Men are far more likely to start new ventures than women. We argue that one explanation of this gap is that women respond differently to signals of past entrepreneurial success due to the "male hubris, female humility" effect. We argue that as a result women are disproportionately less likely to persist in second founding attempts than men when they have succeeded or failed by large margins. Using a data set of serial founders in crowdfunding, we find evidence supporting this prediction. We then turn to a unique survey of founders in crowdfunding in order to examine alternative explanations. We find support for a variety of systematic differences between male and female founders, but the persistence effect remains. While decreased persistence in the face of low quality opportunities benefits women individually, we argue that it disadvantages women as a group, as it leads to 25.3% fewer female-led foundings in our sample than would have occurred if women reacted similarly to men. Abstract: Men are far more likely to start new ventures than women. We argue that one explanation of this gap is that women respond differently to signals of past entrepreneurial success due to the "male hubris, female humility" effect. We argue that as a result women are disproportionately less likely to persist in second founding attempts than men when they have succeeded or failed by large margins. Using a data set of serial founders in crowdfunding, we find evidence supporting this prediction. We then turn to a unique survey of founders in crowdfunding in order to examine alternative explanations. We find support for a variety of systematic differences between male and female founders, but the persistence effect remains. While decreased persistence in the face of low quality opportunities benefits women individually, we argue that it disadvantages women as a group, as it leads to 25.3% fewer female-led foundings in our sample than would have occurred if women reacted similarly to men. * Authorship is alphabetical; both authors contributed equally in developing this paper. The authors would like to thank Adam Cox,
Disciplines
Management Sciences and Quantitative Methods