Research Highlights Featured Chart

June 10, 2026

Unionization and wage inequality

The impact of unionization on the salaries of faculty at Canadian universities.

Tyler Smith

Source: Joe Piette

The decline of organized labor is frequently linked to widening income gaps, yet credible evidence on how unions affect the full distribution of wages remains scarce. It has been difficult for studies to distinguish compositional effects (the systematic differences of workplaces that tend to select into unions) from causal effects and to avoid measurement errors when relying on changes to union status over time.

In a paper in American Economic Review: Insights, Michael Baker, Yosh Halberstam, Kory Kroft, Alexandre Mas, and Derek Messacar address these challenges by using administrative salary records for full-time faculty at Canadian universities between 1970 and 2022. They link these records to data on union certifications and the first union salary contracts and exploit the staggered timing of unionization across universities to trace its effects over time.

Figure 2 illustrates the study's central findings across three panels. The horizontal axis in each panel marks years relative to union certification, with year zero indicating the certification date (dashed vertical lines).

Figure 2 from Baker et al. (2026)

Panel A plots the effect of unions on the average of the logarithm of salaries. Prior to unionization the estimates cluster around zero, indicating that faculty who unionized were not on different salary trajectories prior to unionization. After certification, average salaries rose 2.4 percent in the first year and reached 6.1 percent by year six.

Panel B disaggregates salary effects into two subperiods. The estimates for institutions unionizing between 1970 and 1995 (blue) are roughly one percentage point above the overall average, while estimates for 1996 through 2022 (red) hover near zero and remain statistically insignificant. This panel illustrates that wage gains were largely concentrated in the earlier era.

Panel C displays effects at five different percentiles of the salary distribution, showing that the effects decline at higher percentiles. Gains reached 12.4 percent at the tenth percentile but became indistinguishable from zero at the ninetieth. 

Overall, the findings indicate that unionization raised faculty salaries and compressed their distribution by lifting the lowest salaries. Ultimately, the authors trace this compression to contractual salary floors negotiated by the unions, which appeared in nearly all of the first contracts they examined.