Why Can Losing $50 Feel Weightier Than Gaining $50?

Imagine a coin-toss gamble: heads gives you $60; tails costs you $50. The average outcome is positive, yet many people would still hesitate because the possible $50 loss carries more psychological weight than the possible $60 gain. This tendency for a loss to influence choice more strongly than an equal gain, measured against the same reference point, is called loss aversion.

The reference point is crucial. People often treat the status quo, an expectation or a newly possessed state as zero: outcomes above it feel like gains and outcomes below it feel like losses. In prospect theory, Kahneman and Tversky represented this asymmetry with a value function that is steeper on the loss side. This does not mean every emotional response to loss is precisely twice that of a gain. Nor is it a theorem about how people ought to choose. It is a descriptive model of many observed choices.

Loss aversion is also not identical to every form of reluctance to let go. Continuing a failing plan because time has already been invested may reflect sunk cost; valuing an object more after acquiring it may reflect the endowment effect. These phenomena may be related to loss aversion, but they are not interchangeable. Before applying the label, identify the reference point and the structure of the options.

The evidence has boundaries. A large interdisciplinary meta-analysis published in 2024 examined 607 estimates from 150 articles and reported substantial loss aversion overall. Another 2024 meta-analysis, restricted to risky-choice data suitable for fitting individual parameters, found a smaller average effect and noted limited, uneven-quality data. Loss aversion therefore has empirical support, but its measured strength changes with the task, model, stakes, experience and person. It is useful for checking how “losing” frames a judgement, not for explaining every cautious choice as a fixed psychological instinct.

https://www.aeaweb.org/articles?id=10.1257/jel.20221698
https://doi.org/10.1016/j.joep.2024.102740