At diversity conferences, I sometimes hear statistics like, “in the United States, a woman earns roughly 77 dollars for every 100 dollars earned by a man.” As someone trying to be an ally on diversity, I cringe. In my head, I hear all the objections male executives use to ignore fair pay issues. The gap isn’t really that big, they’ll say. More women go into low-paying careers, which is true (though that ignores the possibility those jobs are paid less because women are more likely to work in them). Or they say it’s because women take time out from their careers to raise children, and thus have less work experience, which is more true for women than men.
If you’re going to convince those executives gender is a factor, you have to provide apples-to-apples comparisons, and preferably, an evidence-backed solution. Fortunately, Canada gave researchers a real-world test of the gap and the impact of salary transparency on it. Six of the 10 provinces had passed laws by 2019 requiring universities to publish the salaries of employees earning above certain amounts, starting with three provinces in 1996.
Researchers at the Univ. of Toronto and Princeton used a database of almost all faculty salaries in Canada to compare salaries at the same levels (assistant professor to assistant, associate to associate, etc.), in the same departments, by province, before and after salary transparency was required. Only base salaries were compared. The study also accounted for each person’s age, amount of time at the university, and time since getting their last degree, all of which could have explained a gender gap. More than 50,000 individuals were covered, a quarter of them women over the 20-year study period. (In the last years that had risen to 40%.)
Even with all of those factors controlled, there remained a 6 percent gap at the start of the period that could only be explained by gender. This had dropped to 1 percent by the end of 2018. “A consistent result… is that the pay transparency laws reduced the gender gap,” the scholars said, causing between 20-40 percent of the drop depending on how they looked at the data. Because the study compared data before and after laws were passed, at different times, the “consistent result” makes it likely salary transparency actually caused part of the gap reduction. No other laws affecting the gap were passed, making this a clear test of transparency.
The data gave some hints about how the gap closed:
- Many universities conducted internal studies in response to the laws, to figure out how to address their gaps.
- In the first years after a disclosure law went into effect, men’s average wage growth slowed.
- The data suggested fewer low-level professors might have been hired, leaving more money for increases for females at higher ranks, where the pay gap was higher.
- Unionized universities accounted for most of the gap closure.
- The effect was strongest where the initial gaps were smaller; logically, more actions are required where gaps are large.
Obviously, many of my fellow males will object to slowing their salary growth to close the gap. What my peers are ignoring is the flip-side of the argument: that we were paid too much relative to women with the same credentials and experience for years, so we are losing nothing on balance. On average it only took four years for things to level out.
When I was a middle manager at Los Alamos National Laboratory, all salaries were public knowledge. It made life so much easier for us managers during the annual raise cycle, I cannot fathom why any manager opposes transparency (unless they know they are paid more than they should be relative to workers!). I heard far fewer complaints about pay raises each year than I did later in the private sector.
Not only does salary transparency reduce those manager headaches, we now we have solid evidence it is an effective tool for improving pay equity.
Source: Baker, Michael, Yosh Halberstam, Kory Kroft, Alexandre Mas, and Derek Messacar, Pay Transparency and the Gender Gap (Cambridge, MA: National Bureau of Economic Research, May 2019), p. w25834 <http://www.nber.org/papers/w25834.pdf> [accessed 3 June 2022].