Does who we work with matter? The Effect of Coworkers on Income

I study the role of coworkers by estimating individual-level coworker quality for all Swedish workers. High-quality coworkers increase earnings of those they work with. I develop a novel method combining penalized estimates with group fixed effects in an AKM framework to identify individual and coworker quality. Coworker quality explains about 10% of earnings variations, 2/3 of the variation otherwise explained by firms. I estimate social returns to education using my coworker quality measure with an instrumental variable strategy based on distance to new college openings. Receiving a college degree boosts coworkers' earnings by approximately 8%, implying social returns to a college degree are 20% higher than individual returns alone.

Herstory: The Rise of Self-made Women

We analyze women's status globally throughout history using a new database of seven million notable individuals. We measure women's prominence as their share among the most notable people. Historically, women's status was linked to nepotism, with higher female representation in influential families. Self-made women emerged among writers in 17th century Protestant Europe due to informal education and new public spheres. A broader rise began with the 1800 birth cohort, starting with artists and scholars, then politicians. This trend has persisted, creating cross-country differences despite widespread growth in the 20th century.

The Freezing Point of History: How Adverse Weather Shocks Affect the Emergence of Leaders

We examine how harsh early-life environments affect exceptional outcomes crucial for long-term growth. Using US temperature data since 1790, census information, and measures of historical prominence, we find that severe winters in the first year of life reduce the proportion of a birth cohort achieving advanced education, top-quartile earnings, and historical significance.

Human Biographical Record (HBR)

We present the Human Biographical Record (HBR), a new dataset of over seven million notable individuals throughout history. Built on Wikidata and enhanced with multiple digital sources, HBR uses machine learning and text analysis to extract key biographical information. This chapter details HBR's construction, assesses its completeness and accuracy, and compares it to existing datasets.

Social Inclusion: A Measure of Equality of Opportunity

We introduce "social inclusion" (SI) as a new, measurable concept of equality of opportunity. SI is defined as the equality of outcome distribution given all predetermined characteristics, or the unpredictability of outcomes based on circumstances. We estimate SI in four applications, addressing challenges of over-fitting and specification error using out-of-sample estimation and decision tree learning.

Adoption of Medical Innovations Across Hospitals and Socioeconomic Groups: Evidence from Sweden

We study innovation adoption in healthcare using Swedish data on 58 novel medicines for 47 conditions. We find significant variation in adoption rates across hospitals and socioeconomic groups, with a positive correlation between patient income rank and adoption rates. A case study of an antiplatelet drug suggests that equalizing adoption rates between income extremes could have reduced mortality during the study period.