Researchers at the University of Bath and the London School of Economics and Political Science say they have developed a new evolutionary explanation for one of human psychology’s most persistent quirks: our tendency to remain overconfident even when inflated self-belief can lead to serious mistakes. Their study argues that overconfidence may survive not in spite of its costs, but because those costs make confidence informative to other people.
Published in Psychological Review, the research presents overconfidence as a form of costly biological signalling. In evolutionary biology, a signal is considered credible when it is difficult for everyone to produce or maintain. The classic example is the peacock’s tail: its size and brightness can attract predators and consume energy, yet those same burdens make it a potentially reliable indicator of the bird’s condition. The researchers propose that exaggerated human self-belief may operate according to a similar principle.
Overconfidence refers to believing that one is more capable, knowledgeable or likely to succeed than objective evidence warrants. It is associated with failed businesses, excessive risk-taking, poor investment decisions and inaccurate judgments about personal performance. Yet the researchers argue that these disadvantages are not distributed equally. Individuals with greater underlying ability may be better able to absorb the consequences of overconfidence because they make fewer errors, recover more effectively from failure or derive greater benefits when confidence persuades others to offer them opportunities.
This creates what the researchers describe as a separating signal. If confidence had no consequences, everyone could claim to be exceptionally talented, making confident statements worthless. But if exaggerated self-belief creates real risks, the most capable individuals may be more able to sustain it without suffering catastrophic losses. Less capable individuals, by contrast, may pay a larger price when their beliefs lead them to pursue unsuitable opportunities, take unnecessary risks or ignore warnings. The signal does not need to be perfectly accurate; it only needs to contain enough information for observers to distinguish, on average, between different levels of ability.
“Overconfidence is strikingly common and often costly,” said Chris Dawson, a behavioural economist at the University of Bath’s School of Management. “It contributes to failed businesses, excessive risk-taking and poor decisions. But the real puzzle is why evolution hasn’t eliminated it. Our research suggests that those costs are exactly what makes it useful.”
Confidence can produce social and material rewards before ability has been independently verified. A person who presents themselves as capable may be more likely to obtain a leadership role, attract investment, receive a promotion or gain access to valuable relationships. Once the opportunity is secured, however, actual performance becomes harder to conceal. In this framework, confidence helps open doors, while ability determines whether someone can remain successful after entering. The more capable therefore receive both a greater potential benefit from projecting confidence and a lower expected cost from being wrong.
The study builds on an influential theory proposed by evolutionary biologist Robert Trivers, who suggested that self-deception may have evolved because it helps people deceive others. Conscious deception can produce detectable signals of dishonesty, including hesitation, nervousness or unnatural vocal patterns. If individuals genuinely believe their own exaggerated claims, they may suppress some of these cues and become more convincing. The new research addresses a problem left open by that theory: if people know that confidence can be deceptive, why do they not simply discount every confident claim as cheap talk?
David de Meza of LSE’s Department of Management said the answer lies in the relationship between confidence and its consequences. People may discount self-beliefs, but rejecting all confident claims would also lead them to underestimate individuals who are genuinely capable. Because excessive self-belief carries a price, it can retain informational value. The boldest claims are not automatically trusted, but they may be treated as more credible when observers understand that only some individuals can afford the risks associated with making them.
The researchers also connect their model to loss aversion, the well-established psychological tendency to experience losses as more significant than equivalent gains. Loss aversion may act as a hidden restraint on overconfidence. People can speak boldly and cultivate an image of high status while remaining cautious when actual losses are possible. This combination allows confidence to generate social benefits without encouraging every dangerous action implied by a person’s public self-image. Because the caution may not be visible to observers, it can reduce destructive risk-taking without completely weakening the outward signal.
The study further proposes an evolutionary explanation for why men, on average, often report higher levels of overconfidence than women. The authors argue that historical differences in reproductive and social incentives may have shaped which qualities required signalling. They suggest that status, commitment, resource acquisition and social influence were difficult for potential partners to observe directly, creating pressure for men to advertise underlying capability through confident behaviour. The researchers emphasize that this is a population-level evolutionary account rather than a prediction about every individual, and that social norms and institutions can strongly influence confidence today.
By combining overconfidence with loss aversion, the model challenges the assumption that every psychological bias should simply be removed. Interventions designed to eliminate overconfidence entirely could make people less persuasive, less willing to pursue opportunities or less capable of attracting support for promising plans. Removing loss aversion could create a different danger by encouraging excessive risk-taking. The researchers do not suggest that costly mistakes are desirable, but argue that the apparent irrationality of human self-belief may reflect a deeper strategic balance: confidence helps people compete for opportunities, while caution limits the damage when their beliefs exceed reality.
Subject of Research: People
Article Title: Talking the Talk, Not Walking the Walk: The Coevolution of Overconfidence and Loss Aversion
News Publication Date: 20-Jul-2026
Web References: https://doi.org/10.1037/rev0000644
References: Psychological Review, DOI: 10.1037/rev0000644
Keywords: overconfidence, loss aversion, evolutionary psychology, costly signalling, self-deception, behavioural economics, social status, human behaviour, decision-making, evolutionary biology
