We’ve been asked the following question a few times: is unionizing likely to increase my take-home salary? As academics, our impulse was to answer this question by turning to the academic literature. We’ve provided links to the articles we found which address the effect of academic unions on academic salaries in Canada, and briefly summarized their findings, below. We also included one article about the effect of academic unions on non-salary aspects of institutional decision making, and one article which is a meta-analysis on whether there is bias in the literature on this subject.
The salary effect of faculty unionism in Canada (1995) PDF
This paper finds a 2-3% salary premium from unionization, based on Statistics Canada data from 1972-1991.
Unions without rents (2004) PDF
This paper also finds a small positive effect of unions on salaries in Canada – somewhat larger, around 5%, at medical-doctoral universities such as McGill. It also finds a small negative effect on research output (as measured by papers per year) and a small positive effect on research quality (as measured by average number of citations per paper).
The causal effect of faculty unions on institutional decision making (2013) PDF
This paper isn’t directly about salaries; it finds that “unionization greatly increases faculty influence over institutional decision- making, both in compensation and in areas outside of compensation”
The Impact of Unions on Wages in the Public Sector – Evidence from Higher Education (2025) PDF
This paper studies Canadian data from 1970-2022. It finds that “unionization increases salaries on average by 2 percent in the first year post-unionization and by 6 percent after 6 years”. It finds the largest wage increases for those in the lowest half of the salary distribution, and postulates this is largely due to the introduction of floors on compensation.
Aside: When lower wages increase more than higher wages, so that the ratio of highest to lowest wages decreases, the literature sometimes seems to refer to it as wage compression.
Publication bias in union-productivity research (2025) PDF
This paper does a meta-analysis which shows no significant statistical evidence of publication bias in the union-productivity literature. Quoting the paper:
“This finding of no significant publication bias in the union-productivity literature should be seen against the backdrop of other investigations for publication bias. In all other applications in economics conducted thus far, evidence of publication bias has been claimed (Card and Krueger, 1995; Ashenfelter, Harmon and Oosterbeek, 1999; Gorg and Strobl, 2001; Doucouliagos, 2005). Hence, our finding of no publication bias in this area of industrial relations research is rather exceptional”.
The paper does find some statistical evidence of a small amount of publication bias in U.S. studies; but none in non-U.S. studies.
One Comment