I love statistics and so can you! The gender pay gap edition.
Last week I wrote about the gender pay gap — that is the fact that women make less than men. There are also racial components to this pay gap, with Latina women earning less than half of what men make (only 48 cents to the dollar) and black women making just 63 centers for every dollar men make. The gender pay gap in Georgia is 88 cents to the dollar, across all racial groups.
Someone in the comments asked: “Could you expand upon the gender wage gap? Is it just a man to woman comparison? Or is it a comparison of men and women with equal degrees or experience and doing the same type of job?”
This is a common question.
So how does the data work and what is true?
It turns out all these factors matter, but even when women have the same degrees and the same job, they are still paid less. Human capital variables — like education and years of experience — as well as industry and occupation explain some of the gender pay gap, but not all of it. There’s something else going on, and that something else is gender discrimination in pay. What’s more, these other variables are all influenced by gender (and race for that matter), thanks to lots of different subtle and unconscious ways bias and discrimination work.
Francine Blau and Lawrence Kahn are some of my favorite labor economists. They developed a really robust model looking at gender and pay, one that they’ve continued to update over the years, keeping both the data and the model relevant as the economic theory changes.
Blau and Kahn estimate that 8 to 18 percent of the gender pay gap is unexplained by factors like experience, occupation, education and industry, leaving them to conclude that gender discrimination plays a role. And while that discrimination factor had been shrinking over the past three decades, it’s not really shrinking anymore.
Gender and gendered expectations also drive what kinds of social and academic behaviors are recognized and rewarded (and so what skill sets people develop before going into the workforce), what kinds of fields people are encouraged to go into and who is expected to take time off from work to care for kids. This is also all very well-documented.
So not only are women paid less, even when accounting for these other factors, but women face barriers to getting the same education and experience as men. Bias and discrimination against women leads them to have lower access to high-paying jobs, and to be paid less when they do get these jobs. Bias and discrimination also impact pay for people of color, with women of color facing disparities due to both gender and race.
All the studies and reports linked throughout this post use different methodologies. But I particularly appreciate that Blau and Kahn use really solid data and robust modeling, so that 8 to 18 percent (of the gender pay gap is due directly to discrimination) figure is very reliable.
Blau and Kahn use two large data sets, one maintained by the University of Michigan and another from the U.S. Census Bureau and the U.S. Bureau of Labor Statistics. The idea is, if enough people’s data is included, researchers can make generalizations about what the data shows for the general population. (There are also statistical tests researchers do on data to make sure it represents enough people, and that it doesn’t fall prey to other problems).
Blau and Kahn look at human capital variables, like number of years in the workforce and education, as well as the industry and occupation for each person. They use these variables to figure out which ones are most likely responsible for the gender pay gap that we observe, and how responsible they are.
And the answer is: all of them. All of these variables have an impact and discrimination is happening.