Women, Black, and Latinx workers in retail jobs similar to cannabis retail sales benefit strongly from unions: Union wage premium for proxy cannabis sales workers by gender and race

Specification All Women Workers of color Black Latinx
(1) Basic controls 8.8% 9.0% 12.3% 9.3% 15.4%
(2) With additional demographic controls 7.6% 8.4% 10.8% 9.1% 13.5%
(3) Controlling for industry and occupation (no additional demographic controls) 10.7% 10.2% 13.5% 9.4% 17.4%
(4) Controlling for full demographics, industry, and occupation 9.5% 9.5% 12.1% 9.3% 15.5%

Notes: The union premium is how much more in wages workers covered by a union contract earn than similar nonunionized workers in cannabis retail proxy jobs. All results p-value < 0.01. For a full discussion of the methodology, see the Appendix.

Extended notes: The premium is calculated as exp(b)-1, where b is the coefficient on union status with log of wages as the dependent variable. All models include basic controls for educational attainment, age, full-time/part-time work status, usual weekly work hours, year, and state. Additional demographic controls include gender, race/ethnicity, marital status, and citizenship status. Our preferred estimation (Specification 3) does not include these additional demographic controls. While such demographic characteristics are associated with differences in pay, these differences often reflect labor market discrimination rather than differences in job skills and qualifications. To estimate the union difference for a specific demographic group, the sample is limited to respondents of that group (female workers, Black workers, etc.) before calculating the premium. For a full discussion of the methodology, see the appendix. 

Source: EPI analysis of 2010–2019 Current Population Survey ORG data (EPI 2021a).

View the underlying data on epi.org.