Unionization, Employer Opposition, and Establishment Closure∗

We study the effect of private-sector unionization on establishment employment and survival. Specifically, we analyze National Labor Relations Board (NLRB) union elections from 1981 to 2005 using administrative Census data on the universe of establishments in the U.S. Our empirical strategy extends a difference-in-differences design with regression discontinuity extrapolation methods. This strategy allows us to estimate treatment effects that include elections that win by large margins of support. We show that unionization decreases an establishment’s employment and likelihood of survival. We hypothesize that two reasons for these effects are firms’ ability to avoid working with new unions and managers’ opposition to unions. We test this hypothesis for unionization in manufacturing. There, the negative effects are significantly larger for elections at multi-establishment firms. Additionally, after a successful union election at one establishment, employment increases at the firms’ other establishments. Both pieces of evidence are consistent with firms avoiding new unions by shifting production from unionized establishments to other establishments. Finally, we find larger declines in employment and survival following elections when managers were more opposed to the union. To support this, we estimate treatment effect heterogeneity based on two proxies for managers’ opposition: delays during the election process and the lack of other unionized establishments at the firm. Taken together, our results are consistent with firms’ union avoidance tactics contributing to the overall negative effects of unionization. ∗We are grateful to Daron Acemoglu, David Autor, and Simon Jäger for guidance and advice throughout this project. We thank Josh Angrist, Jon Cohen, David Hughes, Sylvia Klosin, Tom Kochan, Felix Koenig, Mike Piore, Frank Schilbach, Garima Sharma, Martina Uccioli, John Van Reenen, Michael Wong, and Josef Zweimüller and seminar participants at the U.S. Census Bureau for helpful comments. This paper benefited greatly from Henry Hyatt and Kirk White’s data expertise. We thank Stephanie Bailey, Jim Davis, and Nathan Ramsey for their assistance with the data access and the disclosure process. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No. (1745302). All errors are our own. Disclaimer: Any views expressed are those of the authors and not those of the U.S. Census Bureau. The Census Bureau’s Disclosure Review Board and Disclosure Avoidance Officers have reviewed this information product for unauthorized disclosure of confidential information and have approved the disclosure avoidance practices applied to this release. This research was performed at a Federal Statistical Research Data Center under FSRDC Project Number 2389 (CBDRB-FY22-P2389-R9311 and CBDRB-FY22-P2389-R9358). †Massachusetts Institute of Technology. Emails: swang1@mit.edu and sgyoung@mit.edu.