We study the role of diversity and performance in the entrepreneurial teams. We exploit a unique dataset of MBA students who participated in a required course to propose and start a real microbusiness that allows us to examine horizontal diversity (i.e., within the team) as well as vertical diversity (i.e., team to faculty advisor) and their effect on performance. The design of the course allows for identification of the causal implications of horizontal and vertical diversity. The course was run in multiple cohorts in otherwise identical formats except for the team formation mechanism used. In several cohorts, students were allowed to choose their teams from among students in their section (roughly 90 students). In other cohorts, students were randomly assigned to teams based upon a computer algorithm. In the cohorts that were allowed to choose, we find strong selection based upon shared attributes. Among the randomly-assigned teams, greater diversity along the intersection of gender and race/ethnicity significantly reduced performance. However, the negative effect of this diversity is alleviated in cohorts in which teams are endogenously formed. Finally, we find that teams with more female members perform substantially better when their faculty section leader was also female. Because the gender of the faculty section leader is exogenous to the gender make-up of the entrepreneurial team, the positive performance effects can be interpreted as causal. These findings suggest that diversity policies should take adequate consideration of the multiple dimensions of diversity.

This paper shows that local productivity spillovers propagate throughout the economy through the plant-level networks of multi-region firms. Using confidential Census plant-level data, we show that large manufacturing plant openings not only raise the productivity of local plants but also of distant plants hundreds of miles away, which belong to multi-region firms that are exposed to the local productivity spillover through one of their plants. To quantify the significance of plant-level networks for the propagation and amplification of local productivity shocks, we develop and estimate a quantitative spatial model in which plants of multi-region firms are linked through shared knowledge. Our model features heterogeneous regions, which interact through goods trade and labor markets, as well as within-location, across-plant heterogeneity in productivity, wages, and employment. Counterfactual exercises show that while knowledge sharing through plant-level networks amplifies the aggregate effects of local productivity shocks, it widens economic disparities between individual workers and regions in the economy.

How do geographic and historical-cultural factors shape new business formation? Using novel data on new business registrations, we document that 75% of the variation in new business formation is explained by time-invariant county-level factors and examine the extent to which such variation is driven by historical, cultural, and geographic factors. Current-day new business formation is positively related to historical attributes that presage individualist culture: frontier experience and historical birthplace diversity, as well as the county’s topographical features. The relation holds when we exploit plausibly exogenous variation in frontier experience driven by shocks to the settlement process that arise from historical immigration flows. Our study points to the fundamental role of geographic and historical-cultural features, especially rugged individualism, in explaining contemporary new business formation in the U.S.

This paper studies how demand for labor reacts to financial technology (fintech) shocks based on comprehensive databases of fintech patents and firm job postings in the U.S. during the past decade. We first develop a measure of fintech exposure at the occupation level by intersecting the textual information in job task descriptions and fintech patents. We then document a significant decline of job postings in the most exposed occupations, and an increase in industry as well as geographical concentration of these occupations. Firms resort to an upskilling strategy in face of the fintech disruption, requiring “combo” (finance and software) skills, higher education attainments, and longer work experiences in the hiring of fintech-exposed jobs. Financial firms and those with high innovation outputs are able to offset the disruptive effect from the fintech shock. Among innovating firms, however, only inventors (but not acquisition-driven innovators) experience growth in hiring, sales, investment, and enjoy better returns on assets.

This paper studies how entrepreneurs form new beliefs after making forecast errors. I use survey-based micro data that are representative of the population of French entrepreneurs, and I find that 21% of entrepreneurs make optimistic errors, while 36% make pessimistic errors, suggesting that a minority of entrepreneurs are initially well-calibrated. Although optimism and pessimism are persistent types over time, I show that the likelihood of making errors declines within individuals over time. After overestimating their development and hiring prospects, optimistic entrepreneurs revise their beliefs downward, whereas pessimistic entrepreneurs, who underestimate their prospects, revise upward. The evidence is consistent with entrepreneurs who learn from their past errors. In addition, the ability to correctly forecast sales and employment and revising beliefs are correlated with better performance and growth, and even more so for entrepreneurs who started with pessimistic beliefs.