Selected Work-in-Progress

New Employer Taxes and Entrepreneurship: Evidence from Unemployment Insurance,
with Audrey Guo
Are Immigrant Entrepreneurs Magnets for Foreign Investors?,
with Mahdi Eghbali and Livia Yi

Working Papers

Entrepreneurial Spillovers Across Coworkers

Tracking the employment and entrepreneurship of over forty million Americans, I find strong evidence of entrepreneurial spillovers across coworkers. An employee whose coworkers have more prior entrepreneurship experience is more likely to become an entrepreneur themself within the next five years. These spillovers are strongest across coworkers with similar jobs and demographics. Furthermore, an individual is more likely to become a successful entrepreneur if those coworkers were themselves successful entrepreneurs, suggesting spillovers also can affect productivity. Finally, while these spillovers are potentially accessible to people of different backgrounds and resources, these spillovers are primarily at play for the traditionally represented groups in entrepreneurship: White and Asian men.

Supplementary Material and Media

Online appendix
Jordi Blanes i Vidal's The Visible Hand podcast
Matt Clancy's New Things Under the Sun #1 and #2

Pay, Productivity and Management. [NBER #29377],
with Nicholas Bloom, Scott Ohlmacher, and Cristina Tello-Trillo

Using confidential matched employer-employee earnings data from the U.S. Census Bureau, we find that employees at more productive firms and at firms with more structured management practices have substantially higher pay, both on average and across every percentile of the pay distribution. This pay-performance relationship is particularly strong amongst higher paid employees, with a doubling of firm productivity associated with 11% more pay for the highest-paid employee (likely the CEO) compared to 4.7% for the median worker. This pay-performance link holds in public and private firms, although it is almost twice as strong in public firms for the highest-paid employees. Top pay volatility is also strongly related to productivity and structured management, suggesting this performance-pay relationship arises from more aggressive monitoring and incentive practices for top earners.

Other Versions and Media

CES working paper
Harvard Law School Forum on Corporate Governance
Harvard Business Review

Peer-Reviewed Publications

The Slow Diffusion of Earnings Inequality. [Official Link, PDF, NBER #30977],
with Isaac Sorkin, 2023, Journal of Labor Economics.

Over the last several decades, rising pay dispersion between firms accounts for the majority of the dramatic increase in earnings inequality in the United States. This paper shows that a distinct cross-cohort pattern drives this rise: newer cohorts of firms enter more dispersed and stay more dispersed throughout their lives. A similar cohort pattern drives a variety of other closely related facts: increases in worker sorting across firms on the basis of pay, education, and age, and increasing productivity dispersion across firms. We discuss two important implications. First, these cohort patterns suggest a link between changes in firm entry associated with the decline in business dynamism and the rise in earnings inequality. Second, cohort effects imply a slow diffusion of inequality: we expect inequality to continue to rise as older and more equal cohorts of firms are replaced by younger and more unequal cohorts. Back of the envelope calculations suggest that this momentum could be substantial with increases in between-firm inequality in the next two decades almost as large as in last two.


Bureau of Labor Statics' Beyond BLS Series

The Impact of Homelessness Prevention Programs on Homelessness, 2016, Science, 353(6300), 694-699,
with Bill Evans and Jim Sullivan

Despite the prevalence of temporary financial assistance programs for those facing imminent homelessness, there is little evidence of their impact. Using data from Chicago from 2010 to 2012 (n = 4448), we demonstrate that the volatile nature of funding availability leads to good-as-random variation in the allocation of resources to individuals seeking assistance. To estimate impacts, we compare families that call when funds are available with those who call when they are not. We find that those calling when funding is available are 76% less likely to enter a homeless shelter. The per-person cost of averting homelessness through financial assistance is estimated as $10,300 and would be much less with better targeting of benefits to lower-income callers. The estimated benefits, not including many health benefits, exceed $20,000.

Supplemental Material and Media

Materials and Methods Supplementary text
Data and Programs
Media coverage by the University of Notre Dame>, "What Would You Fight For?" series
Media coverage by Science Magazine
Media coverage by WGN Radio, Chicago

Previous Work