Cognitive biases in investing: what the research actually shows

A cognitive bias is a systematic error in judgment: not a random slip, but a mistake that many people make in the same direction under the same conditions. Investing is full of the conditions that produce them. Outcomes are uncertain, feedback is noisy, prices move every second, and the same decision can make or lose money for reasons that have nothing to do with its quality.

Behavioral finance is the field that studies these errors and their effect on markets. It grew out of psychology experiments in the 1970s, moved into economics in the 1980s, and was tested on millions of real brokerage trades in the 1990s and 2000s. Two of its founders won the Nobel Prize in economics: Daniel Kahneman in 2002 and Richard Thaler in 2017.

This pillar guide is the map. It explains where the research came from, what prospect theory says, and what the best-known studies of real investors found, with the numbers. Each major bias then gets its own deeper article in this group. For a shorter, trader-focused version with practical habits, see Trading psychology: loss aversion, overconfidence and other biases.

Where the research came from

For most of the 20th century, economics assumed a rational decision maker: someone who weighs every option by its expected value, updates beliefs correctly when new information arrives, and is not swayed by how a choice is described. That model is elegant, and it still underpins much of finance, including the efficient market hypothesis.

In 1974 the psychologists Amos Tversky and Daniel Kahneman published a paper in Science that described three mental shortcuts, or heuristics, that people use to judge probabilities: representativeness (judging by resemblance to a stereotype), availability (judging by how easily examples come to mind) and anchoring and adjustment (starting from a number and moving away from it too little). The shortcuts work well enough most of the time. The paper's point was that they also fail in predictable ways.

Five years later, in Econometrica, the same pair proposed prospect theory, a model of how people actually choose between risky options. It became one of the most cited papers in economics. In the 1980s Richard Thaler and others carried the ideas into finance, and in the 1990s researchers gained access to large sets of brokerage account records. That let them test the laboratory findings against what real investors did with real money, and most of the laboratory findings held up.

Prospect theory in four ideas

Prospect theory is the backbone of behavioral finance, so it is worth understanding properly. It makes four claims about how people evaluate risky choices.

1. Reference dependence. People evaluate outcomes as gains or losses relative to a reference point, not as levels of total wealth. For investors the reference point is often the purchase price, sometimes a recent high, sometimes the value on the last statement. A portfolio worth $95,000 feels like a loss if it was worth $100,000 last month and like a win if it was worth $90,000.

2. Loss aversion. Losses weigh more than gains of the same size. In their 1992 follow-up, Tversky and Kahneman estimated the median loss-aversion coefficient in their experiments at 2.25. The detail and the evidence are in Loss aversion: why losses hurt twice as much as gains please.

3. Diminishing sensitivity. The first $100 of a gain or loss matters more than the next $100. The value curve is concave for gains, which makes people risk-averse when choosing between gains, and convex for losses, which makes them risk-seeking when choosing between losses. Facing a sure loss, many people would rather gamble on breaking even.

4. Probability weighting. People do not treat stated probabilities at face value. Small probabilities get too much weight and moderate-to-high ones too little.

The figure draws the weighting function from the 1992 paper's published formula and parameters. With those values, a 5% chance of a gain carries the decision weight of about 13%, and a 1% chance carries about 5.5%. At the other end, a 90% chance is weighted like about 71%. The curve crosses the diagonal at roughly one-third.

Two investing patterns follow almost directly. Overweighting small probabilities makes lottery-like assets attractive: penny stocks, far out-of-the-money options and the hope that a small holding becomes the next hundred-bagger. Underweighting high probabilities makes a very likely but uncertain gain feel less valuable than a certain one, which feeds the urge to lock in profits early.

The biases that matter most, and the evidence

The figure below groups the best-documented biases by the stage of an investment decision where they tend to act. The groupings are a reading aid; many biases act at several stages.

The disposition effect

Hersh Shefrin and Meir Statman coined the term disposition effect in 1985 for the tendency to sell winners too early and hold losers too long. They explained it with prospect theory (a gain is in the risk-averse zone, so locking it in feels good; a loss is in the risk-seeking zone, so holding on for a rebound feels better than accepting it), plus mental accounting, regret aversion and weak self-control.

Terrance Odean tested it in 1998 with the records of 10,000 accounts at a US discount broker from 1987 to 1993. On days when an investor sold anything, he compared the proportion of gains realized with the proportion of losses realized. Across the year, investors realized 14.8% of the gains available to them and 9.8% of the losses. Only in December, when selling losers saves taxes, did the pattern flip.

The habit was costly. Odean found that the winners investors sold went on to outperform the losers they kept, by about 3.4 percentage points over the following year. Holding losers also defers a tax benefit: in a taxable account a realized loss can offset gains (see Tax-loss harvesting: how it works, the wash-sale rule and its limits), while realizing a gain early brings the tax bill forward.

Overconfidence and overtrading

Overconfidence has two forms. Miscalibration is believing your estimates are more precise than they are. The better-than-average effect is believing your skill is above the median; in a well-known 1981 survey by Ola Svenson, 93% of the American students sampled placed themselves in the top half of drivers for skill.

In markets, overconfidence predicts trading. Brad Barber and Terrance Odean studied 66,465 US households at a large discount broker from 1991 to 1996. The average household earned 16.4% a year after costs, against 17.9% for the market. The fifth of households that traded most earned 11.4%. Their stock picks before costs were not much worse than anyone else's; commissions and bid-ask spreads on frequent trades made the difference. A 2001 follow-up found that men traded about 45% more than women, and that trading cut men's net returns by 2.65 percentage points a year against 1.72 for women, a pattern the authors attributed to greater overconfidence among men. Overconfidence and overtrading: the research on trading too much covers the mechanisms and the later evidence.

Herding and attention

Investors buy what grabs their attention: stocks in the news, with unusual volume or with big one-day moves. Barber and Odean showed in 2008 that individual investors were heavy net buyers of such attention-grabbing stocks, because there are thousands of stocks to buy but only a handful an individual already owns and might sell. Herding adds a social layer: when others are buying, their behavior looks like information. Sometimes it is. Often it is the same information counted many times. Herd behavior and FOMO in investing: cascades, manias and panics covers informational cascades, the fear of missing out and the 2021 meme-stock episode.

Anchoring, representativeness and extrapolation

Anchoring is relying too much on an initial number. In Tversky and Kahneman's 1974 experiment, participants watched a wheel of fortune, rigged to stop at 10 or 65, before estimating the percentage of African countries in the United Nations. The median estimates were 25 for those who saw 10 and 45 for those who saw 65, from a number they knew was random. In investing, purchase prices, past highs and analyst targets act as anchors; see Anchoring and confirmation bias: how investors fool themselves.

Representativeness is judging probability by resemblance. A company with several years of strong growth "looks like" a great company, so investors extrapolate the past too far. Werner De Bondt and Richard Thaler tested this in 1985 by forming portfolios of the biggest prior losers and winners over three-year periods. Over the next three years the losers outperformed the winners by about 25%, which the authors read as overreaction followed by correction. Later research has debated how much of that effect reflects risk rather than bias, which is a useful reminder that market-level evidence is rarely clean.

Confirmation, hindsight and self-attribution

Once an investor holds a view, confirmation bias steers attention toward evidence that supports it. After the outcome, hindsight bias makes the result feel more predictable than it was ("the crash was obvious"), and self-attribution credits wins to skill and losses to bad luck. The three together are why many investors learn less from experience than they think they do: the record that should correct them is edited by the same biases.

Framing, mental accounting and the status quo

The same choice can be framed as a gain or a loss, and the frame changes the decision. Mental accounting, Thaler's term, describes how people sort money into separate "accounts" and treat each by different rules: "house money" from a winning trade is spent more freely, and each position is judged on its own rather than as part of the portfolio. The status quo bias and the related endowment effect make people value what they already hold more than what they could hold instead, which slows needed changes such as Rebalancing a portfolio: calendar vs threshold, drift and taxes.

How the biases fit together: a worked example

Biases rarely act alone. Consider an illustrative investor, Dana, who buys 100 shares of a software stock at $80 after reading three bullish articles in a week (availability and attention). She expects about a 30% gain within a year and feels sure of it (miscalibration). The stock climbs to $96, and she sells 50 shares to "lock in" an $800 gain (the disposition effect, driven by diminishing sensitivity to gains).

The stock then falls to $60 on weak guidance. Dana's reference point is still $80, so the remaining position shows a $1,000 loss. Selling would make it "real" (loss aversion), the company's past growth still looks representative of its future (representativeness), and she mainly reads the bullish takes (confirmation bias). She holds and adds 50 shares at $60 to lower her average cost. If the stock recovers, she will credit her conviction (self-attribution). If it does not, the story will be that nobody could have known (hindsight).

Nothing in that sequence required ignorance. Each step felt reasonable at the time. That is what makes the research useful: it gives names to steps that feel like judgment.

Does behavioral finance mean markets are inefficient?

Not necessarily, and the distinction matters. Individual biases can exist without creating easy profits for anyone else, because betting against mispricing is risky and costly. A stock that is overvalued because of investor enthusiasm can become more overvalued before it corrects, as many professional short sellers learned during the The dot-com bubble: the 1995–2002 boom and the 78% Nasdaq fall. Economists call this the limits to arbitrage.

So the strongest, least disputed findings of behavioral finance are about investors' own results rather than market-beating strategies. Trading more tends to earn less. Selling winners and holding losers costs money and taxes. Chasing recent performance means buying after prices have risen. Those findings are useful to anyone, whatever they believe about market efficiency. For the evidence on how hard it is to beat a benchmark, see Index investing: passive vs. active funds and what the data shows.

What helps: designing around biases

Willpower is a weak defense against a bias, because the bias works by making the wrong choice feel right. The defenses that have the best support are structural: they change the situation rather than the person.

Three general habits cover much of the table. Rules made in advance, in a written plan (see Trading plan and trading journal: what to write down), move decisions to a calm moment. Diversification and Dollar-cost averaging vs lump sum: how DCA works and what data show shrink the cost of any single mistake and reduce the number of decisions you have to make. Honest records, such as a journal of what you expected and why, counter hindsight and self-attribution because they cannot be edited later.

How Finabase helps

Finabase is a research app. It does not place trades or give advice, but several features support the defenses above. Portfolios record each holding's cost basis and gain at live prices, so you see the whole portfolio rather than the few positions you remember. Alerts notify you when a price, a percent move or an indicator level is reached, which removes the need to watch screens and reduces impulse trades. Each symbol page has a Notes tab for a written thesis (for example SPY's notes), and a Valuation tab that shows a company's multiples over up to ten fiscal years, a useful check against anchoring on a single price.

Frequently asked questions

What are the most common cognitive biases in investing?

The best-documented are loss aversion, the disposition effect, overconfidence, herding, anchoring and confirmation bias. Loss aversion and the disposition effect have the strongest evidence from real account data, and overconfidence is the bias most clearly linked to lower returns through overtrading.

What is prospect theory in simple terms?

Prospect theory says people judge outcomes as gains or losses from a reference point, feel losses about twice as strongly as gains, become less sensitive as amounts grow, and give small probabilities too much weight. Kahneman and Tversky proposed it in 1979 and refined it in 1992 as cumulative prospect theory.

Who founded behavioral finance?

No single person did. Its foundations are the psychology research of Daniel Kahneman and Amos Tversky in the 1970s, and its move into economics and finance owes most to Richard Thaler, along with researchers such as Robert Shiller, Hersh Shefrin and Meir Statman. Kahneman (2002) and Thaler (2017) received the Nobel Prize in economics; Tversky died in 1996, and the prize is not awarded posthumously.

Can you overcome cognitive biases?

Rarely by effort alone. The research suggests that changing the decision environment works better than trying to think harder: writing rules before acting, trading less often, holding diversified funds and keeping records that make it hard to rewrite history.

Summary

Cognitive biases are systematic, predictable errors, and investing triggers many of them at once. Prospect theory explains a large share: people judge outcomes from a reference point, feel losses about 2.25 times as much as gains and misjudge probabilities. Studies of real accounts show the costs. Investors realize gains far more readily than losses, sell winners that go on to beat the losers they keep, and the most active traders earn several percentage points a year less than the market. The best defenses are structural rather than mental: plans written in advance, diversification, fewer decisions and honest records. The articles on Loss aversion: why losses hurt twice as much as gains please, Herd behavior and FOMO in investing: cascades, manias and panics, Overconfidence and overtrading: the research on trading too much and Anchoring and confirmation bias: how investors fool themselves go deeper on each.