ProBackend
Psychology & Behaviour
5 days ago6 min read

Why Confidence and Caution May Have Evolved Together

Why Confidence and Caution May Have Evolved Together

What is behavioral economics? A look at Dawson and de Meza’s hypothesis that overconfidence can signal ability while loss aversion limits its costs.

Behavioral economics and the puzzle of confidence

Why would a mind carry two tendencies that seem to pull in opposite directions? Overconfidence can push someone toward a risky choice; loss aversion makes a potential loss loom large and can hold them back. That apparent contradiction is the starting point for a new evolutionary argument by behavioural economists Chris Dawson and David de Meza.

Behavioral economics studies how people actually make choices, including the ways judgment and motivation depart from the idealized rational decision-maker in standard economic models. It brings psychological processes into the study of economic behaviour. The field doesn’t imply that every bias is useful, or that a pattern observed in people has a single evolutionary explanation. In this case, the narrower question is why overconfidence and loss aversion might persist together.

In their paper, “Talking the Talk, Not Walking the Walk,” published in Psychological Review, Dawson and de Meza propose that the two tendencies can form a functioning pair. Their proposal is an evolutionary signaling theory, not a settled account of human psychology. The authors argue that overconfidence may advertise ability because its decision-making costs are harder for highly able people to bear than for less able people; loss aversion may then limit some of the damage. (Paper: DOI 10.1037/rev0000644; University of Bath research announcement; Neuroscience News report.)

What is behavioral economics in this example?

The familiar economic ideal is a decision-maker who weighs information consistently and chooses the option that best serves their interests. Real people don’t reliably behave that way. They may be overconfident about their skill, or react more strongly to a possible loss than to a comparable gain. Behavioral economics asks how such patterns shape decisions rather than treating them as mere noise.

That framing matters here. Overconfidence is not simply “being confident.” It means holding a more favourable belief about one’s ability or prospects than the evidence warrants. Loss aversion, meanwhile, describes the tendency for losses to carry greater psychological weight than equivalent gains. The researchers treat these as potentially interacting tendencies, not two independent quirks to be labelled and dismissed.

The paper also sits within a broader debate about self-deception. One influential evolutionary account, associated with Robert Trivers, proposes that self-deception can make deception of others more effective by reducing tell-tale signs of insincerity. Dawson and de Meza identify a further puzzle: if people are often self-deceived, why should observers trust confident signals? Their alternative is that inflated self-belief may convey information, rather than merely conceal a lie. That is the authors’ proposed explanation, not proof that confidence is generally informative.

How costly confidence could signal ability

The logic resembles a costly signal in biology. A peacock’s conspicuous tail is an intuitive comparison: it takes resources to maintain and may bring disadvantages, so the signal is difficult for a weaker animal to imitate. The Bath announcement uses that analogy to explain the researchers’ idea. Human overconfidence, they argue, could work similarly if acting on overly positive beliefs creates errors that are less burdensome for more able individuals.

The key is the difference in costs, not confidence alone. Suppose two people hold equally inflated beliefs. If one has greater underlying ability, the resulting mistake may be less damaging on average; the other may pay more. If being perceived as capable also brings greater benefits to the more able, the theory predicts that a costly display of confidence could carry information about ability. In that model, overconfidence is not a reliable certificate of competence in every person or situation. It is a proposed signal whose credibility depends on unequal costs and benefits.

That distinction is easy to lose in a catchy analogy. A person’s confidence does not establish that they are skilled, and a bold claim can still be wrong. The researchers’ argument concerns why such a trait might persist across a population under particular evolutionary incentives. It does not give observers a shortcut for judging an individual, nor does the press coverage establish that the model explains every form of overconfidence.

Loss aversion as a private brake

If confidence encourages action and public persuasion, why wouldn’t it push people into ruinous risks? Dawson and de Meza’s answer is that loss aversion can act as a counterweight. The paper proposes that it partially reduces the decision costs of overconfidence: someone may project confidence outward while remaining sensitive to losses when making choices.

The signaling account depends on that restraint being at least partly hidden. If others see the confident display but not the full extent of private caution, loss aversion need not erase the signal. In the authors’ framing, the traits can be complementary: confidence may help someone seek opportunities or convince others, while sensitivity to losses curbs some costly follow-through. Neuroscience News describes this as an external “megaphone” paired with an internal “handbrake”; the metaphor is memorable, but the underlying point is a theoretical relationship, not a measured rule for every decision.

This is the paper’s most interesting move. Instead of asking which bias to delete, it asks whether the combination can produce a better outcome than either tendency alone. That does not mean the combination is always beneficial. Overconfidence can still contribute to poor decisions, and loss aversion can still block worthwhile opportunities. The theory proposes a possible evolutionary balance, not a universal prescription.

What the evidence does—and doesn’t—say

Dawson and de Meza report that evidence supports their interpretation, and their article presents the signaling logic as an explanation for the coexistence of the two tendencies. The available university research record identifies the work as a peer-reviewed article in Psychological Review and gives its DOI. The University of Bath announcement and Neuroscience News account explain the proposal for a general audience.

Those sources should be read with care. An evolutionary model can show how a trait might persist if its costs and benefits differ across people; it cannot, by itself, establish that this is the sole history of the trait or that every confident individual is signaling ability. The paper’s abstract states the authors’ claim about a credible signal and a role for loss aversion. That is a research argument to evaluate, not a settled fact to apply casually.

The practical implication is therefore modest. It may be too simple to assume that removing every cognitive bias is automatically helpful, because a tendency can have both costs and functions in a particular model. But the study does not show that people should cultivate overconfidence or ignore evidence of risk. Better decisions still require checking beliefs against outcomes, understanding the stakes, and noticing when confidence has outrun what one actually knows.

A useful hypothesis, not a confidence test

Behavioral economics is valuable partly because it asks what psychological tendencies do in real choices, not just whether they violate an ideal. Dawson and de Meza offer a striking answer to one puzzle: overconfidence may persist because its costs can make it a credible signal, while loss aversion may help keep those costs from becoming overwhelming.

It’s an elegant hypothesis. It is not a licence to treat confidence as proof of competence, and it doesn’t settle the evolutionary story. The more careful takeaway is that apparently opposing tendencies may interact in ways worth studying—and that explaining a bias is different from endorsing it.

behavioral economics and the puzzle of confidencebehavioral economics and the puzzle of confidence