversion and under prospect theory. What trading activity costs is measured under excess trading. Whether Meera Sundaram was right to sell Suvarna Chemicals Limited on 12 October or right to keep Kesari Logistics Limited is not settled here, and no rule follows about when anything ought to be sold by anyone. The tax treatment of a realised loss depends on the tax law in force, and the Securities and Exchange Board of India at sebi.gov.in is where any such requirement is confirmed before it is relied on.

Sources

SourceDocumentSite
Shefrin and Statmanthe paper that named the pattern of realising gains too readily and losses too slowly, Journal of Finance, 1985ssrn.com
OdeanAre Investors Reluctant to Realize Their Losses, Journal of Finance, 1998, which measured the pattern against opportunity rather than against salesssrn.com
Barberis and Xiongthe paper setting out realisation utility as a term attaching to the act of closing a position, Journal of Financial Economics, 2012ssrn.com
Kahneman and Tverskythe 1979 paper setting out prospect theory, Econometrica, named here only to say where the mechanism is treatednber.org
Securities and Exchange Board of Indiaconduct, suitability and disclosure requirements applying to registered intermediaries, and the place to confirm any tax treatment of a realised losssebi.gov.in
Association of Mutual Funds in Indiainvestor-facing practice guidance for distributors and registered advisersamfiindia.com

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.

Covered in this topic

Subtopics

Realisation UtilityDisposition Effect vs Tax-Loss Harvesting