Estimating Effects of Wages on Smoking Prevalence Using Labor Unions as Instrumental Variables

Abstract

Objectives To test for the effects of wages on smoking using labor unions as instrumental variables. Methods We analyzed four waves of the Panel Study of Income Dynamics (2013 to 2019 alternate years). The overall sample included workers aged 18 to 70 years in 2013 and subsamples within blue + clerical/white-collar and private/public sector jobs (N = 37,117 to 8446 person-years). We used two instrumental variables: worker's union membership and states' right-to-work laws. Results $1 (2019 US dollars) increases in wages-per-hour resulted in 1.3 (P < 0.001) percentage point decreases in smoking prevalence (8.2% decreases at the smoking mean). Larger effect sizes and strong statistical significance were found for blue-collar + clerical and private-sector subsamples; smaller sizes and insignificance were found for public-sector and white-collar subsamples. Conclusions Unions increase wages, and higher wages, in turn, reduce smoking. Wages and labor unions are underappreciated social determinants of health.