Research
Research
How Much Do Workers Benefit from Labor Unions? Evidence from Union Formation Elections (with David Wasser)
Abstract: This paper estimates the causal effects of newly formed labor unions on worker outcomes using a difference-in-differences design that compares workers in establishments where unions won or lost their elections. Using linked records from the Census Bureau’s Longitudinal Employer–Household Dynamics program and the National Labor Relations Board, we construct a panel of more than two million incumbent workers employed at over 4,000 establishments that held union elections between 2002 and 2010. We track these workers beginning four years before the election to ten years afterwards, across jobs, industries, and states. We find four main results. First, incumbent workers in newly unionized establishments do not experience long-run gains in earnings or employment. Second, the effects of new unions on wages are heterogeneous but rarely positive: we find no large “winners” and some evidence of losses among workers in small or goods-producing establishments. Third, new unions do not increase the likelihood that workers remain at the establishment. Fourth, we show that union organizers tend to target establishments already on high-growth trajectories, and that post-union changes in hiring and productivity may dampen wage gains. Pre-election establishment characteristics explain the entire observed wage gap in our data, implying limited causal benefits of new unionization in today's U.S. economy.
Changes in the relationship between income and life expectancy before and during the COVID-19 pandemic, California, 2015-2021. Schwandt, H., Currie, J., Von Wachter, T., Kowarski, J., Chapman, D., & Woolf, S. H. (2022). JAMA, 328(4), 360-366.
Abstract: This retrospective analysis of census tract–level income and mortality data in California from 2015 to 2021 demonstrated a decrease in life expectancy in both 2020 and 2021 and an increase in the life expectancy gap by income level relative to the prepandemic period that disproportionately affected some racial and ethnic minority populations.
Interview: Capital Public Radio (July 2022): Segment on Covid Life Expectancy Research
The Failure of Life Expectancy to Fully Rebound to Prepandemic Levels. Schwandt, H., Currie, J., von Wachter, T., Kowarski, J., & Woolf, S. H. (2025). JAMA.
Abstract: This study found that California life expectancy remained lower in 2024 than in 2019, due primarily to non–COVID-19 causes. Although a prior analysis of California data reported that the life expectancy–income gradient increased during 2020-2021, by 2024 the gap between Q1 and Q4 returned to prepandemic levels. Life expectancy in the Black population was much lower than in other racial and ethnic populations, and Black and Hispanic populations experienced the largest deficits.
Media Coverage: New York Times (07/09/25): Life Expectancy in California Has Not Rebounded After Covid
The Role of Firm Turnover in Private Sector Union Decline: 1948-2019 (Revise and Resubmit at the ILR Review)
Abstract: From 1948 to 2019, the share of unionized US private sector workers fell from 35% to 7%. This paper examines a potential explanation for this decline: the dissipation of exceptionally high WWII Era union formation due to the entry and exit of firms. US labor unions exist within firms. To maintain high union membership share in a firm-based union system, new firms need to be unionized each year to compensate for the entry of new non-union firms. I find that union formation has been insufficient to compensate for non-union entry since at least 1949. Using a novel accounting framework and dataset, I estimate that almost half (47%) of labor union decline between 1948 and 2019 is explained by the entry of new firms in the US economy. Firm entry is more important than post-1949 declines in union formation rates (43%) or low employment growth rates of union firms (10%).
The Distribution of Rents from Employer Mergers (with Ellie Prager, David Wasser, Kevin Rinz, and Wes Yin)
Abstract: We study the extent to which merger-induced gains in product market power are passed through to workers as a function of their skill level and bargaining leverage. Using external data on mergers, we document that mergers defined based on firm IDs in US Census data suffer from meaningful measurement error. We then use our improved measure of mergers to examine effects on workers.
Labor Spillovers of Environmental Regulations: Evidence from the 2004 CAA expansion (with David Wasser)