Loss aversion in marketing: the asymmetry that sells urgency, and the fight over whether it exists

Origin

Daniel Kahneman and Amos Tversky published Prospect Theory: An Analysis of Decision under Risk in Econometrica in 1979 [1]. The problem they were solving was that expected utility theory, the standard account of rational choice, did not describe how people actually chose between risky options. People were not consistently risk-averse or risk-seeking. They were risk-averse for gains and risk-seeking for losses, and they treated a loss as worse than a gain of the same magnitude.

The paper’s value function is S-shaped: concave for gains, convex for losses, and steeper for losses than for gains. The steepness is loss aversion. Their phrasing: “losses loom larger than gains” [1].

The 1979 paper did not put a number on it. Tversky and Kahneman’s 1992 paper on cumulative prospect theory estimated the loss aversion coefficient at around 2.25 from their experimental data, which is where the “losses hurt twice as much” shorthand comes from [2]. The shorthand drops the context the estimate came from, which was monetary gambles in a laboratory.

What the evidence shows

The asymmetry holds up in gambles. Across many replications of the original paradigm, people do demand more to accept a risky bet than expected value would predict, and the asymmetry between gains and losses recurs [1][2].

The endowment effect follows from it. Thaler coined the term in 1980 for the observation that people value things more once they own them [4]. Kahneman, Knetsch and Thaler’s 1990 experiments gave participants mugs and found they demanded roughly twice as much to sell a mug as others would pay to buy it [5]. The gap is loss aversion applied to ownership: giving up the mug is a loss, acquiring it is a gain.

The S-shape is the model: concave for gains, convex for losses, and steeper below the reference point. No loss-aversion coefficient is drawn, because the size of the asymmetry is the contested part and varies by setting.

Source: Illustrative. After Kahneman and Tversky, Prospect Theory, 1979. Brave New Media, 2026.

Framing effects are large and replicable. The same outcome described as a loss or as a gain produces different choices. Tversky and Kahneman’s Asian disease problem is the canonical demonstration, and framing has replicated in medical, financial and consumer contexts [6].

The marketing applications are documented but mostly correlational. Hardie, Johnson and Fader found loss aversion in consumer brand choice data in 1993 [7]. Price increases hurt more than equivalent decreases help, in field data, which is consistent with the theory and also with several other explanations.

The evidence is strong that people respond asymmetrically to losses and gains in many situations. It is weaker on whether this is one mechanism with one coefficient, and that is where the challenge comes in.

The challenge

Gal and Rucker’s 2018 paper, The Loss of Loss Aversion, argued that the evidence for loss aversion as a general principle has been overstated [3]. Their case: many of the findings attributed to loss aversion have alternative explanations, the effect does not appear reliably when stakes are small, and in some conditions gains are weighted more than losses. They did not argue that the asymmetry never occurs. They argued it is a contextual effect dressed as a law.

The reply came quickly. Mrkva and colleagues in 2020 and others defended the broader reading, pointing to the breadth of settings in which the asymmetry appears and arguing that Gal and Rucker’s alternative explanations do not cover the full record [8]. The field has not settled this and a marketer should not pretend it has.

A second challenge is about the coefficient. The 2.25 figure came from one set of monetary gambles. Estimates in other domains range widely, and treating any single number as the exchange rate between loss and gain is a precision the data does not support.

The honest position for this section: the asymmetry is real and common. The claim that it is universal, or that it has a fixed size, is contested by serious people and the contest is ongoing. That is how the entry should read and how a marketer should use it.

Where I see this go wrong

Every offer is framed as a loss. The deadline, the expiring discount, the “only three left”. Used once, with a real constraint behind it, this works. Used on every email, buyers learn that the constraint is fictional and the brand that cried wolf has spent its credibility on urgency. The mechanism depends on the loss being believed.

The coefficient is quoted as a law. “Losses hurt twice as much as gains” goes into the deck as if it were a physical constant. It was one estimate from one paradigm and the field is arguing about whether it generalises. A number that precise, in a dispute that live, is not a number to build a strategy on.

Loss framing is applied to things nobody owns. The endowment effect works because the buyer already has the thing. Telling a prospect they will “lose” a discount they never had is not loss aversion. It is a sentence with the word loss in it.

I spent years around investment before I ran an agency, and loss aversion is the one finding from this section that trading floors had already priced in. Nobody there needed Kahneman to know that a loss felt worse than a gain. What they needed, and what marketing needs, is the second half: that the asymmetry is unreliable, that it shrinks when people are experienced, and that a strategy built on it is a strategy built on buyers staying naive.

What to do with it

1. Use loss framing only where the loss is real. A trial that actually expires, a price that actually rises, a capacity that is actually limited. The mechanism is powerful when believed and destructive when discovered to be theatre.
2. Count how many of your communications in a quarter use urgency. If the answer is most of them, the mechanism has stopped working on your list and started costing you.
3. Prefer endowment to threat. Let the buyer own something, a trial, a configured demo, a draft plan with their name on it, before asking them to give it up. That is the documented version of the effect.
4. Do not quote 2.25, or “twice as much”, to a board as a law. Say that losses are weighted more heavily than gains in most studies, that the size varies, and that serious researchers dispute how general it is. That is the accurate sentence.
5. Test the framing in your own category before relying on it. Gal and Rucker’s point is that the effect is contextual. Your context is the one that matters and nobody has measured it for you.
6. Read the 1979 paper and the 2018 critique together. Twenty-eight pages and thirty. Most of what is said about loss aversion in marketing is said by people who have read neither.

If you run the company rather than the marketing

You will see loss aversion in two places. In your marketing, as urgency: deadlines, expiring offers, scarcity. And in your own decisions, as reluctance to cut a project that has already consumed money, which is the same asymmetry pointed inward.

On the marketing. Ask how often the urgency is real. If every campaign has a deadline, the deadline has stopped being information and started being a tic, and the buyers who have noticed are the experienced ones you most want. One real constraint a quarter does more than twelve fictional ones.

On the number. If anyone tells you losses hurt exactly twice as much as gains, know that the figure comes from one laboratory estimate in 1992 and that the field has spent the years since 2018 arguing about whether the effect is even general. It is a tendency, not a constant, and a strategy that treats it as a constant is overfitting to a disputed decimal.