Does economics imply a narrowing gender gap?

In my latest New York Times column for The Upshot, I look at some evidence on the gender gap.  Here is the bad news:

In one set of these experiments, called the dictator game, women were found to be more generous than men. Players were given $10 and allowed but not required to hand out some of it to a hidden and anonymous partner. Women, on average, gave away $1.61 of the $10, whereas men gave away only 82 cents.

In another test, called the ultimatum game, one player received $10 and then decided how much of it to offer to a partner. (Let’s say the first player suggests, “$8 for me, $2 for you.” If the respondent accepts the offer, that’s what each gets. If the respondent is offended by the unequal division or dislikes it for any other reason, he or she may refuse, and then no one gets anything.)

The depressing news was this: Both men and women made lower offers, on average, when the responder was female. Male proposers offered an average of $4.73 to male respondents, but only $4.43 to women. More painful yet was the behavior of female proposers, who, on average, offered $5.13 to men but only $4.31 to women. It seems that women were seen as softies who were willing to settle for less — and the discrimination was worse coming from the women themselves.

I am nonetheless optimistic about longer-term trends, and here is one specific example I give:

As a former chess player, I am struck by the growing achievements of women in this great game — one in which men were once said to have an overwhelming intrinsic advantage. (Among the unproven contentions was that men were better at pattern recognition.) Although women were never barred from touching the chess pieces, strong female players were few in number.

These days, many more women play very well, and the gap between the top men and women in the game is narrowing. The main driver of the change appears to be that more and more women are playing chess, creating a cycle of positive reinforcement that encourages ever more women to excel. We’ve seen a similar dynamic in the workplace, as more women have made great strides in the areas of law, medicine and academia. And this process may spread to other sectors of the economy as well, such as technology industries.

Do read the whole thing.


Judit Polgar and ... ?

Polgar recently announced her retirement from chess. The current best active female player is Yifan Hou (2663 elo, #87 on the fide rating list), the only active female player with a rating above 2600 (top ~250 or so).

My impression is that a lot of fields have become more male-dominated in recent years: computer programming is an obvious example.

Please read When Computers Where Human. There was a time when coding was low status and women did that boring job.

"was a time" < pre 1965?, probably even earlier - when there was no software to speak of. also i assume the numbers of women "programmers" even back then are highly exaggerated.

I was thinking of the COBOL Era: c. 1970 to Y2K. I knew lots of female COBOL programmers during that era, including, to pick one example, my wife. COBOL, the most widely used language in American business, was pulled together at the end of the 1950s by a woman, Grace Hopper.

That wasn't coding. That was taking complicated problems and breaking them down to simple steps that any fool could do. Many of those fools were female and all of them were called computers. But they weren't in the modern sense.

The status issue hints at a truth or a question. A man pursues a challenging career for X+mate. A female is supposed to do it just for her version if X.

That's not coding, that's computing, i.e. factoring numbers and doing calculations. They were doing the job of a Ti-85 calculator.

No, genuine female computer programmers were common in the American corporate world at least from the point I started in 1982.

My sister started in 1981 with IBM. She's had an IT career ever since, but hasn't been a programmer for probably 20 years.

If this is true then it's might be just an American thing. In Germany the proportion of female compsci majors has been rising for years:

Still only about a bit more than a fifth of all compsci students but much more than a few years ago.

(I know, I know compsci degrees don't necessarily lead to programming jobs but I'm pretty sure there's at least some correlation)

When Steve says "recent years" here, he means the last 20-30 years. Compare the fraction of women in programming in the 70s and early 80s to now...

Right, in America wasn't the peak percentage for women getting computer science degrees about 30 years ago?

One thing: fix academia where the career head start happens when women should be having kids.

To review some very good jobs, directing movies remains extremely male dominated. (They finally gave an Oscar to a woman, but it was the Boys Club's favorite lady director, Kathryn Bigelow, for a war movie, The Hurt Locker.) Screenwriting for the movies may well be more male than in the 1930s-40s, when male-female teams were common. I don't believe a woman has ever even been nominated for a Best Cinematography Oscar.

The producer was female too, though her money comes from her father Larry Ellison.

My general impression is that young women rapidly narrowed career gaps in the 1970s, but not much has happened among young people since then. By the time I got my MBA in 1982, it was simply assumed that women could, would, and should go into all careers men went into. (The only word of caution I can recall being given to female MBA students at UCLA in 1980-82 was that women, especially gentile women, shouldn't try for a career in Los Angeles department stores: you had to be a Jewish man to get to the top in L.A. department stores.)

I don't see much change in the numbers for young people in the three decades since then. If there had been much change since then, there wouldn't be the 2012-2014 hysteria about male privilege, would there?

The way to understand the narrowing of the gender gap is to simply see that as populations grow, the tails of the bell-shaped curve get bigger so naturally you'll get more top level players. But it's still unfair that women get the short end of the stick.

Bonus chess read:

Women, on average, offered more than half to male opponents in the ultimatum game? I'm confused as to why anyone would do that, ever.

Well, you should always offer 1 penny because it is irrational to reject any amount, right?

I can see why someone might reject a lopsided ultimatum out of spite. I can't see why someone would reject an even split. On average, women offered $5.13 to men, keeping $4.87 for themselves. What I don't understand is the thought process that led to some subset of women saying "I'd better offer him $6 and keep just $4 for myself."

Maybe women are wired to overshare. I bet men are wired to over-spite, but maybe not to the same degree and maybe not in a game where he can't reach the other guy's throat.

Or maybe women are more wired to treat it as a learning game.

Hmmm. Don't you mean maybe women are wired to overshare with men?

According to this, women are hard on women.

No behavior seems particularly irrational when the amounts involved are so small. If I were ever involved in such an experiment my dominant consideration would be my contempt for dumb social science experiments.

I once told my girlfriend about the Ultimatum game, and how if there were big money involved, the outcome might diverge from what we see in the lab. She replied that when she had gotten divorced many years ago, she refused her ex-husband's request to file as "married" during their divorce year, even though it would mean saving both of them money. Women can be spiteful too!

Is it bad to be nice and to value relationships over money? That seems to be the thinking behind considering the results of this experiment "depressing". Is it depressing that women are more generous and perceived as less likely to get offended and reject a deal?

It doesn't say it's bad. It does suggest it's not behaviour that will close the gender gap.

Surely Tyler calling it "depressing" implies that it is bad.

And as for the gender gap, the same applies: if many women are being less aggressive in raising their salary and getting ahead, is that actually bad? Maybe this behavior is more commendable.

Clearly, if discrimination against women is causing the gender gap, the gender gap is bad--it is unequal opportunity. But evidence is accumulating that the gap exists even when there is not discrimination, and therefore the discussion is shifting to behavioral differences. To the extent that it is a story of women making different choices, there is no obvious reason for considering these choices worse than men's choices. Indeed, in other contexts it is often considered bad to be motivated by grabbing power and making gobs of money, while opting for a less lucrative career path is commendable.

If women are less aggressive in seeking salary increases, one would expect savvy wage conscious employers would be hiring more of them.

What I think may be the problem is that persons may be focusing on just one element of the pay package: money,--.

Ignoring that women might exchange pay for other benefits, such as job flexibility.

Economists tend to measure the world with rulers and yardsticks,

when real humans consider consider distance, time and temperature, the latter two of which are not measured.

"If women are less aggressive in seeking salary increases, one would expect savvy wage conscious employers would be hiring more of them."

And a lot of employers do hire huge numbers of women. So, we're moving toward a society in which the Top Dogs are lavishly rewarded and are overwhelmingly male, huge numbers of women have moderate paying jobs, and large numbers of men aren't employed because they are too obstreperous.

Bill & Steve, nice points.
I read, "women might exchange pay for other benefits, such as job flexibility." and thought "Or not working with pricks."
Then I read Steve's comment. I guess obstreperous is a good synonym.

Perhaps there are just a lot more women and a lot less men playing chess now.

I am an officer in the best chess club (IMO) in one of the top 10 metropolitan areas in the US. Currently our membership has one female and close to 100 males.

Tyler Cowen is wowed by this female success.

I think behavioural economics reveals some very striking truths. But I am in the camp of those unconvinced by those little experiments they like to run. I don't think they're worthless, but I do think it's always a big step from the experiment to application in the real world. What I mean is, maybe all we have learned is that women will offer more money to men (than to other women) when they participate in this experiment.

"it’s always a big step from the experiment to application in the real world."

I agree. I see two major problems here:

1) Reasoning from an artificial situation to the real world. Not enough weight is given the the fact that the participants are playing a game. In Risk, I try to ruthlessly eliminate competitors, but I'm not like that in real life (I hope). I consider deception to be immoral in real life, but good strategy in many games. And so on.

2) Inferring particular motivations and thought patterns from behavior. Often this rests on lots of assumptions, and there are many other possible explanations for the results.

There are other more-real experiments that find similar results. I recall a paper being published on a "interviewee quality" experiment where participants were asked to evaluate a job candidates interview and subsequent benefits negotiation.

On average, the observers viewed males that attempted to negotiate their benefits package a positive light (smart, aggressive).
On average, the observers viewed females that attempted to negotiate their benefits package in a negative light (bitchy, aggressive).

The observers consisted of both male and female study participants.

I do believe that this is more-real than the ultimatum game, but both play on the same prejudices. Both women and men *expect* women to be more mild-mannered and accommodating.

The little experiments are worthless. A game situation doesn't parallel a life situation, it's only a game. In a given pool of game participants a very small number of aberrant choices are statistically misleading and skew the results. They can't be correlated to society as a whole even if they are valid from a behaviorial aspect.

"On average, women offered $5.13 to men, keeping $4.87 for themselves." This is playing with phony numbers.

I want to see some results from experiments with substantial money in play. Like say ultimatum game with $1,000. Is anyone really going to turn down a $900-$100 split due to unfairness?

I would also like to volunteer as a participant in such an experiment.

See my comment above for one anecdote.

There is a varint on the ultimatum game that I would be interested to see: rather than the offeree deciding to accept or reject, a third party decides. If males know a third party woman is deciding, they may offer the same to both sexes.

+1 Or, if the company is doing 360 degree performance reviews and you know those above, to the side, and below you get to comment on your team spirit or your ability to work in teams.

Isn't the gender pay gap under 10 cents now?

And what of the gender incarceration gap?

And the longevity gap. Men might get paid more than women but women inherit it in the end and get eight more years to spend it. By the way, I thought the pay gap goes away when accounting for the time spent child rearing.

Why does it matter? Seriously, why is it necessary, or beneficial to society, that men and women earn the same amount on average?

Other than some upper-middle-class women feeling "unfulfilled," what problem is solved by eliminating the "gender gap"?

If you stop thinking of the gender gap as a problem that good people are trying to solve and instead see it as an excuse to do evil, such as harm women who would rather be home with their children, it makes sense.

Interesting that male prejudice for the "correct" answer is displayed in this post--

that somehow being more selfish in each game is the best way to play either game, particularly when they are not one shot games.

You see, I have used both games in a graduate pricing class involving vertical channel coordination and the operation and formation of joint ventures.

In those extensive games, what the students learn is that channel and joint venture partners will not succeed in the game unless they have trust and behave cooperatively, even sometimes not taking advantage of a partner. Those who take a longer view and cooperate do better than those who do not.

You can play this game as well: play the dictator or ultimatum game with the students. Observe the persons who didn't play fairly with their partner (if you can use such terms with these games). Later in the class, introduce a vertical coordination game (e.g., manufacturer has to rely on dealer to promote a product at an early stage, or a short run shortage arises and dealer has the opportunity to switch suppliers rather than wait). Observe how the person who screwed his partner in the first game gets treated when that person now has to rely on the other.

The world is not one shot games. There is a reason why man evolved in tribes and not alone in some hut by himself and a spouse in the jungle.

By the way, I learned this experiment from a Caltech experimental economist who played it on me and another lawyer.

Thinking that the best way to play the game was to give as low an offer as I could to the other lawyer, I did.

Then, the economist followed up: he interviewed the person whom I screwed on how he would get back at me, and whether he would now trust me or whether if there were a contract which did not cover all of the possible eventualities and now there was a need for one side to cooperate with the other by making a concession it was not contractually obligated to make (with little cost to him but much greater cost to the other party) whether he would make a concession that cost him little but me a lot..

That resulted in a long discussion.

By the way, remember, chess is a two person, one shot, zero sum game.

And, don't forget, the results of these games are culturally you might not be very successful in international distribution or marketing if you play the distribution or joint venture game as if it were chess and this was a simple zero sum game.

Here is a little lawyer knock knock joke

Knock Knock, who's there, Sue, Sue Who? Bill, Bill Who? Sue

Why was tyler promoting feminist nonsense?

The bias and bigotry here are entirely Tyler's.

Tyler believes that if men are performing better than women at something -- whether it be chess, or negotiating, or probably a thousand other things -- that this is "depressing" and represents an obvious and alarming problem that must be fixed. Until that day comes he must find ways to show examples of this terrible problem, while also reassuring everyone that he is "optimistic" because "the gap is narrowing."

This bias and bigotry will get you published in the NYT, but has no other purpose.

To the extent that we still see areas where women go from underperforming to overperforming when fairly cheap and reasonable accomodations are made, it is depressing. It's depressing not just because of it's effect on the women in question (although basic human decency requires being depressed by that). More importantly, it's depressing because it implies enormous stocks of potential human capital were left to rot. Worse, they are still being left to rot. Less depressingly, it implies enormous sums being left on the table, ready to be grabbed by anyone with enough creativity to think outside traditional gender roles. I'm already starting to see this in my chosen field (software engineering), where outreach to women is starting to result in great new crops of female developers. I'm happy that my company is both outreach to women this because it's the right thing to do, and because the software development marketplace is such that I would cheerfully hire a coelacanth if it could ship code weekly.

I went to the informative site that Tyler linked to in his NYT piece, about women in chess. In the tabulated women FIDE rankings shown there, Judit Polgar is on #1, and her sister Susan is #5. Little sister Sofia is also outstanding.

I'm a big tennis fan, and I'm also American. It never ceases to fascinate me that in the last 12 or so years of tennis, America has been carried--how many millions are we now?--chiefly by four individuals:

Serena Williams
Venus Williams

Bob Bryan
Mike Bryan

Serena is by far the greatest woman tennis player of her era, and definitely is one of the top five women ever, arguably in the top three ever. Venus was at times #1, and probably is one of the top seven women players of all time.

The Bryan brothers are by far the greatest men's double team of their era, and probably the greatest team of all-time.

It is really quite striking how weak America has been in tennis in the past 12 years or so--after the pair of siblings, we are talking about John Isner and then really big steps down, both in men's and women's. It is quite amazing, truly.

In the NYT piece Tyler says the gender gap in chess has been narrowing, but doesn't say what the evidence for that is. I suggest that sibling-hood is factor. If you took the Polgar sisters out of the data (Judit, btw, recently retired), would there be any evidence of a gender-gap narrowing in chess?

Minor correction: When I speak of "big steps down" I should confine that to at present; over the past 12 years there's been Andy Roddick, a former world #1, and probably others in the top ten.

As I mention above, Judith Polgar recently announced her retirement from chess - she's no longer playing. The truth of the matter is that there is only one [1] female player in the top-250 among FIDE-rated players.

"In the NYT piece Tyler says the gender gap in chess has been narrowing, but doesn’t say what the evidence for that is" - I'd suggest that the simplest explanation for this omission is probably the correct one; he doesn't have any evidence. If he did, he'd have included it. He probably figured that it would be a bad idea to start talking about the actual data, given what the data about the top players actually look like.

There may be more female IM's and GMs and so on, that's probably true. But aside from completely exceptional cases, they're still nowhere to be found in top chess tournaments or rating lists.

Another way to think about this would be to consider the average ratings of top male and female players. Go to and look at the stats. The top 100 list (whether or not Yifan Hou is included there or not is unclear, as it's ambiguous whether this is a top list of men or just a top list of players (it's called one thing in the headings but another in the hyperlink) - however her inclusion or exclusion doesn't really matter as there are plenty of males around that level) has an average rating of 2703. The female top 100 has an average rating of 2441. This is a rating difference of 262. A rating difference of 250 implies that the stronger player has an expected score of 0.81 - i.e. he or she will win more than 4 out of 5 games. It may be possible to argue that the difference used to be even larger, but considering that potential narrowing of performance differences as a good example of a wider trend along similar lines is, I think, damning women with faint praise. Most of the best female chess players in the world would simply get completely crushed if they were to participate in a top tournament playing against people like Caruana or Carlsen, and this has most certainly not changed during the last 30 years.

All professional fundraisers know that successfully soliciting any non-tax or estate planning driven gift is far easier from ANY woman than from ANY man. Retired men are a distant second. Exceptions occur, but experience is overwhelming that women are more charitably inclined.

Although I no longer work in that field, just this week I asked an attorney making well into six-figures for $25 for a service club to which we both belong. His response: give me the form. I'll see if the firm will pay it. I was not surprised.

Really? I'd assumed that the truly massive name-chiseled-on-the-building donations tend come from men. I've been told that the stereotype among college development offices is that the whales are usually white male businessmen, probably legacies, who played a little ball back in the day.

"It seems that women were seen as softies who were willing to settle for less — and the discrimination was worse coming from the women themselves."

How is this the only interpretation? Perhaps women are seen as more rational and cool-headed than men who might be more easily offended/stubborn, and therefore willing to accept a lower amount (since this is the 'rational' thing to do).

Overall, pretty ridiculous to play an ultimatum game with trivial sums of money. We all know if the pie were $1million that offers wouldn't be anywhere near 50%.

It is worth repeating this talking point from NOW: women make 77 cents for every dollar that men make doing the EXACT SAME WORK.

lolbait. No two individuals do the EXACT SAME WORK, hence the use of the term individual.

The 77 cents is before adjusting for sex differences in career choices, hours worked, etc.

Imagine what a beautiful world it would be if women had the exact same personalities as men.

The amount of promiscuity would be comparable to that of homosexuals.

Male homosexuals, that is.

Oh, come now- straight women are already at least as promiscuous on average than gay men. For one, it's vastly easier to find a partner. Those ultra-promiscuous gay men account for only about 2% of the total, but get widely written about for obvious reasons.

And I suspect male and female personality distributions are already more similar than existing research demonstrates. To me, the results of this tiny experiment seem to demonstrate women acting in their own self-interest, but within the framework of existing power dynamics. We're all more inclined to be generous with people we see as more powerful than ourselves. Women may be somewhat more attuned to social interaction and obligation over the long term. We also tend to perceive men as having more power, therefore when we have control over a resource, we consciously or subconsciously want to get on the good side of a counterpart we see as dominant, hoping it will come back to us down the line. This doesn't always work out for us in the artificial world of corporate hierarchies, but it's not a bad strategy for a social animals evolved for a degree of interdependence. But you could also conclude that women are at least as economically rational as men; we're confident that women will accept less, so that's what we offer. Which is why the pay gap exists; we will do the same work for less money. Fair? Maybe not. But it is predictable enough that you can usually get away with lowball salaries for nearly every woman you hire, and this will be true indefinitely. However, the things women are more inclined to demand from their workplace do not reduce as nicely into quantifiable numbers, so things may actually become more fair in a holistic sense without a considerable reduction in the pay gap.


The depressing news was this: Both men and women made lower offers, on average, when the responder was female. Male proposers offered an average of $4.73 to male respondents, but only $4.43 to women. More painful yet was the behavior of female proposers, who, on average, offered $5.13 to men but only $4.31 to women. It seems that women were seen as softies who were willing to settle for less — and the discrimination was worse coming from the women themselves.

How many of the people, when offered 2$, choose to reject it. Probably miniscule. Would you reject it? I'd like to see this finding replicated.

Maybe out of spite. It's only $2, after all. That's not even a coffee and a doughnut. On the other hand, I might not turn down an 80-20 split on $1,000--or even $100, since that's "real money"

Yeah these experiments might be better framed as "How much are participants willing to forego (and how much do they expect others are willing to forego) to spite someone doing something unfair?" And while I'd expect some scaling-up as the amount involved becomes larger, it won't scale up by much. I might be willing to forego $100 to spite someone who has a $100,000 to split and gives me $100 and keeps $99,900. But $99,000-$1,000? I don't know. Maybe if I get to see their face when they find out they don't get to keep $99,000.

There may be a worse field than chess to use to support an argument that the gender gap is closing. But I can't imagine what it is.

Comments for this post are closed