Under the skin of the gender pay gap

Googling “gender pay gap”, it returns 226.000.000 results. The topic is not new, yet as persistent as an athlete’s foot. Probably it is the same issue like with the fungus, trying to do the symptom curing will not eradicate the root-cause. We need to get under the skin.

“Despite a long record of research on the sources of the gender wage gap, a large fraction of gender wage differences remains unexplained. In this paper, we propose gender differences in social capital as a novel explanation for the gender wage gap.” British Journal of Sociology, Collischon, Eberl, 2021

Well, I have been stealing it, but …

One of the?value-creating use cases of Human Network Analysis?is pattern detection. One of the patterns we had already identified was the buildup of Social Capital over time. Another pattern that easily comes to mind is Social Capital differences between groups. Which groups? For example, males and females.

Let’s put the puzzle together:

Recently we have shown that higher Social Capital correlates with better performance ratings (Bridgebuilders, a new species), and from better performance ratings it is just a tiny step towards better pay (at least in theory, hahaha).

What does the data tell us about the Social Capital of males and females?

Males have a higher Social Capital in Ericsson than females*.

?

The over-representation of males becomes apparent when filtering for higher and higher Social Capital (with roughly 25% females in the company, one would expect in equal representation 1 pink dot for 3 green dots, which is definitely not the case for higher Social Capital, reading the picture below from left to right):

And the same disproportion we detect when we look at those who are sticking out with their Social Capital (the long tail of the power-law distribution, the bridgebuilders (1,2,3).

Females are less represented proportionally in the bridgebuilders with extraordinary collaboration importance than males.

Can we understand the difference in Social Capital? There is a blog post yet to come: Twin Talk. Twin Talk summarizes the observation that we talk over-proportionally to those who are alike. We observe Twin Talk in various dimensions, relevant here: Twin Talk applies to genders: ?Women talk to women, men talk to men (over-proportionally).

Or as McKinsey* puts it:

We seem to see this in Ericsson, when looking at where the discrepancy in Social Capital between males and females appears in the hierarchy. While on lower hierarchy levels there is not a clear difference visible, the scissor opens on higher hierarchical levels***.

Another indication of the reasons for the Social Capital gap we see, when at various tenure groups:

The biggest gap appears in the tenure range of 5-10 years. When checking with India, Sweden and China, the 3 countries with the largest Ericsson employee base, it is in this tenure window, when on average women take maternity leave.

In summary we see that Human Network Analysis and Social Capital can be a very helpful concept in the analysis of DEI (diversity, equity & inclusion) on gender. We are just at the beginning.

?

?

*Appendix:

Social Capital is short hand for the Human Network Analysis metric betweenesscentrality, which is applied to the collaborative meetings, which serve as a proxy for collaboration. Of course we are aware of the simplification this model applies.

?

**Appendix:

It is well known that there are other parameters that influence the Social Capital: Tenure, hierarchy, functional area. When comparing males and females these other parameters were kept comparable, that is males and females groups have been showing the same distribution in tenure, hierarchy and functional area.

***Appendix:

Lx denote the hierarchical layers in the company, where x denotes the distance from the CEO, direct reports to the CEO are L1, their reports L2 ... the higher x the further away hierarchically from the CEO.