Do Losses Hurt Twice as Much as Gains? The Evidence Is Messier Than the Slogan

Checked against primary sources by the Editorial Team. Every figure in this article was traced to the study it came from, including the ones that contradict what we previously published. How we review.

This article contains affiliate links. If you buy something through one of them, we may earn a commission at no extra cost to you. It never changes what we recommend or what we say about it. Advertiser Disclosure.

You have probably seen the line: losing $100 hurts twice as much as gaining $100 feels good. It sounds tidy, and it has traveled a long way. The research behind it exists and matters, but the "twice" part has been tested many times since, and the answers are not as neat as the slogan.

Where the idea comes from

Daniel Kahneman and Amos Tversky introduced prospect theory in a 1979 paper in Econometrica. Kahneman later received the 2002 Nobel Memorial Prize in Economic Sciences; the Nobel committee's citation credits him for integrating insights from psychological research into economic science, especially on human judgment and decision making under uncertainty. The prize citation does not mention the number two.

The number comes from a later paper. In 1992, Tversky and Kahneman published a refined version of the theory, called cumulative prospect theory, in the Journal of Risk and Uncertainty. Later authors who quote that paper report a median loss aversion coefficient of 2.25. In plain terms, a coefficient of 2.25 means a loss gets weighted about two and a quarter times as heavily as a gain of the same size. Round it down and you get "twice as much."

Two things are worth noticing. First, 2.25 was a median from one set of experiments, not a constant of human nature. Second, a coefficient is a fitted number inside a mathematical model. It depends on how the model is set up and on what choices people were asked to make.

What the large reviews find

Two meta analyses published in 2024 pooled earlier work, and they do not agree on the size of the effect.

Brown, Imai, Vieider, and Camerer, in the Journal of Economic Literature, analyzed 607 empirical estimates from 150 studies. They report a mean loss aversion coefficient of 1.955, with a 95 percent probability interval from 1.820 to 2.102. They also report that few observable features of study design were strongly related to differences in the estimates. On their numbers, "about twice" is a fair summary of the average across that literature.

Walasek, Mullett, and Stewart, in the Journal of Economic Psychology, took a stricter route. They looked at 17 published studies (19 datasets) that estimated the coefficient in cumulative prospect theory using mixed gambles, which offer both a possible gain and a possible loss. Their aggregate estimate was 1.31, with a 95 percent interval from 1.10 to 1.53, well below the 2.25 from 1992. The authors also report that estimates ranged from 0.65 to 3.45, that twelve of the nineteen datasets had intervals spanning both loss neutrality (about 1) and substantial loss aversion (above 2), and that a lot of the available data had measurement problems.

The two reviews differ in scope and method, so the gap between 1.955 and 1.31 does not mean one team made an error. It does show that the exact number depends on which studies and which methods you trust.

The challenge to the whole idea

A sharper dispute is about whether losses loom larger at all. David Gal and Derek Rucker argued in a 2018 paper in the Journal of Consumer Psychology that losses do not generally loom larger than gains. As described by Simonson and Kivetz in their published reply, Gal and Rucker contend that the main evidence for loss aversion, the endowment effect and status quo bias, can be explained by other mechanisms. We could not open the Gal and Rucker paper itself, so we describe it here only as the reply and a later commentary characterize it.

Simonson and Kivetz accept part of the critique. They maintain that loss aversion probably does occur, but that it is contingent and shaped by moderators, not a universal law. They also question some of Gal and Rucker's experimental procedures.

Herrmann, Mrkva, Gaechter, and Johnson responded with data. In the Journal of Consumer Psychology in 2020, they report five experiments with more than 17,700 participants. They found that people of all knowledge and experience levels were loss averse. They also found that more domain knowledge and experience were associated with lower loss aversion, and that older consumers showed higher loss aversion across their measures. Their title says it plainly: loss aversion has moderators, but reports of its death are greatly exaggerated.

Why the number moves

Several lines of work point to context.

• The range of amounts you see. Walasek and Stewart ran four experiments in which they changed the range of gains and losses participants encountered. They report producing loss aversion, loss neutrality, and even the reverse of loss aversion, which suggests that the experienced mix of outcomes shapes the effect (Journal of Experimental Psychology: General, 2015).
• The size of the stakes. Mukherjee, writing in Frontiers in Psychology in 2019, argues that loss aversion depends on magnitude: gains can equal or exceed losses for small monetary amounts, while losses dominate at larger ones. That is his reading of the literature, and it is a position, not a settled result.
• Knowledge and experience. As noted above, Herrmann and colleagues report lower loss aversion where people knew more about the domain.

Put together, the fair reading is that many people weigh losses more than equal gains, often by a noticeable margin, and that the margin is a moving target. The "twice" in the old headline is a reasonable ballpark for one prominent dataset and for the average in one large meta analysis. It is a poor guide to what any one person will feel in any one situation.

What this means for how you think about money

This article offers no advice about what to buy, sell, or hold. The point is narrower: when someone tells you that your brain makes you weigh losses exactly double, you can treat that as a rough average with a wide spread around it.

It also helps to separate the feeling from the decision. Based on the moderator findings above, one way to act on them is to notice that the same dollar loss can land differently depending on what you have recently seen, how big the stakes are, and how familiar you are with the choice. If you are making a decision that matters, writing down the actual amounts on each side can show whether the feeling of loss is carrying more weight than the numbers.