Why Losses Feel Stronger Than Gains

Ask someone to flip a coin for a fair bet: heads you gain $100, tails you lose $100. Most people refuse, even though the expected value is exactly zero and the math is neutral. Daniel Kahneman and Amos Tversky found the reason in a series of experiments that became one of the most cited results in behavioral science: losses register in the mind roughly one and a half to two and a half times as intensely as equivalent gains, a pattern they called loss aversion.

What loss aversion actually is

Loss aversion is the tendency for a loss to produce more psychological pain than an equivalent gain produces pleasure. It is not the same thing as being cautious or risk-averse in general. A person can accurately calculate the odds of a bet, understand that the expected value is favorable, and still turn it down, because the anticipated sting of losing outweighs the anticipated satisfaction of winning.

Kahneman and Tversky first documented this systematically in their landmark paper, Prospect Theory: An Analysis of Decision under Risk, published in Econometrica in 1979. The paper challenged the dominant economic model of the time, expected utility theory, which assumed that people evaluate outcomes based on final wealth states and weigh probabilities in a straightforward, consistent way. Kahneman and Tversky's experiments showed something different: people evaluate outcomes as gains or losses relative to a reference point, and they treat the two categories asymmetrically.

The 1979 paper became one of the most cited works in economics and psychology, and it laid the groundwork for what eventually became the field of behavioral economics. Kahneman received the 2002 Nobel Memorial Prize in Economic Sciences largely on the strength of this research program. Tversky, his longtime collaborator, had died in 1996; the Nobel committee does not award prizes posthumously, so the recognition went to Kahneman alone, though he has repeatedly credited the work as a joint achievement.

What made the finding useful, rather than just interesting, is that it was reproducible across a wide range of contexts: gambling, employment decisions, insurance purchases, negotiation, and everyday consumer choices. The asymmetry between losses and gains kept showing up, which is what turned a single experiment into a durable theory.

Tversky and Kahneman's follow-up paper in 1992 estimated the loss aversion coefficient at approximately 2.25, meaning a loss needed to be offset by a gain roughly two and a quarter times its size before people treated the gamble as psychologically even. Later replications have found the ratio ranges from about 1.5 to 2.5 depending on the stakes and the domain tested.

Prospect theory and the loss aversion coefficient

Prospect theory replaced the older utility-curve model with something closer to how people actually behave. Instead of asking "does this leave me wealthier overall," the mind asks "relative to where I started, is this a gain or a loss," and then applies very different psychological weighting to each answer.

In the original 1979 experiments, participants were presented with paired gambles and asked which they preferred. A typical setup offered a certain loss of a fixed amount against a probabilistic gamble with a larger potential loss but a chance of losing nothing. Kahneman and Tversky found that people became risk-seeking when facing sure losses, gambling to avoid locking in the loss, even when the gamble had worse expected value. That is the mirror image of how the same people behaved when the outcomes were framed as gains, where they became risk-averse and preferred the sure thing.

The follow-up paper, Advances in Prospect Theory: Cumulative Representation of Uncertainty, published by Tversky and Kahneman in the Journal of Risk and Uncertainty in 1992, refined the original model and produced the widely cited loss aversion coefficient of approximately 2.25. That number is often rounded to "roughly two to one" in popular summaries, but the actual estimate carries a decimal, and subsequent studies using different stakes, populations, and elicitation methods have found values anywhere from about 1.5 to 2.5. A meta-analysis is not the same as a single fixed constant, and researchers studying loss aversion generally treat the ratio as a robust range rather than a single precise number.

It's worth being precise about what the coefficient measures. It is not a claim that losing $100 feels "2.25 units" bad in some absolute sense. It is a relative measure: on average, a loss of a given size requires a gain of roughly 1.5 to 2.5 times that size before people rate the two as emotionally or decision-theoretically equivalent. The ratio is a behavioral summary of thousands of paired-choice responses, not a law of physics.

How was the loss aversion coefficient actually measured?

Researchers typically present participants with a series of mixed gambles, for example a coin flip that could lose $X or gain $Y, and vary Y until the participant is indifferent between taking the gamble and declining it. The ratio of Y to X at that indifference point is the loss aversion coefficient, which Tversky and Kahneman's 1992 estimate placed at approximately 2.25.

The value function: why the curve isn't symmetric

Kahneman and Tversky's prospect theory value function is usually drawn as an S-shaped curve: concave for gains, convex for losses, and distinctly steeper on the loss side of the origin. Picture a two-lane bridge crossing a canyon. The lane heading toward gains slopes gently upward, so each additional dollar of gain adds a little less subjective pleasure than the one before it, consistent with diminishing marginal utility. The lane heading toward losses drops off steeply right past the starting point, so each additional dollar of loss subtracts more subjective pain than a dollar of gain would have added.

This asymmetry is why a $20 loss and a $20 gain do not cancel out psychologically, even though they cancel out arithmetically. Someone who loses $20 and then recovers exactly $20 typically does not feel neutral about the sequence. They still feel behind, because the initial loss carried more subjective weight on the way down than the recovery carried on the way back up. The reference point itself can also shift once a loss is absorbed, which is part of why "getting back to even" after a bad trade or a bad month can feel so difficult even when the raw numbers say the account has recovered.

A second useful analogy: think of the value function like a thermostat with different sensitivities in each direction. Turning the temperature up from a comfortable baseline barely registers until it climbs several degrees, but turning it down from that same baseline produces a noticeable shiver almost immediately. The nervous system, in this analogy, is not symmetric in how it registers deviation from a set point, and prospect theory says something structurally similar happens with money, time, and other resources measured against a reference point.

The reference point itself is not fixed. It usually defaults to the status quo, but it can be reset by expectations, recent history, or explicit framing. A trader who expected a 10 percent quarterly return and receives 3 percent may experience that outcome as a loss relative to the expected reference point, even though the account technically gained value. This is one reason expectation management matters so much in negotiation and in workplace performance reviews: the number that gets compared against is not always the number people assume it is.

Endowment effect: loss aversion applied to ownership

A closely related demonstration of the same underlying asymmetry comes from Richard Thaler's work on the endowment effect, later tested experimentally by Kahneman, Jack Knetsch, and Thaler in Experimental Tests of the Endowment Effect and the Coase Theorem, published in the Journal of Political Economy in 1990. The endowment effect is the tendency to value an object more once you own it than you would value the identical object before owning it.

In the study's best-known design, researchers gave half of a group of participants a coffee mug and then asked them to name a selling price. The other half, who received no mug, were asked to name a buying price for an identical mug. Classical economic theory predicts these two prices should converge, since the mug's utility to a random participant shouldn't depend on which half of the room they happened to sit in. Instead, the sellers demanded roughly twice the price that buyers were willing to pay, a gap the researchers replicated across multiple rounds of the experiment with different objects.

Nothing about the mug itself changed between the two groups. What changed was the framing: giving up the mug was coded as a loss for the owners, while acquiring the mug was coded as a gain for the non-owners. Because losses are weighted more heavily than equivalent gains, owners priced their mugs as though giving them up would hurt more than acquiring an identical mug would help, which is exactly the same asymmetry documented in the gambling experiments a decade earlier.

The endowment effect has since been replicated with a range of goods beyond mugs, including pens, sports tickets, and lottery ticket entitlements, and it shows up in real-world negotiations. A homeowner listing a house they've lived in for years will frequently anchor on a price well above comparable market value, not necessarily out of greed but because the prospect of parting with the property is processed, psychologically, as a loss rather than as a neutral transaction.

Does the endowment effect apply to non-physical possessions?

Yes. Studies have extended the endowment effect to intangible entitlements such as legal rights, insurance policy terms, and even assigned tasks at work, where people resist giving up something they've been assigned or granted even when a neutral evaluation would favor trading it away. The underlying mechanism is the same: relinquishing something already held is coded as a loss.

Where loss aversion shows up in investing

Loss aversion is not confined to laboratory gambles and mug experiments. It shapes measurable financial behavior, most famously in a pattern called the disposition effect, where investors sell winning stocks too early to lock in a gain while holding losing stocks too long in the hope of avoiding a realized loss.

The disposition effect was documented rigorously by Terrance Odean in Are Investors Reluctant to Realize Their Losses?, published in the Journal of Finance in 1998. Odean analyzed the trading records of more than 10,000 individual brokerage accounts and found that investors sold winning positions at a meaningfully higher rate relative to gains than they sold losing positions relative to losses, even though the tax code in most years gives investors a financial incentive to do the opposite, since realizing a loss can offset taxable gains.

Odean's explanation traces directly back to prospect theory. Selling a losing stock forces the investor to convert a "paper loss" into a realized loss, closing the mental account at a defeat. As long as the position stays open, the investor can tell themselves the loss isn't final, that the stock might recover. Selling a winning stock, by contrast, locks in a gain and lets the investor close the mental account as a clear win, even when the underlying evidence suggests holding the winner longer would have been the better financial decision.

Odean's dataset showed the effect held even though selling losers, not winners, was usually the better move for taxes in the years he studied. Investors were paying, in effect, to avoid the psychological discomfort of admitting a loss.

The disposition effect has been replicated in other markets and asset classes since 1998, including studies of real estate transactions, where sellers who bought a property above its current market value tend to set asking prices anchored to their original purchase price rather than to current comparable sales, delaying sales and sometimes accepting worse outcomes than a faster, loss-accepting sale would have produced.

Loss aversion also affects how retail investors respond to market downturns more broadly. Because a portfolio decline is processed as a loss relative to a recent high-water-mark reference point, investors are more likely to sell into a falling market near the bottom, which is the opposite of the buy-low behavior that would benefit them financially, and more likely to hold on stubbornly during a slow bleed if selling would require formally realizing the loss.

Domain	Manifestation of loss aversion	Key source
Gambling decisions	Requiring a gain roughly 1.5 to 2.5 times a potential loss before accepting an even-odds bet	Tversky & Kahneman, Journal of Risk and Uncertainty, 1992
Ownership (endowment effect)	Demanding roughly twice the market price to sell an owned mug versus buy an identical one	Kahneman, Knetsch & Thaler, Journal of Political Economy, 1990
Investing (disposition effect)	Selling winning stocks early, holding losing stocks too long, across 10,000+ brokerage accounts	Odean, Journal of Finance, 1998
Negotiation	Concessions framed as a loss feel worse than an equivalent gain feels good	Kahneman & Tversky, Econometrica, 1979
Insurance purchasing	Overpaying for coverage against small, low-probability losses relative to their actuarial cost	Kahneman & Tversky, Econometrica, 1979

Documented loss-aversion effects span gambling, ownership, and financial markets, but the magnitude of the asymmetry varies by domain and stakes.

Why do losses feel stronger than gains psychologically?

Kahneman and Tversky's prospect theory found that the subjective value function is steeper for losses than for gains of equal size, meaning a loss of a given amount produces more negative feeling than a gain of the same amount produces positive feeling. Tversky and Kahneman's 1992 estimate placed this loss-to-gain sensitivity ratio at roughly 2.25, with later studies replicating a range of about 1.5 to 2.5 depending on context and stakes.

What people get wrong about loss aversion

A common misreading treats loss aversion as evidence that people are simply irrational or bad at arithmetic. That is not what the research shows. The asymmetry between how losses and gains are weighted is not a computational error; it is a consistent, structural feature of how the mind assigns subjective value relative to a reference point. Participants in Kahneman and Tversky's experiments could often calculate the expected value of a bet correctly and still decline it, because the decision wasn't being made on expected value alone. It was being made on subjective, reference-dependent value, which is a different quantity entirely.

A second common mistake is treating the 2-to-1 ratio, or the more precise 2.25 figure from the 1992 paper, as a fixed universal constant that applies identically to every person, every domain, and every stake size. Kahneman himself cautioned against this in Thinking, Fast and Slow, published in 2011, noting that the exact ratio shifts with the size of the amounts involved, the domain being tested (money behaves somewhat differently than time or health outcomes), and individual differences between people. Some individuals show much higher loss aversion than the group average; a smaller number show close to none. Treating 2.25 as gospel for every scenario overstates the precision of what is genuinely a robust, but variable, behavioral tendency.

A third confusion worth clearing up: loss aversion is not the same concept as risk aversion, even though the two frequently travel together and get used interchangeably in casual conversation. Risk aversion describes a general preference for certain outcomes over uncertain ones of equal expected value, and it can apply symmetrically to both gains and losses. Loss aversion is specifically about the differential weighting of losses versus gains, independent of whether the outcome is certain or probabilistic. A person can be loss averse without being especially risk averse, and the prospect theory experiments actually found the opposite pattern for losses under risk: people become more risk-seeking, not less, when trying to avoid a certain loss, which is the reverse of what a simple risk-aversion model would predict.

Finally, some critiques of the original research point out that most of the foundational experiments used hypothetical, relatively small-stakes gambles with student or general-population samples, rather than large real-money stakes with professional traders. Later field studies, including Odean's brokerage-account analysis and subsequent real-estate and insurance-market research, addressed this by measuring actual behavior with real financial consequences, and the core asymmetry has generally held up, even if the exact magnitude varies by setting.

What this means for everyday decisions

Understanding loss aversion doesn't make the asymmetry disappear, but it does explain a lot of behavior that otherwise looks inconsistent. It explains why a modest market downturn triggers outsized anxiety relative to an equivalent rally's mild pleasure. It explains why negotiators who frame a concession as "you're losing X" get more resistance than negotiators who frame the identical concession as "you're not gaining X." And it explains why people cling to underperforming investments, subscriptions, or relationships longer than a purely rational cost-benefit calculation would recommend: walking away requires converting an ambiguous, ongoing situation into a definite, realized loss.

None of this means loss aversion is a flaw to be engineered away. It is, if anything, a sensible evolutionary bias in contexts where losses genuinely carried more survival weight than equivalent gains, a point Kahneman has made in discussing why the asymmetry likely evolved in the first place. The practical value of knowing about it is recognizing when the feeling of loss is distorting a decision that would otherwise be close to neutral on the numbers, whether that decision is about selling a stock, ending a subscription, or walking away from a bad negotiation.

The takeaway

Loss aversion is not a quirk restricted to gambling tables or lab experiments. It is a structural feature of how the mind assigns subjective value relative to a reference point, and it surfaces whenever a decision can be framed as a potential loss rather than a foregone gain. Kahneman and Tversky's 1979 and 1992 papers gave this pattern a name and a rough magnitude, somewhere between 1.5 and 2.5 times, and later research from the endowment effect to the disposition effect in investing kept finding the same asymmetry showing up in different clothes, from a coffee mug in a Cornell classroom to ten thousand brokerage accounts.

---

Read next:

• Why Choice Overload Makes Decisions Harder
• Why Goals Fail Without Feedback
• Why Confirmation Bias Is Hard to Notice

FAQ

Q: What is loss aversion?

Loss aversion is the tendency for losses to carry more psychological weight than equivalent gains. Daniel Kahneman and Amos Tversky documented it in their 1979 prospect theory paper, finding that people typically need a potential gain roughly 1.5 to 2.5 times the size of a potential loss before accepting an even-odds gamble.

Q: What is the exact loss aversion coefficient found by Kahneman and Tversky?

Tversky and Kahneman's 1992 paper in the Journal of Risk and Uncertainty estimated the loss aversion coefficient at approximately 2.25, meaning a gain needed to be about 2.25 times the size of a potential loss before people treated a gamble as roughly even. Later studies have found this ratio ranges from about 1.5 to 2.5 depending on stakes and context.

Q: What is the endowment effect and how does it relate to loss aversion?

The endowment effect is the tendency to value something more once you own it. Kahneman, Knetsch, and Thaler's 1990 mug experiments, published in the Journal of Political Economy, found owners demanded roughly twice the price to sell a mug that non-owners were willing to pay to buy one, because selling framed the outcome as a loss.

Q: Does loss aversion affect investing?

Yes. Terrance Odean's 1998 analysis of more than 10,000 brokerage accounts, published in the Journal of Finance, found investors tend to hold losing stocks too long to avoid realizing a loss, while selling winning stocks too quickly to lock in a gain, a pattern known as the disposition effect.

Q: Is the 2-to-1 loss aversion ratio always exactly true?

No. Experimental estimates place it in a range of roughly 1.5 to 2.5, centered near 2.25 in Tversky and Kahneman's 1992 estimate, and Kahneman noted in Thinking, Fast and Slow (2011) that the exact ratio shifts with stakes, domain, and individual differences, so it should be treated as a consistent tendency rather than a fixed constant.

Q: Is loss aversion the same as risk aversion?

No. Risk aversion is a general preference for certainty over uncertain outcomes even at equal expected value. Loss aversion is specifically about losses feeling worse than equivalent gains feel good, and it can actually make people more risk-seeking when trying to avoid a certain loss, the opposite of what simple risk aversion would predict.

Q: Were the original loss aversion experiments tested with real money?

The 1979 and 1992 prospect theory experiments primarily used hypothetical gambles. Later field research, including Odean's 1998 brokerage-account study and subsequent real-estate market studies, tested the same asymmetry using real financial stakes and found the core pattern held up, though the exact magnitude varies by setting.