Union wage effects in Sweden: Evidence from the interwar period

Abstract

In this paper, I use a new plant-level dataset to investigate the relationship between wages and the regional strength of unions. Using a shift-share or ’Bartik’ instrumental variables approach, I disentangle the causal effect of union strength on wage levels. I find statistically significant and economically substantial, heterogeneous union wage effects for men with the bottom of the distribution of plants being impacted by union density and the top two-thirds being unaffected. I find a statistically weaker negative effect on wages for women and argue that unions, in general, were uninterested in the issues of women. The paper contributes to the literature by providing the only evidence of a union wage effect in Sweden and, the earliest identified union wage effect anywhere—highlighting the importance of unions in shaping labor market outcomes in the early 20th century and showing that union wage effects are products of their historical and institutional context.