Mental accounting

An investor keeps £15,000 in a cash ISA labelled “house deposit”. It earns far less than the interest on a credit card balance she is carrying. In a separate stocks and shares ISA she holds a few thousand pounds she thinks of as “play money”. That pot has doubled, so she adds a speculative position to it. She moves no money between the accounts, because they feel like different kinds of money.

Nothing in her bank statements treats the pots differently. Pounds are pounds. But the labels change what she is willing to risk, sell and spend.

Mental accounting is the set of habits people use to sort, track and evaluate money, usually by source, purpose or time period, rather than treating it as one interchangeable pool.

What the research shows

Richard Thaler, who won the 2017 economics Nobel for his contributions to behavioural economics, introduced the term. Thaler (1999) describes mental accounting as the cognitive operations households use to organise, evaluate and keep track of financial activity. He identifies three components: how outcomes are perceived and evaluated, how money is assigned to categories and budgets, and how often accounts are evaluated. Each component breaks the economic principle of fungibility, the idea that a pound in one account is a perfect substitute for a pound in another.

Mental accounting sits within prospect theory, so that needs a quick definition. Kahneman and Tversky (1979) showed that people judge outcomes as gains or losses relative to a reference point, not as changes in total wealth. They also showed that losses weigh more heavily than equivalent gains. A mental account is, in effect, a reference point with a label on it.

The best-known investing-related evidence concerns how earlier results change risk-taking. Thaler and Johnson (1990) reported a house money effect, meaning more risk-seeking after a prior gain. They also reported break-even effects, where after a loss an option offering a chance to get back to even looks especially attractive. According to the abstract, the paper includes real-money experiments alongside its other choice evidence. These are laboratory studies with small stakes, not portfolio data. Their explanation is that a gain is booked separately, so losing it feels cheaper, while a loss stays open until it is recovered.

The evidence is not uniform. Peng, Miao and Xiao (2013) ran two-stage gambles with 915 university students in China, with stakes in yuan. They found only a moderate house money effect, and their participants were less risk seeking than Thaler and Johnson’s. They speculated that people may be more cautious when they can lose all of their prior gains than when they can lose only part of them. They also noted that their design was a simulated gamble, not a real-world decision. Their third study found that gambling winnings and normal income appeared to open different mental accounts, which participants spent quite differently. The effect looks real but sensitive to how the task is set up.

Shefrin and Statman (1985) linked mental accounting to the disposition effect, the tendency to sell winners early and hold losers. Their framework combines mental accounting with regret aversion, self-control and taxes. They argued that tax considerations alone cannot explain the pattern. A share bought at £10 and now worth £7 is an open account in the red. Selling closes it and makes the loss official.

Hartzmark and Solomon (2019) looked at dividends. They found that many individual investors, funds and institutions trade as if dividends and price changes were unrelated, without fully allowing for the fact that share prices fall when dividends are paid. They also report that this behaviour aggregates into predictable price effects around large dividend payments. The evidence here is statistical. It shows patterns in trading and prices, not what any one investor is thinking.

How it shows up in portfolios

Treating dividends as free income

A share price tends to fall by roughly the dividend when it goes ex-dividend, so a dividend moves value from one pocket to another rather than adding return. Investors who file dividends under “income” and price moves under “capital” may spend one and protect the other. Hartzmark and Solomon (2019) found this separation in the trading of professional funds as well as private investors.

Holding losers until they recover

The disposition effect is mental accounting applied to individual positions. Each holding gets its own account, judged by whether it sits above or below the purchase price. A holding may be kept because selling would “lock in” a loss, even though the loss exists whether or not it is recorded.

Playing with house money

After a good run, an investor may treat the gains as a separate stake and take risks with them that the original capital would not have been exposed to. Thaler and Johnson (1990) documented the underlying pattern in experiments. How strongly it carries over to real portfolios is less certain, given the mixed results in Peng, Miao and Xiao (2013).

Rigid buckets across a balance sheet

Holding a cash ISA earmarked for a goal while carrying higher-rate debt is a common example. Each pot is judged against its own purpose, not against the household’s net position. The separation can also reflect a sensible wish to keep money out of reach, so the cost depends on the size of the gap between the savings rate and the borrowing rate.

Checking too often

Thaler (1999) lists evaluation frequency as part of mental accounting. Benartzi and Thaler (1995) built on this. They argued that investors who are loss averse and who also evaluate their portfolios frequently will find equities less attractive than long-term returns would justify. In their calculations, a loss-averse investor who evaluates annually fits the historical equity premium. The more often an account is balanced, the more often a loss is seen.

Why it happens

Categorising money is a cognitive shortcut. Real finances are complicated, and budgets, labelled pots and “money for X” make decisions manageable. Thaler (1999) treats this partly as a way of coping with limited attention and self-control, not simply as an error.

The shortcut has a cost when categories start to override arithmetic. The reference-point logic of prospect theory supplies the emotional layer. Closing an account at a loss feels worse than the same loss left open, so purchase price becomes the anchor for each account.

Thaler (1999) is balanced about whether mental accounting is good or bad for people. Separate pots can help with self-control, for example by keeping a pension out of easy reach. The same separation can cause avoidable costs elsewhere.

Ways to counter it

These are general techniques that behavioural finance researchers and practitioners discuss. They are not personal advice. Where a technique has no tested effect size, this is stated.

• Review the whole balance sheet. Investors who periodically list every asset and debt in one view, across ISAs, SIPPs, workplace pensions, cash and borrowing, can see trade-offs that separate accounts hide. This is a commonly discussed practice, not one with a measured effect.
• Look at total return. Adding price change and income together removes the artificial split between dividends and capital gains. This follows from the finding that prices fall when dividends are paid (Hartzmark and Solomon, 2019). Whether framing returns this way changes behaviour is a separate question that this article does not claim to answer.
• Use the fresh-money question. A practitioner heuristic asks whether a holding would be bought today, at today’s price, with new cash. It is meant to strip out the purchase price as a reference point. There is no tested effect size for it.
• Set rules in advance. A written policy on what happens to dividends, bonuses and gains, decided before they arrive, reduces the chance that each windfall gets its own improvised category. Peng, Miao and Xiao (2013) found that windfall money is treated differently from normal income in an experimental setting.
• Fix a review schedule. Checking at a set interval, instead of whenever an app prompts, changes how often accounts are balanced. Thaler (1999) and Benartzi and Thaler (1995) both identify evaluation frequency as a driver of how losses are felt.
• Keep buckets that serve a purpose. Separate pots for distinct goals can support discipline. The aim is to notice when a label is driving a decision that the numbers do not support.

Related biases

• Disposition effect: the trading pattern that results when each holding is its own account and closing one at a loss feels worse than the loss itself.
• Loss aversion: the prospect-theory asymmetry that makes open losses uncomfortable and gives the break-even effect its pull.
• Myopic loss aversion: loss aversion combined with frequent evaluation, closely tied to how often accounts are balanced.
• Sunk cost fallacy: letting money already spent shape a forward-looking decision, as happens when a purchase price becomes the benchmark for holding or selling.