Alex Tabarrok points to a paper by Uri Gneezy, Muriel Niederle, and Aldo Rustichini showing that although women solve a particular class of problems about as well as men on average, men improve their scores more than women when there is a winner-take-all tournament. The authors believe that this shows that men are more competitive than women.
I think that there is something to this. I remember noticing at the very first get-together of (mostly male) first-year graduate students in economics how the guys all seemed to be trying to impress and intimidate one another. In business meetings later in my career, I observed the same phenomenon–women tended to be task-focused while men seemed to be more concerned with showing who was on top. Particularly traumatic for me was a meeting to plan a merger between two small businesses. The testosterone contest between my CEO and the other company’s CEO was disgusting.
Tabarrok sees this as a potential explanation for why most CEO’s are male, since the CEO is chosen in what resembles a competitive tournament. That may be right, although I have always believed a different explanation. The variance of the distribution for men in such characteristics as IQ tends to be higher than that in women. If you have a tournament with one winner, you are selecting one end of the distribution, which is likely to include a greater percentage of males than the middle of the distribution.
For Discussion. How could one test the various possible explanations for the high proportion of male CEO’s?
READER COMMENTS
Jacqueline
Apr 25 2004 at 1:36pm
How much of it is just that women take lots of time off to have babies, and thus can’t compete with men who have been on the job the whole time? Who would you promote to CEO — the guy who has worked steadily for the past several years, or the woman who takes several months off every three years, then spends less time at work when she gets back?
Shank
Apr 25 2004 at 5:23pm
Unfortunately, I think the previous commenter is correct. And I don’t know how to solve the issue.
When a dual income household decides to have children, even if both persons decide to return to work full time after the baby can go to day care, there are things that come up that one of the parents needs to deal with during work hours. Such as when the baby gets sick. The parents need to decide which parent will take off work for that day. Chances are, because things are *already* unequal and the man is likely making more money than the woman, the opportunity cost will be lower if the woman takes off from her job. Women would therefore miss more work and likely look less responsible to a person who could hire them for a management or CEO position.
Once her kids are older and out on their own, the woman is still a few rungs behind a man who was in an equal position before she had children.
This is probably true even if this competing man has children, because chances are, he was making more money than *his* wife, and was able to be at work more because the wife took care of things that came up with the kids. So, wage and position inquality begets more wage and position inequality.
Maybe if a woman is interested in being a CEO *and* a mother, she should choose a mate who wouldn’t have a problem missing a few days here and there from *his* job. Or, as the study may show, are men too competitive and macho have a CEO for a wife?
Mats
Apr 26 2004 at 3:01am
Probably, statistics is important here. And you don’t have to assume any large difference in variance between the male and female population’s abilities. All you need to do is to assume a small difference in the mean. This is just the way the normal distribution works. Very small differences in the mean leads to large ratios of probablitities in the tails.
From “a mathematician reads the newspaper”, I forgot the name of the author alas.
dsquared
Apr 26 2004 at 11:23am
Mats: But we know it isn’t a difference in the mean; if it was, this would imply large differences in the lower tail which are at odds with casual empiricism.
Jacqueline
Apr 26 2004 at 12:56pm
The gender pay gap also almost disappears if you compare childless women and men. (98%, see Christian Science Monitor.)
I think it’s unrealistic to think that a woman can both have children and a career without cheating one or both.
Jacqueline
Apr 26 2004 at 12:59pm
Oh, and until you teach a man to lactate, the woman is generally the best one to stay home with the baby. She’s going to have to take at least a week or two off to give birth anyway.
Boonton
Apr 26 2004 at 2:43pm
No one can have both a family and a career without cheating one or both to some degree. There will always be someone who will work that extra 15 minutes to the exclusion of his wife, his kids, her husband or even having a normal social life.
On Saturday I was in borders reading The Letters of Abelard and Heloise, which was mentioned on last week’s Sopranos episode. In it, Abelard was about to marry his student & lover Heloise. She, even though she loves him, insists that he not marry her since marriage would distract him from using his intellectual gifts to write & teach philosophy.
Talk about a different way of thinking but there is a point to the Medieval mindset. Obviously a person who is able to devote themselves 100% to something has more resources than the person who is only able to devote themselves 75%.
A CEO is not a normal career path for most people. It is more like being a superstar or sports hero. Unless a person has great natural gifts, it very well may require people to sacrifice family to their goal (look at the divorce rate among Hollywood stars).
For most jobs, taking off a week or two is not a career diaster. Lactation is no longer as essential as it used to be since we seem to have gotten pretty close to perfecting bottling milk…
Brad Hutchings
Apr 26 2004 at 4:14pm
I think this explains the first season of The Apprentice. The all-girls team clobbered the all-boys team at first, then when Trump mixed things up, the boys rose to the top.
(ducking)
-Brad
David Tufte
Apr 26 2004 at 6:23pm
Mats and DSquared don’t have the math quite right.
What Arnold is talking about is a higher variance in the distribution of important variables for men (I first heard Camille Paglia make this point, but I don’t know if she invented it).
A higher variance for men will produce a 99% percentile for women that matches up with a much lower percentile for men. If CEO’s are chosen from the top of some distribution, then the pool will be dominated by men.
This argument goes both ways, and has been used to explain why so many more men are in jail too.
Bernard Yomtov
Apr 26 2004 at 8:47pm
Most CEO’s are in their 50’s or 60’s, so their careers began 25 to 40 years ago. What was the male/female ratio in business, law, and engineering schools then, and others who sought a business career?
I suspect the ratio was tilted heavily towards males, so it shouldn’t be surprising that most of those who made it to CEO are male. That’s not going to be the only reason, but it’s an important one, I suspect.
David W. Crawford
Apr 26 2004 at 11:44pm
I think both David Tufte and Bernard Yomtov are
correct.
The first paper I saw on this ‘the class with the
high variance will be over represented in the set
of winners; and males have high variance’ meme was
in the early 1990s by Eddie Dekel. A
later-published cite is:
Dekel, E. and S. Scotchmer. (1999).
On the Evolution of Preferences in Winner-Takes-All Games.
Journal of Economic Theory, 87, 125-43.
dsquared
Apr 27 2004 at 4:45am
This argument goes both ways, and has been used to explain why so many more men are in jail too
Exactly; that’s how we know that the ratio in the uper tail is being driven by a difference in the variance rather than a difference in the mean.
Eric Krieg
Apr 27 2004 at 8:01am
>>I think it’s unrealistic to think that a woman can both have children and a career without cheating one or both.
Well, what I see in my workplace is that when women have children, they get on a “mommy track”. They don’t take or don’t get the tough assignments, the ones that build your reputation as a high performer and a problem solver.
I see a dichotemy at my workplace. On the one hand, they force working mothers into the mommy track. But on the other hand they play lip service to diversity, and seem to be making an effort to get more women on the board of directors. You would think that it would be obvious that killing the mommy track and being more accomodative to working mothers might help solve the second problem.
Shank
Apr 27 2004 at 2:46pm
Same thing at my workplace. The male bosses think they are being nice and considerate for giving mothers flexible schedules and easy assignments. They probably have the best of intentions and think they are being chivalrous. They don’t extend the same concern to fathers, however.
It is up to both parents to think about how they are treated at work. If a father wants to be more active in the lives of his children and isn’t given as much flexibility as mothers, then he needs to point that out. If a mother wants to focus on her career and let the father be more responsible for the children, then she needs to stand up for herself and insist on getting the tough assignments that she needs to further her career.
boonton
Apr 27 2004 at 5:29pm
I think there is general agreement that:
1. Equal pay should be given for equal work.
2. People should be neither rewarded or punished for having children.
3. As long as #2 is being followed, employers should be free to pay employees based on their performance & other factors that are legitimately connected to their job (such as the assignments they accept, tenure etc.).
If a woman decides to not have children (or not be the primary caregiver to her own children) in order to focus on her career, then she should advance just like a man who makes the same decision. It’s unfair to her (& men) if the policy is to promote other women just because they are women.
That being said, we are a long way from really being a world of fully rational economic actors. I believe there are many times an employer may incorrectly assume a woman who just had a child should not be promoted because she isn’t 100% focused while they will overlook this so-called ‘lack of focus’ in male candidates who goof off. So I’m not quite willing to accept the proposition that just because gender pay differences disappear when you control for childbirth that means this is fully the result of free actors making decisions that best benefit themselves.
Patri Friedman
Apr 27 2004 at 5:56pm
dsquared writes: “Mats: But we know it isn’t a difference in the mean; if it was, this would imply large differences in the lower tail which are at odds with casual empiricism”
The empirical evidence I have seen in articles on this subject says the opposite of what you observe. Consistent with the variance theory, there are large differences in the lower tail as well, ie there are many more males with moron-level intelligence than there are females.
For those who believe in evolutionary psychology, this makes sense because men are higher-variance creatures because of the fertility differences. (a highly successful man can have many more children than an unsuccessful or average one, so its worth shooting for).
dsquared
Apr 28 2004 at 3:10am
For those who believe in evolutionary psychology, this makes sense because men are higher-variance creatures because of the fertility differences. (a highly successful man can have many more children than an unsuccessful or average one, so its worth shooting for).
Patri, this makes no sense; how could an individual’s genes code for variance, which is a property of populations?
Leon
Apr 29 2004 at 8:40am
How could the population’s genes code for variance, if not via the functioning of the individual’s genes? What other kind of genes does a population have?
One could even say that a ‘population’, as a homogenous ‘thing’, is a side-effect of the genes of a large quantity of individuals – almost like an “extended phenotype” of their collective gene pool.
An individual’s genes can contain certain ‘instructions’ or ‘propensities’, while other individuals, carrying out other roles, carry a complementary set of genes which produce certain propensities for certain behaviors. These complementary instructions combined, form a genetic coding which transcends the physical boundaries of one individual, much as a grid computer network can solve large computational problems across the physical boundaries of individual computers.
It is well known that genes can be recessive, and hosted by either parent, of either sex, then passed down to a child of either sex. This is seen clearly in physical characteristics. Why would a 40-year-old man carry genes for a certain behavioral trait normally used by 2-year-old child? He wouldn’t, but his genes are only as “concerned” about their current incarnation as they need to be in order to successfully propagate themselves to the next generation, then the next, then the next, and so on.
Populations get their characteristics from the aggregate behavioral traits of multiple individuals each playing a crap game, balancing the odds between taking this action or that (which, incidentally, may go towards proving that evolution isn’t about survival of the fittest, but about survival of those that fit best).
Comments are closed.