Disposition Effect Investing: The $1,785 Mistake Hiding in a Routine Withdrawal

An investor needs $9,800 out of a taxable brokerage account for a roof repair. Two positions could fund it: one up $7,500, one down $5,200. Almost everyone reaches for the winner. In this case that instinct costs $1,785 in a single tax year — and, according to the largest study ever run on the question, it also picks the worse stock to sell.

This is the disposition effect: the tendency to sell winners too early and hold losers too long. Disposition effect investing behavior is one of the few biases with a precisely measurable price tag, because the tax code puts a number on it. Below is the full case study — the account, the two paths, the research on 10,000 real brokerage accounts, and the five-step process that keeps the bias from touching your sell decisions.

The account: three positions, one uncomfortable decision

Here is the taxable account, simplified to the parts that matter. All positions have been held more than a year, so any gain is long-term.

The investor needs $9,800. Selling roughly half of Position A raises it. Selling all of Position B raises it. Both produce identical cash in the checking account. Only one of them feels good.

Selling Position A means booking a win. The confirmation is immediate: you were right, and here is the proof in dollars. Selling Position B means signing your name to a $5,200 mistake and admitting the thesis didn’t work. That asymmetry — not the math — is what drives the decision for most people.

What disposition effect investing costs in this exact scenario

Run the tax consequences on both paths. Assume a 15% long-term capital gains rate and a 22% ordinary income bracket, which covers a large share of middle-income households.

The gap is $1,785 on identical cash, plus a $2,200 loss carryforward that keeps working in future years. Those mechanics come straight from the IRS: per Topic No. 409, net capital losses in excess of gains can offset up to $3,000 of ordinary income per year ($1,500 if married filing separately), with the remainder carried forward indefinitely.

Worth being precise about what this comparison is and isn’t. It is a worked example with stated assumptions, not a universal result — an investor in the 0% long-term capital gains bracket would see no tax cost from selling the winner at all, and someone with large realized gains elsewhere would apply the loss differently. The general shape holds, though: the disposition effect systematically pushes people toward the more expensive of two identical-cash options.

What 10,000 brokerage accounts revealed

Hersh Shefrin and Meir Statman named the disposition effect in 1985. Terrance Odean measured it. In “Are Investors Reluctant to Realize Their Losses?” (Journal of Finance, 1998), Odean analyzed every trade in 10,000 accounts at a large discount brokerage from 1987 through 1993 — 10,967 realized gains, 36,033 paper gains, 9,476 realized losses, and 51,502 paper losses.

He built two ratios. The Proportion of Gains Realized is realized gains divided by all gains, paper and realized. The Proportion of Losses Realized is the same calculation for losses. If investors ignored whether a position was up or down, the two numbers would match.

Two findings deserve emphasis.

First, the December reversal proves this isn’t ignorance. The same investors who spend eleven months avoiding their losers suddenly realize them in December, when the tax benefit is salient and the deadline is close. They know the rule. They apply it once a year, under pressure, and ignore it the rest of the time.

Second, and more damaging: the stocks they sold went on to beat the stocks they kept. Odean found that winners which were sold outperformed losers which were held by an average of 3.4 percentage points over the following year, measured against the CRSP value-weighted index. The disposition effect is not a defensible tradeoff between taxes and returns. It loses on both.

Why disposition effect investing survives contact with the evidence

The engine underneath is loss aversion — the finding that a loss registers roughly twice as strongly as an equivalent gain. An unrealized loss is, psychologically, still provisional. Selling converts it into a fact, and the mind will pay real money to postpone that conversion. We went deep on the mechanism and how to design around rather than against it in our piece on how loss aversion affects budgeting.

Layered on top is a second bias with its own price tag. Holding the loser is usually justified with “I’ll sell when it gets back to what I paid” — a sentence that treats a cost basis as though it carries information about future returns. It doesn’t. The market has no memory of your purchase price. That is the sunk cost fallacy operating on a brokerage statement, and it converts a bad position into a long one.

There is also a self-image component. Selling the loser closes the case on a stock you researched and chose, which is uncomfortable in proportion to how confident you were going in — a dynamic that compounds with the pattern we described in overconfidence bias and stock picking. Holding keeps the verdict open, and an open verdict feels better than a bad one.

Five steps that take the bias out of the sell decision

1. Decide what to sell before you look at gains and losses. Open the position list sorted by ticker, not by return. Ask one question about each holding: if I had this much cash today, would I buy this position at this price? Anything that fails goes on the sell list. Only after that list exists do you look at cost basis — and then only to sequence the sales for tax efficiency, not to revise the list.

2. Raise cash from the highest-basis lots first. Most brokerages let you specify which tax lots to sell instead of defaulting to first-in-first-out. Selecting specific lots — shares purchased at the highest price — minimizes realized gain on any given sale. This is a settings change and a two-minute habit, and it is free money for anyone selling from a taxable account.

3. Harvest the loss without leaving the market. The objection to selling losers is that you give up the recovery. You don’t have to. Sell the position, book the loss, and immediately buy a similar-but-not-identical fund — a different index tracking the same asset class. The IRS wash sale rule disallows the loss if you buy a “substantially identical” security within 30 days before or after the sale, so the replacement has to be genuinely different. Whether the effort pays off at smaller account sizes is a separate question, and we ran the numbers on that in our analysis of tax loss harvesting for small portfolios.

4. Set rebalancing bands and let them fire. A rule that says “rebalance whenever any asset class drifts more than 5 percentage points from target” removes the decision entirely. Bands sell whatever is overweight, which is mechanically the opposite of the disposition effect and requires no courage.

5. Stop treating December as the tax month. The December reversal in Odean’s data is the tell that people know what to do and just don’t do it on a normal Tuesday. Put a recurring calendar entry on the last business day of each quarter to review unrealized losses. Four low-stakes reviews beat one deadline-driven scramble.

I hit this one personally a few years back, holding a single position well past the point where I would have bought it fresh, entirely because selling meant admitting the pick was wrong. What eventually fixed it wasn’t insight — it was writing the rebalancing bands into a script that flags drift on a schedule and mails me the list. As a software engineer I have a reflex toward automating anything I’ve already gotten wrong once, and behavioral economics gave me a name for why the manual version kept failing. Most of my money now sits in index funds and tax-advantaged accounts precisely because those structures make fewer decisions available to me. No advisor, no stock screens, just fewer opportunities to be the weak link.

Three situations where holding the loser is the right call

A bias having a name doesn’t make every instance of the behavior wrong. There are legitimate reasons to keep a down position, and they are worth naming so the correction doesn’t overshoot into reflexive selling.

The position still belongs in the portfolio. A broad index fund that is down 12% in a drawdown is not a mistake to be corrected. It is the asset class doing what asset classes do. The test in Step 1 — would I buy this today at this price? — returns yes, and a yes means hold regardless of what the cost basis says.

You have no gains to offset and no ordinary income to shelter. The $3,000 annual limit on deducting net capital losses against ordinary income caps the immediate benefit. If you have already used it this year and hold no realized gains to offset, the near-term value of harvesting drops to a carryforward — still worth something, but not urgent enough to justify a trade you would otherwise not make.

The replacement would be worse. Harvesting only works if you can maintain equivalent market exposure with a fund that isn’t substantially identical. In a narrow corner of the market — a specific sector, an illiquid holding, a fund with no close analogue — the tracking difference or the bid-ask spread can eat the tax benefit. Run the comparison before assuming the harvest is free.

What none of these justify is the version most portfolios actually contain: a position held solely because selling would confirm a loss. That distinction — real reason versus retroactive justification — is the whole game, and the fastest way to tell them apart is whether the reason existed before you looked at the cost basis.

Key takeaways

Nothing here is a recommendation about any particular security, and tax outcomes depend on your bracket, your state, and the rest of your return — worth confirming with a tax professional before acting on a large position. But the underlying point survives all of those caveats: the position you least want to sell is frequently the one you should, and the discomfort you feel looking at it is information about your psychology, not about the stock.

Photo by Tech Daily on
Unsplash

Written by Chris Steve

Chris Steve is a software engineer with a deep interest in personal finance, behavioral economics, and AI. He started Money & Planet to share clear, research-backed money guides — the kind that explain the math instead of pushing products. His writing focuses on long-term wealth building, the psychology behind spending and investing decisions, and the practical tools regular people can use to make smarter financial choices.