Men's almost universally greater tendency to overestimate their abilities is likely the heritage of a long evolutionary urge to achieve the upper hand in the mating game

The Cocksure Conundrum: How Evolution Created a Gendered Currency of Corporate Overconfidence. Richard Ronay, William W. Maddux & William von Hippel. Adaptive Human Behavior and Physiology, Aug 19 2022. https://rd.springer.com/article/10.1007/s40750-022-00197-5

Abstract: Biological differences between men and women mandate that women’s obligatory investment in reproduction is significantly greater than that of men. As a result, women have evolved to be the “choosier” of the two sexes and men have evolved to compete for female choice. To the degree that overconfidence is an effective tool for attracting mates and driving away competitors, greater competition among men suggests that they should express more overconfidence than women. Thus, sexual selection may be the primary reason why overconfidence is typically more pronounced in men than it is in women. Sexual selection may also be a distal, causal factor in what we describe as a cult of overconfidence pervading modern organizations and institutions. Whereas overconfidence was once regulated and constrained by features of ancestral life, levels of social mobility and accountability in contemporary society and modern organizations make it increasingly difficult to keep this gendered bias in check.

Why are People Overconfident?

Psychological explanations for the widespread tendency towards self-aggrandizement have focused primarily on intrapersonal hedonic benefits, such as higher self-esteem for those who believe they are better than others, and reduced reactivity to stressful events (e.g., Dunning et al., 1995; Taylor & Brown, 1988). An alternative possibility, however, is that overconfidence may be an evolved strategy of considerable utility for achieving status and other types of social currency, such as alliance formation, persuasion and influence, romantic attraction, and ultimately reproduction (von Hippel & Trivers, 2011a).

Individual differences in traits that lead to differential “fitness” within a given ecology result in reproductive variance (i.e., natural selection; Darwin, 1859; Andersson & Iwassa, 1996)). As a result, certain individuals’ expressed traits enable them to leave a greater genetic footprint than that of others. Across generations, this process leads to a proliferation of whatever genes assist in survival and reproduction. And so the footprint grows. The pan-cultural nature of the overconfidence bias (Sedikides et al., 2003) suggests that it may well be one such adaptive trait, selected for over many thousands of generations (Johnson & Fowler, 2011), likely predating the separation of our hominin ancestors from the ancestors of modern chimpanzees (e.g., Moore et al., 2009; Noë et al., 1980). Indeed, Daniel Kahneman speaks of overconfidence as being fundamentally built into the structures of human reasoning (Shariatmadari, 2015).

Given that overconfidence can lead to faulty assessments of one’s circumstances, and hence potentially perilous decisions, it seems odd that overconfidence might have provided an advantage in the context of natural selection. Believing oneself single-handedly capable of bringing down a woolly mammoth was unlikely to have been a winning attitude for our ancestors. Nonetheless, in competitive settings marked by uncertainty, overconfidence has the potential to maximize individual outcomes, so long as the associated costs of failure are outweighed by the benefits of possible success (Johnson & Fowler, 2011; Számadó, 2000). In other words, when the potential gains of achieving a particular goal outweigh the potential costs of failing, risky strategies such as overconfidence have adaptive utility (Adams & Mesterton-Gibbons, 1995; Számadó, 2003).

This perspective is consistent with Error Management Theory (Haselton & Buss, 2000), which predicts the emergence of psychological biases when; (1) the decision had recurrent impacts on fitness (reproductive success), (2) the decision is based on uncertain information, (3) the costs of false positives and false negatives were recurrently asymmetrical over evolutionary time. Overconfidence in one’s abilities meet these criteria quite well under many circumstances.

First, overconfidence potentially impacts fitness by helping individuals compete for sexual opportunity (Murphy et al., 2015) as well as other material and social resources that contribute to reproduction, such as status, prestige, and material rewards (e.g., Henrich & Gil-White, 2001). Second, uncertainty would have been a necessary condition for overconfidence to evolve, as certainty does away with the need for competition and signaling. In such cases, the strongest or most obviously qualified rival simply takes the desired resource. Indeed, as the uncertainty of contested outcomes increases, so too does the utility of overconfidence (Johnson & Fowler, 2011). Third, the costs in lost opportunities associated with being underconfident, or even accurate, are often greater than the costs associated with being overconfident, particularly when people compete with each other over limited resources (Soldà et al., 2021).

As a result of these processes, in situations where the potential gains of overconfidence outweigh the potential risks of overclaiming, overconfident individuals may have an advantage. If so, then overconfident displays may be a functional adaptation that help individuals acquire material and social benefits, depending on the relative magnitude of the risks and rewards that a given situation affords (Adams & Mesterton-Gibbons, 1995; Számadó, 2000, 2003). Consistent with the above reasoning, overconfident people gain a host of social and material benefits, such as increased perceptions of competence and a rise in social status and perceived leadership potential (Anderson et al., 2012; Ronay et al., 2019). As status increases, physiological markers of stress such as cortisol decrease (Sherman & Mehta, 2020) and dopamine sensitivity increases (Morgan et al., 2002), providing proximate, secondary mechanisms for the relationship between overconfidence and the maintenance of positive affect in response to social stressors (Ronay et al., 2019). Thus, it comes as no surprise that overconfidence is selected for in CEO appointments, despite the higher probability of overconfident leaders initiating value-destroying investments (Goel & Thakor, 2008) and financial reporting fraud (Schrand & Zechman, 2012). And perhaps, as we outline below, no surprise that fewer than 5% of CEO positions in the US and Europe are held by women (Edgecliffe-Johnson, 2018).

Why are Overconfident People So Often Men?

This distally focused, status-enhancing account of overconfidence has at least one important moderating factor: The available evidence strongly suggests that men tend to be more overconfident than women. For example, men exhibit more overconfidence than women in academic achievement (Bengtsson et al., 2005), finance and trading (Cueva et al., 2019; Prince, 1993), conflict and competitions (Johnson et al., 2006; Niederle & Vesterlund, 2007), science and mathematics (Ehrlinger & Dunning, 2003; Hyde et al., 1990), past performance (Reuben et al., 2012), intelligence (Steinmayr & Spinath, 2009), and on general knowledge and cognitive tasks (O'Laughlin & Brubaker, 1998; Pallier, 2003). By way of example, in one study, 70% of men and 30% of women, overestimated their work performance and professional skills (Lindeman et al., 1995). While under-confidence is generally the exception, it is more often women than men who err on the side of excessive humility (Lenney, 1977; Small et al., 2007), underestimating their chances of success across various outcomes (Erkut, 1983; Mura, 1987). Even successful women are more likely to attribute their triumphs to external causes, such as others in their team, or luck, rather than to personal aptitude (Campbell & Hackett, 1986; Haynes & Heilman, 2013; LaNoue & Curtis, 1985).

One potential origin of these observed gender differences is biased sampling, in that prior research has often assessed overconfidence in what are considered traditionally masculine domains. However, we argue that overconfidence is not a direct product of domain importance, expertise, or even stereotypicality; rather it is a product of the desire to persuade others of one’s competence in a given domain (Hoffman & Yoeli, 2022; von Hippel & Trivers, 2011a). As such, it is not so much the male-bias of the domains that matters, but the degree to which perceived ability in the domain can help people compete with members of the same sex or attract members of the opposite sex.

Thus, our theorizing also suggests that in contexts that stimulate competition between women, we might see stronger expressions of female overconfidence. For instance, given the importance of social support to female reproductive success (Campbell, 2004; Taylor et al., 2000), we might expect greater female overconfidence in domains related to emotional intelligence, such as empathy (Muthukrishna et al., 2018). And given that attractiveness is a primary dimension of competition for women (Blake et al., 2018; Buss, 1989), we might also see greater overconfidence in attractiveness among women than among men. A finding that is potentially consistent with this possibility is that the correlation between self- and other-ratings of physical attractiveness is substantially lower for women (r = 0.29) than it is for men (r = 0.53) (Marcus & Miller, 2003; see also Pereira et al., 2019).

Despite these possibilities, there are two important caveats that suggest male overconfidence is more important than female overconfidence in attracting a mate. First, with greater variability on a trait, competition for that trait becomes more important. In short-term mating the playing field among females is much more equal than it is among males (Brooks, 2021), suggesting that overconfidence is more likely to be wielded by males than females in short-term mating contexts. In long-term mating, competition among females is more focused on male traits that confer status, and as such, we might expect female status competition to increase with greater variability in male status and income. Support for this prediction can be seen in the finding that women’s self-sexualization occurs to a greater extent in environments that are economically unequal (Blake et al., 2018). Maximizing one’s beauty is a fruitful strategy for attracting high-status male partners (Udry & Eckland, 1984), which historically, has been an important strategy (indeed, sometimes the only strategy) for female survival and social mobility (Blake & Brooks, 2019).

But this possibility leads to our second point, which is that competition among females for long term mates is more focused on male traits that are not directly observable and hence can only be detected with greater uncertainty (e.g., the capacity to gain resources and assist in parental care giving; Taylor et al., 2000). As noted above, uncertainty magnifies the impact of overconfidence. As a consequence, overconfidence has more potential to enhance perceptions of important male traits (such as competence) than it does to enhance perceptions of important female traits (such as physical attractiveness). Perhaps for this reason, Blake (2018) finds that expenditure at beauty salons and women’s clothing stores also covaries with economic inequality, as adornments may be a more viable means of amplifying physical attractiveness than overconfidence. Women’s relative overuse of image-enhancing filters and photo editing (Dhir et al., 2016) may stem from similar motivations. The bottom line here is that males’ internal assets are more readily distorted via overconfident claims.

These differences in adaptive pressures are not unique to humans, and although cognitive tools such as language, theory of mind, and episodic foresight have dramatically enhanced the scope of human deception (Dor, 2017; Suddendorf et al., 2022), false signaling in the context of sexual competition is widespread. For instance, Noë et al. (1980) examined the role of dominance in explaining social rank within chimpanzee hierarchies, identifying three categories of dominance displays – agonistic, bluff, and competitive. They found that male chimpanzee’s social rank to be tied to displays of agonistic dominance (direct physical dominance) and bluff displays (closest to overconfidence). Females social rank was linked only with dominance in competition for space – such as giving way when another is approaching; and social competition – such as refraining from interacting with another partner because another chimp usually acts as partner to the third. Consistent with the human data, overconfidence has the greatest utility in the context of agonistic confrontations and bluff displays.

We can see similar effects further afield from our genetic roots: consider male fig wasps who signal their fighting ability during territorial competitions by displaying their impressive mandibles (Moore et al., 2009). Wasps with large mandibles are intimidating as they can inflict significant damage on opponents. Capitalizing on this advantage, there is an atypical male phenotype that develops mandibles 50% larger than expected for body size. These males are competitive signalers and engage in fewer fights than typical males but have higher mating success. Nonetheless, when compelled to combat, they fare poorly and incur more injuries than a typical male. Taken together, these examples from our near and distant cousins suggest that male sexual competition frequently takes the form of exaggerated signaling, a strategy that may be well served by self-deceptive overconfidence (see Angilletta et al., 2019).