Friday, February 07, 2014

Christina Hoff Sommers: Nope, Women Don't Earn Less Than Men. Part 2 Of 3.

This is the second post about Christina Hoff Sommers' argument that, essentially, lower wages are fair wages for women.  The first post can be found here, the third here.  I realized, after working on this post, that there needs to be a third one.  Otherwise this  post ends up book length.  So this post talks about occupational segregation between women and men, and the third (and final, promise!) post talks about engineers.


Back to Hoff Sommers.  After addressing the question of the gross gender gap in wages and the net gender gap in wages (in my terminology), she moves on to "choice" by focusing on gender segregation* in the labor markets.   For example:

Much of the wage gap can be explained away by simply taking account of college majors. Early childhood educators and social workers can expect to earn around $36,000 and $39,000, respectively. By contrast, petroleum engineering and metallurgy degrees promise median earnings of $120,000 and $80,000. Not many aspiring early childhood educators would change course once they learn they can earn more in metallurgy or mining. The sexes, taken as a group, are somewhat different. Women, far more than men, appear to be drawn to jobs in the caring professions; and men are more likely to turn up in people-free zones. In the pursuit of happiness, men and women appear to take different paths.

Where to begin with that?  Perhaps here:  One cannot just state that "much of the wage gap can be explained away by simply taking account of college majors."  This is because only 33% of  American adults have a four-year college degree or higher, and even if we include two-year college degrees (which Hoff Sommers does not, in her argument), the majority of American adults (61.3% in 2011) do not have a college degree.

Hoff Sommers' other opinions about college majors are the stuff for my third post.  In this one I wish to look at this quick reference** to gender segregation in occupations:

Women, far more than men, appear to be drawn to jobs in the caring professions; and men are more likely to turn up in people-free zones. 
Hoff Sommers doesn't really go on to talk about gender segregation, because she uses her data for that "choice" argument, but gender segregation in the labor market does matter a lot in explaining why full-time female workers earn less than full-time male workers, on average, and it's worth a short (for me) detour to that topic, to explain, specifically, why "people" are not just "drawn" to certain types of professions but might end up in them for all sorts of reasons. 

 Let's begin by noting that a)  female-dominated occupations pay less, on average, than male-dominated occupations, and b)  there are a lot of women in female-dominated occupations and a lot of men in male-dominated occupations, so that even small differences in the average earnings between occupations cause clear differences in the gross gender gap. 

Put these together, and it turns out that occupational gender segregation accounts for a largish chunk of the gross gender gap in earnings.  Or viewed slightly differently, if we could shift enough people about to make all occupations gender-integrated, the gross gender gap in earnings would be much reduced.

Could that be achieved?  It's hard to say.  The conservative argument, which Hoff Sommers presents in her opinion piece, is that occupational segregation is based on men and women wanting different things in life and on choosing different things when deciding on an occupation.   Usually this "choice" argument is firmly based on the "no-choice" argument of innate differences.  For example, Hoff Sommers argues that women prefer jobs having to do with people, men prefer jobs having to do with things, on average, and that it is these differences, seen as innate and not susceptible to choice, which drive those "choices."

I discuss the problem with the "choice" argument in the previous post a lot so won't bore you with it here again.  But it's worth being reminded that occupational "choice" certainly depends on what else people are expected to do with their time (mind children?  give birth? be able to financially support a large family alone?) during their lives and that it also depends on how well the labor markets have adjusted to letting people carry out  tasks such as child-rearing and the care of elderly relatives while also remaining full-time workers.

That people like different things is obviously true, and it is not impossible that women and men, on average, prefer different types of jobs, either because of innate differences or because of cultural expectations or both.  On the other hand, what we prefer is so intertwined with the society we grow up in that I'd be very careful to simply assume away alternative explanations, at least as a supplement to the "choice" argument.   Have a look at this picture of young American white women somewhere around 1915:

The way they are dressed and the way their hair is done is based on choice.  Nobody probably forced them to wear those specific outfits or hairstyle.  But the way they are dressed is clearly also based on the society in which they were living and their place in it.  The clothing reflects the norms about how women should dress during that time period, the hairstyles reflect the Gibson Girl fashion concept.

Apply this thinking to occupational choice, and consider a culture in which certain jobs are regarded as appropriate for women, certain jobs as appropriate for men.  Those expectations frame what people "choose."  Sure, other "choices" are not illegal, and, sure, there's almost always some scope for individual choice, but in the longer term sense we have all received a lot of invisible training about gender roles and the demands each gender faces, and by the time we make a job choice that training has probably become almost impossible to separate from some innate form of preferences.

On top of all that, historically (at least) occupational gender segregation has also been shown to be linked to direct and explicit discrimination.  Until the 1970s or so, advertisements seeking workers were legally allowed to be separate for men and women.  Women and men wanting to work for the AT&T were once given separate leaflets about available jobs, and the leaflets listed different jobs.   Career guidance counselors routinely made different suggestions to girls and boys, and so on.

Occupations can also be kept gender-segregated by those who don't wish to see increased competition.  For instance, the first women entering a traditionally male occupation are both visibly "different" and a sign of possible extra competition which could lower the average wages, should a lot more women follow the first few.  One response to that can be sexual harassment or other treatments which make the job less appealing to "the others."  That women look different from other new entrants in such a field make them a clearer target for resistance.  Finally, it can be very hard to be the "token woman" or "the token man" at one's job.  That can cause social isolation, and it can make the "token person" regret choosing the field.  It could even make leaving the field more likely, even if the person's initial preferences were for that particular job. 

The point I'm making above is that there are more complicated reasons for occupational gender segregation than simply the idea that women and men might prefer different types of jobs, on average, or even the argument that women and men are just good at different kinds of jobs or both.

 As partial evidence of that, note that there are occupations which  have changed from being almost completely male to currently being fairly gender-integrated*** at the entry level (such as medicine, at the current time)****, occupations which have changed from all-male to predominantly female (veterinarian science now and typists a long time ago), and also note that computer science in the early 1980s was a pretty gender-integrated field (at entry level) but has since become increasingly male-dominated.

Those types of changes are not credible evidence for nothing else explaining occupational gender segregation except for some stable preference difference between the average man and the average woman.

Two further questions matter when we talk about occupational gender segregation.  The first is to simply ask why the female-dominated occupations pay less, on average, than the male-dominated ones.  I'm not going to try to answer that question here, assuming that I could find all the necessary research, but one theory worth thinking about is that women tend to "crowd into" a relatively small number of occupations, for whatever reason*****, and that it is the large supply of willing female workers in these occupations which depresses the average earnings.

The obvious corrective for this is to encourage women to consider a wider range of jobs when they are entering the labor market or thinking about what to study, and more importantly, to consider earnings potential of the different jobs not from the angle of being able to support just one person but from the angle of being able to support a family. I also recommend training in how to negotiate one's earnings (when one has negotiating power) and working to unionize industries.

The second (and final) question about occupational gender segregation has to do with resegregation in the opposite direction:  The idea that a predominantly male/female occupation over time might move towards integration, reach it, but instead of stopping continue changing until the job ends up predominantly female/male.

Some economists argue that there is a tipping point, a particular relative size for the increasing minority group of workers, which causes the previous majority group to start moving out of the job.  In the case of gender segregation, this theory is based on the salaries and status of jobs getting lower because more women enter them. The drop in both status and pay make these jobs then less appealing to men (who can do better elsewhere and/or who are faced with the role expectation of being able to bring home most of the bacon).

I don't know how common resegregation might be in the labor markets, but if it is common, then the lower earnings of female-dominated occupations cannot be fixed by encouraging more women to leave them and more men to enter them.  All this obviously relates to the more fundamental question:  Are women paid less because so many of them work in female-dominated occupations or are those occupations paid less because it is largely women who work in them?

*The term gender segregation in this context doesn't tell anything about the reason for a pattern where many occupations are mostly male and many occupations mostly female.  The segregation could be voluntary or it could be legally required (as is the case in terms of some occupations in countries such as Iran) or it could be based on social and cultural values and sanctions or a complex mixture of all these reasons.

**You want to be careful with just accepting the idea that women are drawn to caring professions and men prefer people-free zones.  The military is not a people-free zone, neurosurgeons are in a caring profession, most politics requires people-skills, ministers/priests/rabbis/mullahs have a lot of contact with people and are often expected to care about their problems and so on.

***A reverse pattern, though much weaker, can be seen in nursing which more men are now entering than in the past. The traditional pattern (of professions becoming female-dominated) can also be seen in psychology.

****I define a fairly integrated occupation one in which the proportion of the less represented gender is at least 30%.  I made that one up (though a proper definition no doubt exists), because it might match with the critical coalition concept which shows that women or men in a job are no longer seen first as representatives of their gender but as individual workers when their overall percentage reaches roughly 30%. 

*****Those reasons could include what jobs a society deems suitable for women vs. men (the early "suitable" jobs for middle-class women with education were secretarial work, nursing, teaching and social work), a pattern of work which is more compatible with child-rearing duties, such as more flextime or shorter working days, preference or taste differences, whether innate or not. 

The reasons could also include discrimination against women who enter male-dominated occupations and tactics like sexual harassment in those, because they increase the psychological costs of working in those jobs and thus make the traditional pink collar occupations look better.

That last sentence is based on the assumption that people judge jobs as a package deal, where what matters is not just the money earnings one gets.  But ultimately how much or how little one is paid should matter in the following economics sense:  If women keep crowding into a small number of occupations the average earnings in them will keep declining.  However much the workers in that job might prefer it to other alternatives, there should be a point at which women start moving out of that job and into other jobs, even if they are less desirable, as long as they are better paid.  Because of the package deal aspect.  When this movement doesn't seem to happen, economists look for any obstacles which keep women away from the better-paying fields.