Deaths, Injuries and the the Gender Gap in Pay

by Mike Kimel

Deaths, Injuries and the the Gender Gap in Pay

A few days ago, Glassdoor.com released a report (Demystifying the Gender Pay Gap) looking at the gender gap in pay, which they claim is about 24%. That is, women make about 76 cents for every dollar made by men. Glassdoor also indicates that adjusting for education, years of experience, location, job title, and job type, the gender gap in the US drops to about 5.4%, which means adjusting for the aforementioned factors, women in the US make about 94.6 cents for every dollar earned by men.

Figure 1. Glassdoor’s Gender Gap

Kimel 1.   Glassdoor Pay Gap by Gender

I skimmed through the report, and two obvious things missing (admittedly hard for Glassdoor.com to capture) from the analysis are death and injury. Men are more likely to get injured or even die on the job. This is true even for the same job.

For example, consider the most white collar office workplace with which you are familiar. Every so often, it is decided that some file cabinet, or some other heavy equipment has to be moved, and it just isn’t worth the requisition or the wait to have someone come up from Maintenance and move it. Regardless of the gender composition of the hyper white-collar workers at the workplace, what is the gender composition of the subgroup that moves the stuff? Because in the one time out of a thousand that something falls on someone, that’s the gender composition of those who will get hurt or worse.

For what its worth, I’ve been doing white collar work for close to twenty years in a variety of locations, and I’ve been one of the designated movers of some heavy items at least once every year in that career. I’ve worked at places where just where women outnumbered men and just about everyone was a self-described feminist, and places where men outnumbered women and held attitudes reminiscent of the 1950s. Without exception (and I do mean without exception), when it came time to move heavy stuff, the gender profile of those doing the moving of the stuff was exactly the same.

Enough anecdotes. Let’s try to quantify this differential in injuries and deaths. We can start by looking at the number of men and women in the workforce, which is among the really cool data available in Figure 2 below:

Figure 2

Kime 2.  Earnings by Gender and Ethnicity (BLS)

Figure 2 copied from: http://www.bls.gov/opub/reports/cps/womensearnings_2008.pdf

Figure 3 shows deaths for injury at work for various years through 2009. However, the 2009 data is incomplete, so we’ll use the 2008 data.

Figure 3.

Kimel 3 Number of Deaths for Injury at Work

Figure 3 copied from: http://www.cdc.gov/nchs/data/nvsr/nvsr60/nvsr60_03_tables.pdf#tabs10,

What are deaths from injury at work? Well, from a different year (2011), those looked like this:

Figure 4

kimel4

Figure 4 from: http://www.bls.gov/iif/oshwc/cfoi/cfch0010.pdf

Divide the number of employed by the number of fatalities from workplace injuries and you get a death rate from work, which for women is 7.9 * 10^-6, and for men is 7.3 * 10^-5. Put another way, work injuries are 9.2 times more likely to claim the lives of men than of women.

But… this isn’t quite right. Take a look at the graph below:

Figure 5.

Kimel 5 Fatal Injury Events by Gender

Figure 5 from: http://www.bls.gov/iif/oshwc/cfoi/cfch0010.pdf

According to this graph, death by homicide is more likely for women than for men, but according to the next graph below, from the same source, a 38% of work related homicides of women involve a relative or a domestic partner. For men, that figure is 2%.

Figure 6

Kimel6 Work Related HomicideBy Gender and Assailant Type
A this point, I believe a rational person has to take issue with the data. Being murdered by a relative while at work is no more a “work injury” than dying in a terrorist bomb blast while eating a meal at a restaurant can be considered food poisoning.

So… the number of homicides of women has to be reduced by 38% (i.e., percent of homicides of women while at work that were committed by a relative or domestic partner) of 20% (i.e., percent of women dying of homicide), or about 14.44%. By contrast, for men, the equivalent is 0.18%.

Reducing the 2008 deaths of women at the office by 14.44%, and of men by 0.18%, gives us an adjusted work-induced death rate of 6.76 * 10^-6 for women, and 7.25 * 10^-5 for men. (Yes, I know we’re mixing years, 2011 and 2008, but presumably the percentages are very similar.) Put another way, on average, a working man is 10.7 times more likely to be killed by his job than a working woman is to be killed by her job.

But I don’t think this is the whole story. After all, the injury related deaths don’t include heart attacks, strokes, and other deaths that occur at the office, and a) at least some portion of those heart attacks and strokes are truly work-induced, and b), anecdotally, at least, it would appear more men die from from these causes at the office than women. But that is hard to quantify.

I would also note, however, this (Occupational injuries, illnesses,
and fatalities among women
):

WOMEN HAVE EXPERIENCED far fewer occupational injuries
and illnesses than their hours worked would suggest. Even more
disparate in relation to employment hours was the female share
of occupational fatalities. Female injuries, illnesses, and fatalities
are not only disproportionately low; they also differ from male
cases qualitatively.

This post is starting to get long, and I am nowhere close to the punchline yet, so let me end by saying I will follow up this post with an estimate of how the cost of deaths and injuries relates to the gender pay gap.