Research

Abstract: This paper investigates the tragedy of ill-informed leadership in a setting where both leaders and followers lack certainty about the returns from a public good. In this context, leaders may unknowingly drive followers to form incorrect beliefs about the state of the world, ultimately leading them to make decisions that are to their disadvantage. Using a laboratory experiment, we show that leaders need not hold perfect information about the state of the world to be influential. When leaders are ill-informed, this influence takes a tragic turn, as leaders who (rationally) follow their information end up misleading followers and generating substantial welfare losses. Specifically, having a misleading leader is associated with a reduction in gains from contributions approximately twice as large as the reduction from halving the return from the public good. Our results underscore the significance of having well-informed leaders.

Abstract: This review examines how virtual embodiment interventions can inform economic research on inequality across social groups. In these interventions, widely used in psychology and related disciplines, virtual reality is used to embody individuals in virtual bodies whose appearance (skin tone, gender, age, etc.) can be experimentally manipulated to represent a member of a distinct social group. We propose a simple framework for understanding how outgroup embodiment may affect prejudice- and identity-related outcomes through other-regarding preferences, beliefs about the represented group, and beliefs about the self. We synthesize the existing literature, provide a practical guide to designing such interventions, and discuss the methodological trade-offs involved in their implementation. The review makes a two-way contribution: embodiment interventions offer economists a promising tool to study the socio-cognitive foundations of inequality, while an economic framework provides a path toward clarifying whether “changing bodies” can change not only minds but also behavior.

Abstract: Women put themselves forward for leadership roles less often than men. To study the role of confidence in one’s judgment as a driver of this pattern, we conduct two controlled laboratory experiments in which leaders provide nonbinding advice that can shape collective outcomes. In Study 1, we introduce a two-stage design that allows us to relate several dimensions of confidence to willingness to lead (WTL), while accounting for objective performance. Relative confidence—participants’ beliefs about how their performance compares with that of others—stands out among these dimensions: it predicts WTL for both genders and accounts descriptively for approximately one third of the performance-adjusted gender gap in WTL. Building on these findings, Study 2 identifies the causal effect of relative confidence on WTL using noisy performance feedback. We find that relative confidence has a positive and statistically significant causal effect on WTL: a one-standard-deviation increase in relative confidence raises WTL by 0.41 standard deviations. Overall, our results position confidence as a determinant of WTL and as a potential mechanism contributing to the gender gap in leadership representation.

Abstract: Gender can typically be inferred from observable cues like physical appearance, dress, or tone of voice. In professional environments that value stereotypically masculine characteristics, women may face pressure to adjust these cues to present themselves in more traditionally masculine ways. This paper asks whether the gender cues individuals perceive about themselves affect their own behavior. I study this question in the context of willingness to compete in the math domain, a career-relevant outcome with well-documented gender differences. In a controlled laboratory experiment, I use virtual reality to assign participants to male- or female-looking virtual bodies, alongside a no-embodiment control condition. I find no evidence that the virtual body’s gendered appearance shapes willingness to compete. However, simply experiencing a virtual body—whether male- or female-looking—reduces women’s math performance but not men’s. These findings provide no evidence that short-run changes in the gender cues people perceive about themselves alter their behavior, while showing that virtual embodiment itself can have unintended gender-asymmetric consequences.