What is the gender pay gap?

The gender pay gap - like any other pay gap - is the difference between what a specific gender group earns relative to another. To put it simply, this can be calculated as follows:

So far, little to no improvement has been made in this area. In 2002, it took 52 extra days a year for a female to make the same amount of money as their male counterpart. 20 years later, females still need 47 extra days to make the same amount a year as males. 5 days less of extra days of work in twenty years is practically negligible.

Recently, President Biden posted a letter he received from a child in his Twitter profile.

And then he received the following response:

We will never know if a real kid asked for the gender pay gap, nor whether Biden is effectively worried about it. But we know that the gender pay gap exists (see here, here, here and here) and we will attempt to explain the issues behind the response that the President of the United States received, along with some extra considerations.

Issues behind the Pay Gap and the way to measure it

• Opportunity cost discussion: the opportunity cost approach usually revolves around domestic work. The backbone for this discussion stems from the claim that yes, women do earn less than men because of motherhood (i.e. women stay home to take care of the children and to run any house chores so men can focus on getting promotions at work, bringing more revenue to the family) and this is somehow fair. Yet again, the idea is ridiculous from any possible point of view. Yes, the forgone income from the “domestic” parent can be offset by the extra income that the working parent makes but it is also true that there are no extra benefits that can offset the forgone opportunities for the parent that stays home. Besides, why on Earth are these roles women-only? If this was a fair claim, then there should be some sort of benefit that rewarded the stay-home parent in at least the monetary value of their forgone work opportunities. And that does not seem to be happening anytime soon.  
  

• Adjusted versus unadjusted pay gap: we can think of two ways of understanding the gender pay gap

1. Unadjusted gender pay gap: the average difference in earnings between females and males, as described above.
2. Adjusted gender pay gap:  the average difference in earnings between females and males when extra variables are considered in the analysis, such as Job Satisfaction, Performance, Education, Age and so on.

Yet we know that we shouldn’t control for the variables that you are trying to measure. Therefore, to address @MattWalshBlog point on the gender pay gap, we can definitely tell them that if one controls for a variable that is expected to explain something the regression might not be reliable or the variable gender might not be statistically significant. Yet this does not mean that the gap does not exist.

• Multicollinearity: this is a statistical concept where several independent variables in a model are correlated. Multicollinearity among independent variables will result in less reliable statistical inferences. Therefore, running simple regressions to capture the effect of gender on salary hikes or average income, might not be the best way to assess whether sexual orientation or gender identity plays a role in how salaries are distributed within an organisation. Let’s think of a practical example: according to recent research, women all over the world are underrepresented in high-profile jobs, which tend to be better paid. Therefore, being a woman determines seniority (correlation can be established between these two variables) and any regression that incorporates these two variables might not be reliable.

• “If you can't measure it, you can't manage it”: following our reasoning behind multicollinearity, we might come across specific variables that correlate with gender or other independent variables but that we can’t properly measure. For instance, working conditions usually affect an employee differently, whether they are males or females. If it is not possible to properly capture these variables, then our model won’t conclude solidly on them. Gender might not seem statistically significant to determine salary hikes, yet this could happen due to all of these variables that were left behind.

What is Equality Group doing about it?

Recently, Equality Group (EG) published their 2023 PE & VC Inclusive Index Score called HONORDEX. This index measures the position of each company within a rank of scores, according to a specific Equality, Diversity and Inclusion (EDI) framework. This means that every year, EG scores an average of 300 firms according to their performance on EDI, considering different areas such as Working Conditions, how Inclusive the work teams are, their EDI approach on public digital channels, and their EDI Leadership support, and critically pay equity data. In other words, EG is making the usually “unmeasurable” variables measurable, providing a quantitative framework for qualitative insights on EDI.

This year, we have also published our own pay gap analysis at Equality Group. By running anonymous surveys across the team, our firm was able to identify that there is a gender pay gap of 0.2% which becomes statistically insignificant when we analyse a set of other relevant variables. Although we are well below the UK government’s suggested threshold of 250 or more employees to report these findings, we believe that showing our gender pay gap data will pave the way for smaller companies to follow our steps. This will, in turn, allow them to identify gender-caused pay disparities and address the matter properly.

References

1. Economic inequality by gender, Ortiz-Ospina & Roser (2018)

2. What is the gender pay gap and is it real?, Gould, Schieder & Geier (2016)

3. Blau, Francine D., and Lawrence M. Kahn. 2017. "The Gender Wage Gap: Extent, Trends, and Explanations." Journal of Economic Literature, 55 (3): 789-865

4. The Enduring Grip of the Gender Pay Gap, Kochhar (2023)