"Talks a good game about freedom when out of power, but once he’s in – bam! Everyone's enslaved in the human-flourishing mines."

Invisible Women

[EDIT: Two readers – Pseudo-Erasmus and Noah Smith – have chimed in with very plausible explanations centering around labor force participation and productivity.]

A commenter asked in yesterday’s discussion on The Two-Income Trap: why did the entry of women into the workforce produce so little effect on GDP? Here’s the graph of US women’s workforce participation:

And here’s the graph of US GDP.

In about 50 years – 1935 to 1985 – we went from 20% of women in the workforce to 60% of women in the workforce. Assuming a bit under 100% of men in the workforce, that’s an increase of almost 50% over the expected trend if number of women in the workforce had stayed constant.

I’ve already admitted I find the fixedness of the GDP trend bizarre. But how on Earth do you unexpectedly raise the number of people in the workforce by 50% and still stick to exactly the same GDP trend? It would be like the US annexed Mexico one day but the GDP didn’t change a bit. Are we imagining here that if women hadn’t entered the workforce, GDP would have suddenly deviated downward from the trend, and totally by coincidence women rushed in to save the day? And then when we ran out of interested women to add to the work force, again by coincidence the GDP stabilized back to its trend line?

Possible solutions:

1. The GDP data is totally false. I am suspicious of that GDP data anyway. Maybe some joker in the Bureau of Labor Statistics just graphed out an exponential function and reported that as our GDP to see if anyone would notice.

2. The GDP data is low resolution. So low-resolution that even a change of 50% is invisible if stretched over a long enough time period.

3. Women contributed through unpaid labor in the home, and their paid labor substituted for that but didn’t add to it. But as far as I know GDP only counts paid-for goods. Not only should women’s labor in the home not have counted, but GDP should overcount the benefits of putting women to work because paid daycare and so on appear as valuable new services.

4. Somehow in total contradiction to usual economic theory, all gains made by women came out of the pockets of men, leaving the same growth as would have happened anyway.

This last one segues into a question asked by another commenter – did women entering the workforce drive down male wages?

This seems a lot like the question “do immigrants entering the workforce drive down native-born wages?” to which economists tend to answer “no” with more or fewer caveats. But the economists’ explanation for the immigrant effect is that in addition to producing, immigrants also consume, increasing demand.

Women were presumably already consuming. Back when they were housewives, they still needed houses, food, clothes, entertainment, et cetera. Their entrance into the workforce may create slightly more demand – for business clothes and office supplies, for example – but nothing like the demand created by immigrants entering the country.

If women increase the labor supply by 50% but don’t change the demand for labor, that seems like it should make wages go way down. Even if women are primarily in different jobs than men, men’s wages should still go down via a substitution effect (that is, even if women all become elementary school teachers, then a man who was planning to be an elementary school teacher before might become a fireman instead, raising the labor supply for firemen and pushing down fireman wages).

But the only study I can find to investigate this says it didn’t happen. And male wages don’t seem to have dropped dramatically starting 1950 or so. They do seem to have stagnated dramatically starting 1970 or so, but I feel like the income inequality explanation for that is on pretty solid ground. Unless you want to argue that the reason for income inequality is that between both sexes there’s now such a large reserve army of labor that the capitalists can get away with paying the workers very little. But I feel like economists would have told us if this were going on.

I hear a lot of conspiracy theories about women in the workforce. On the left, women in the workforce are being exploited and kept down by a sinister patriarchy. On the right, women in the workforce are a satanic plot to weaken our moral fiber.

But so far I’ve never heard the conspiracy theory that women never actually entered the workforce, that all the working women you see around you are animatronic robots or carefully crafted stage illusions.

Maybe this is the surest sign that the conspiracy is working.

This entry was posted in Uncategorized and tagged , . Bookmark the permalink.

105 Responses to Invisible Women

  1. Oligopsony says:

    One hypothesis: not only is the entrance of women into the workforce mostly the replacement of non-market by market forms of production, but it’s part of a longer trend of such, so “the entrance of women into the workforce” isn’t a break in any sort of trendline. You could probably tell a story – I think some of the smarter right-wing Catholics already have – where steady Cthulhu-swimming-commodificationward produces feminism as an afterthought. Good or bad, depending on what you value, but not its own thing at the level of causation.

    • Athrelon says:

      Even so, as Scott notices, replacing non-market labor with market labor should artificially inflate GDP. Even in a world where working women were so incredibly unproductive that they only made enough money to hire each other as daycare workers, that should register as an increase in GDP. (Despite, of course, decreased quality of life all around.)

      • Avery says:

        But GDP did go up.

      • Paul Torek says:

        Sure, it should register as an increase in GDP, but not necessarily a faster-than-before increase. The hypothesis is that commodification takes various forms, but the forms gradually get substituted one for another, while continuing on the same overall trend.

    • Army1987 says:

      You could probably tell a story – I think some of the smarter right-wing Catholics already have – where steady Cthulhu-swimming-commodificationward produces feminism as an afterthought.

      I once heard that in Europe women mostly entered the workforce during the World Wars, because men were needed in the military and someone else had to do their jobs instead. But Fig. 1 in this post doesn’t show anything like that happening in the US.

      • Randy M says:

        That’s definately the story that’s told here too, but the graph doesn’t show it. Maybe that’s because the male workforce participation included those out of work who were drafted?

    • ADifferentAnonymous says:

      This is a good example of the more general hypothesis that the exponential trend is the sum of a steady parade of things that each seem like they just have to cause a discontinuity.

      It also answers the question of why feminism pushed ‘men and women both work’ rather than ‘women can be providers and have housebands’.

  2. Joe says:

    Could the shrinking family have anything to with this? With women in the work force and fewer children to provide for could this lead to a neutralized GDP?

    • Meagan says:

      That was exactly my thought. Birth rates per 1,000 women have fallen by close to half since the 50s/60s… if you’re just replacing having 4 kids where half work with two kids where both work it could make a substantial impact.

  3. Stefan says:

    Another version is that large labor force expansions undercut productivity growth, say because the lower wages and change the incentives for management. There is not that much serious work on this, a nice speculative paper is Paul Romer in the 1990 BPEA, particularly the discussion related to figure 1.

    http://www.brookings.edu/~/media/Projects/BPEA/1990%20micro/1990_bpeamicro_romer.PDF

  4. gunlord500 says:

    I’m glad I was able to inspire such a thoughtful post, though I’m sorry if the manospherians end up jumping all over you for this. I agree with the paper you cite, I’d assume wages would have declined immediately when women started working in the post-war era rather than waiting for the 70s and 80s to do so, if the “economic fempocalypse” theory was correct.

    • Andy says:

      I’m sorry if the manospherians end up jumping all over you for this.

      Yes, you should be sorry. 😛 Kidding, this was an interesting question.
      Or my theory – women, especially younger and poorer women – were in the workforce already, doing low-skill industrial labor, being maids, laundresses, working on family farms, etc etc etc, and the “women in the workforce” – accounted for a much bigger slice of women than the statistics in the graph at the top of the post imply. There’s also that funny gap between 1860ish and 1890.

      • Scott Alexander says:

        Shouldn’t there still be a notable jump from women being able to rise to the level of their skills?

        • Andy says:

          If that was a slow growth, the way women in technical fields supposedly had to be better than men to earn any kind of promotion or recognition, it’s possible the gain would have been slower. Female integration into the workforce didn’t happen all at once, but sector-by-sector.
          And if GDP is measuring potatoes and cars and such across the entire nation, it might not have been noticed the female workforce in all the other Vast Formless Things acting on the economy, like the overall stagnation of wages from jobs being replaced by robots and globalization.

        • Konkvistador says:

          Not if women where already on the level of their skill.

        • Steve Johnson says:

          Now now, Konkvistador – only progressive assumptions allowed.

        • Multiheaded says:

          Have either of you ever met a shrewd woman in the workplace? Man, have I. Maybe you’re just stupefied by the background misogyny?

        • nydwracu says:

          I’ve seen one of my female relatives at work. She spends a lot of time gossiping with other women, and a lot of time checking her email. The part of her job that she talks about the most involves letting contractors into the building.

          Another one of my female relatives has “environmental” in her job title.

          All of my male relatives are engineers, except for one who’s retired now, who was an electrician, and another who died a long time ago, who… well, there’s apparently a room named after him in the RNC’s headquarters. Political media consultant, I think. Never bothered to check but I have a vague suspicion that he might have worked on the Reagan campaign.

        • Josh says:

          No because demand was relatively unchanged. The annexation of Mexico was a false comparison because in that scenario more humans are entering the economy. Here, while more humans enter the work force the actual economy, aka demand, is left unchanged.

        • Nornagest says:

          @nydwracu:

          Female relatives… let’s see. Of those whose professions I can think of offhand, I’ve got an Army officer, a couple of school teachers, an architect, a graphic designer, and a tech writer. Softer jobs by the standards of my male relatives (most of whom are scientists, engineers, or military), sure, but nothing I’d consider unproductive.

          That’s not to say this is representative — I’m sure it’s not. But while we’re throwing anecdotes around…

        • Multiheaded says:

          My mother used to be a radiotech engineer back in Soviet times, as did my father; she worked on some kinda classified naval stuff. In the 90s she became a housewife, then had a little involvement with his unsuccessful hotel business, then taught herself all about about plants and gardening and such, and now they’re running a nursery/flower shop. She’s in charge of the production, he’s in charge of the infrastructure, and the management seems to be deliberative/equally shared. My aunt used to be the chief librarian of an automobile plant. My grandma was a translator.

        • Anonymous says:

          The Right hypothesis of women having no talent is one way of resolving this; the Left hypothesis of relatively flat overal talent endowments is another. The idea that talent endowments are both flat across ascribed identities and highly differential across individuals may just be a justification for Liberal policies of equal opportunity and Careers Open To Talent and so on. But the fact that you can have thriving communities of scientists drawn exclusively from country squires or whatever seems to cut against this at least somewhat.

          e: this is Oligopsony btw

        • Andy says:

          I’ve seen one of my female relatives at work. She spends a lot of time gossiping with other women, and a lot of time checking her email. The part of her job that she talks about the most involves letting contractors into the building.

          I’ve known men who gossip all the time at work and check sports scores. From this we can assume that men are unsuited to an office setting. Some people of both genders work well in an office setting ,some do not, and gender alone is a very poor way to distinguish between the two.

          Another one of my female relatives has “environmental” in her job title.

          I fail to see the relevance. I’ve known some very bright people with “environmental” in their job title. Professor of Environmental Science and Policy, for example. Another very bright lady, and mentor of mine, does not have that word in her title, but teaches several classes with it in their title and is an expert on water in the Western United States.
          What inference are we supposed to draw from these anecdata?

        • Oligopsony says:

          Aren’t Reactionary or affiliated types always complaining that modern schools are oppressive to men because they involve sitting down all day, doing boring tasks, &c.? The implications for suitability to office work are left as an exercise to the reader.

        • pwyll says:

          Re: Multiheaded’s comment:

          Have either of you ever met a shrewd woman in the workplace? Man, have I. Maybe you’re just stupefied by the background misogyny?

          …and Oligopsony’s comment:

          The Right hypothesis of women having no talent

          This characterization of reactionary men as believing that “women have no talent” is unfair. To the extent that a reactionary man is intelligent, he’s likely to have *more* contact than average with smart, capable women – both at work and among his relatives and family. But he understands that the overall distribution of ability, and interest, in the various human pursuits varies by sex… and that trying to ignore these differences will likely have unpleasant side effects.

          The fact that belief in sex differences in personality and cognition is currently controversial, when it’s been obvious throughout history to anyone with experience, is bewildering. See also: Understanding current causes of women’s underrepresentation in science

          and

          Why women see differently from the way men see? A review of sex differences in cognition and sports

        • Oligopsony says:

          Thanks for pointing out the non-symmetry of my phrasing; I allowed my sympathies to seep in. Better phrasing:

          right hypothesis – relatively greater categorical differences in talent endowments
          liberal hypothesis – relatively greater individual and relatively less categorical differences in talent endowments
          left hypothesis – relatively less individual and categorical differences in talent endowments

          N.B. that the right hypothesis here can be complementarian as well as supremacist, though comparative advantage makes the functional difference between the two minimal.

          To the extent that a reactionary man is intelligent, he’s likely to have *more* contact than average with smart, capable women – both at work and among his relatives and family.

          Of course there are multiple ways one could interpret such a bias – if the theory that male talent variance is greater is correct, then people in higher-human-capital social circles should be see an unrepresentatively male-biased section of the total distribution of ability. Of course by Dunning-Kreuger we all think we’re on that side of the bell curve anyway so maybe per Hobbes we shouldn’t take that as our default assumption.

        • Multiheaded says:

          N.B. that the right hypothesis here can be complementarian as well as supremacist, though comparative advantage makes the functional difference between the two minimal.

          In practice, on the other hand, implicit supremacism is of course very heavily used in the ideological justifications for keeping the rate of exploitation so high – but no matter how supremacist the underlying belief structure, we’d expect it to be wrapped in complementarianism.

          And so any attempt to optimize for whatever specialization advantage complementarianism might ascribe to women is semi-implicitly seen as women trying to optimize for their comparative advantage given supremacism – as you noted, we’d expect the practical measures to look similar! – which further reinforces supremacist ideology. See: the Cult of Domesticity.

        • jaimeastorga2000 says:

          I fail to see the relevance. I’ve known some very bright people with “environmental” in their job title. Professor of Environmental Science and Policy, for example. Another very bright lady, and mentor of mine, does not have that word in her title, but teaches several classes with it in their title and is an expert on water in the Western United States.

          I don’t think the point is about intelligence so much as it is about wealth creation. Someone with “environmental” in their job title is unlikely to be creating any value. If anything, it is likely that such a job consists solely of making life difficult for people who are trying to create value.

          • Assuming that the environmentalist is developing or enforcing reasonable restrictions, they should be considered to be contributing value just as much as fire fighters are.

            You can’t live on the disasters that don’t happen, but you still need to do some work to prevent disasters.

  5. Anonymous says:

    http://research.stlouisfed.org/fred2/series/CLF16OV/

    In this graph, it doesn’t look like there was that much of a big spike in the labor force during the 60s through 80s–it seems like the labor force was increasing by a similar amount during the 90s and 2000s, although not after the recent recession.

    • Douglas Knight says:

      If you’re trying to make a point about GDP vs GDP per capita, I’ll pass, except to say that the exponential nature of GDP is crazy.

      If you are trying to reconcile that graph with Scott’s graph, the difference is population. The Civilian Labor Force Participation Rate flattens in 1990, matching the employment of women. But I don’t know its definition, how it could be so smooth (nor your graph). A transparent statistic is the Civilian Employment-Population Ratio. It is what is says, modulo seasonal adjustment.

      • Fnord says:

        But even the Civilian Employment-Population Ratio goes from a low of 55% in 1950 to a high of 65% in 2000 (going from the lowest and highest points, which probably overstates the actual change because of year-to-year variation). Which is quite a bit less than the ~50% increase that the analysis in the OP suggests.

        Post 2010, the difference shrinks even more: the Civilian Employment-Population ratio drops to 58%.

    • Fnord says:

      This seems to be closer to the root, here. Despite the entrance of women into the workforce, there wasn’t an anomalously large growth in the size of the workforce. Of course, the reason for THAT is still a mystery that needs solving, but at least it provides a hint at where to look (a change in the rate of population growth?).

  6. Athrelon says:

    Some thoughts:

    1. Japan and Germany have a much lower fraction of working women, but they have neither fallen further behind the US, *nor*, crucially, have they taken over the world. Naive economic theory would predict that they’d be further behind the US as the US adds more, equally talented people to its workforce. A strong hypothesis that adapting the workplace to women imposes big costs that eat up more than their productivity would predict that lean Japanese and German corporations would gobble up market share from creaky old American ones that have waste Friday afternoons on diversity seminars. Neither seems to be happening.

    2. East Asia, by and large, has a productive class of poor people instead of a ghetto class. (This itself is a noteworthy fact that shatters certain assumptions about how inevitable the Western problems of poverty really are, but that’s a separate post.) But, this amazing advance in social technology doesn’t lead to taking over the world. Instead, what you get is lots of cheap restaurants and convenience stores manned with reliable, polite, inexpensive workers – the fulfillment of urban-boosters’ dream of affordable, walkable neighborhoods. This represents a pleasant quality of life increase, but it’s not really something that you can use to scale.

    One hypothesis that falls out of these observations is that maybe only a relatively small fraction of people are causally responsible for economically growth/progress. This group doesn’t have to be Galt’s Gulch small – it could be perhaps a third of the population – the point is that at some margin adding additional functional workers doesn’t do anything for growth.

    Patriarchy was already quite good at reaping the (relatively small) pool of female geniuses. We had female virtuoso mathematicians and Nobel-quality hard scientists at almost the same rates as we do today. The contributions of these women were valuable, but after you’ve already got the geniuses, adding women does practically nothing (positive or negative). The most you can get from adding non-genius female labor is quality of life upgrades that don’t scale.

    3. Another hypothesis of course is that women did help, but were exactly masked by The Great Stagnation, a decrease in social technology, or your favorite declinist hypothesis. We’re doing less with more – the very definition of losing technology. We really *would* be up to 50% poorer today if we hadn’t suddenly discovered that women could supply valuable labor to help keep the whole thing running.

    • Nornagest says:

      the fulfillment of urban-boosters’ dream of affordable, walkable neighborhoods.

      Having spent some time in Asia, I’ll buy “affordable” but question “walkable”. Might be different in Japan or Singapore, though, neither of which I’ve been to.

      • pwyll says:

        Nornagest, Japan and Singapore are the reverse – generally “walkable”, but not “affordable”.

    • pwyll says:

      Patriarchy was already quite good at reaping the (relatively small) pool of female geniuses. We had female virtuoso mathematicians and Nobel-quality hard scientists at almost the same rates as we do today.

      This seems anecdotally true to me, but I’m curious – are you aware of any hard data on the topic?

      • houseboatonstyx says:

        We had female virtuoso mathematicians and Nobel-quality hard scientists at almost the same rates as we do today.

        The glass ceiling lives on.

      • Athrelon says:

        Quantifying scientific accomplishment so we can figure out how to get more of it seems like an eminently sensible idea. But looking at postwar science it seems to be a disaster second to the implementation of the other nice idea, “wouldn’t it be nice if we shared everything so nobody starves?”

        I’m not quite sure how to make modern scientific institutions less of a hilarious sham (“Bring back the Royal Society” has a nice ring to it!) Until I get closer, I’m going to declare a philosophical objection and quarantine this whole science-quantification business.

        • Multiheaded says:

          The other nice idea didn’t quite work out, but it took you Westerners to kick it into the ground! Although… are you talking more about the likes of GDR or the likes of Maoist China? Because the latter was, however you spin it, an unqualified success. Of course it might’ve been far better had Mao kept the cooperative farming going and went a little easier on the collectivization… but still, China basically jumped out of its misery.

        • pwyll says:

          Douglas Knight comments in the other thread:

          While it may be hard to measure “virtuosos” and “Nobel-quality” scientists, it is a simple exercise to count the proportion of women that actually recognized by the Nobel committee. It has remained roughly constant over time.

          So it would appear there’s at least some evidence for your assertion.

    • johnwbh says:

      I don’t know about Germany, but the Japanese government has recently been putting aggressive efforts into changing attitudes to women in employment as part of their general economic reforms.

      Edit, source: http://www.economist.com/news/briefing/21599763-womens-lowly-status-japanese-workplace-has-barely-improved-decades-and-country

    • Neil says:

      I missed this comment before but its along the same lines I was thinking and the one conclusion still seems about completely right.

      Several people reached this conclusion and it’s just mostly true, most labor contributes very little. Note that as of today 5 million minimum wage US workers would make up less than 1% of GDP and even at higher future minimum wages would still only be 1%. There are even arguments for the large effect of beneficial trade deficits that have been engineered by the US and some specific Western nations over time and exploiting gains and resources elsewhere in the world but very little for labor on overall GDP growth.

      Also, the United States (and some small sections of Western Europe) can be argued to not just have hit the ceiling on productivity of its most talented subset of people in the 20th century but for a long time has clearly been bringing in a lot of the most talented individuals from other parts of the world. The country’s own population growth and shifting career and educational patterns also seem small in absolute magnitude compared to that.

    • Army1987 says:

      Patriarchy was already quite good at reaping the (relatively small) pool of female geniuses. We had female virtuoso mathematicians and Nobel-quality hard scientists at almost the same rates as we do today.

      What? Noether, Franklin, Meitner and Wu all worked in the 20th century. Which earlier genius-level female scientists do you have in mind?

      (Of course now James A. Donald will claim that the women I mentioned were affirmative-actioned into their jobs, but I’ll point out that if that’s why they were there it’d be quite bizarre that three of them were Jewish and the other was East Asian.)

      • pwyll says:

        Perhaps when he says “Patriarchy” he means the system that ended in the 1950s rather than the one that ended in the 1920s.

  7. Sarah says:

    Couldn’t average worker productivity have gone down at around the time women were entering the workforce?

  8. anon says:

    What does it mean to say GDP data is low resolution?

  9. the commenter with the same name says:

    Hypothesis: GDP is an utterly worthless statistic massaged into existence from original data that might have been useful before they got to it. The expectations that the GDP conjurers have to fulfill are roughly exponential in growth.

  10. JTHM says:

    Though the press often trumpets diversity as enhancing the productivity of a group or business, perhaps it does the opposite? Maybe, if you add women to a group of men working on a project, the women are individually productive, but their presence lowers the per-person productivity of the group? It seems like many pre-modern cultures segregated labor by gender; perhaps there’s an evopsych reason why it enhances productivity or something. Does anyone have any studies showing what happens to men’s productivity in a previously all-male group when women are added, either in the workplace, or in school?

    (Disclaimer: I suspect that if true, this effect would not be very strong, but it’s worth investigating nonetheless. It is, after all, a simple hypothesis and one of the easier ones to research.)

  11. Alrenous says:

    The GDP measure is just busted. We have no idea what actually happened.

    Corrected for inflation? GDP can’t grow without effective inflation.

    Toy model: Imagine GDP is $1000. Workers are paid in total that much in wages, which they spend on physical stuff – wealth.
    Women enter the workforce next year in exactly 50% proportion. So they make 50% more stuff. Only, total wages are still $1000. Which means all that stuff is going to be bought for 66% of the previous year’s price. Moore’s law has come to town.

    Inflation can be caused by fractional reserve, by money printing, and possibly by velocity. (Not entirely sold on velocity.) Instead of wages still totaling $1000, the stuff the chicks make is sold to dudes in time to make payroll, and then the chicks buy the stuff the dudes make. Then we get 50% more GDP.

    Which means not only does GDP need to be corrected for inflation (BTW most inflation measures are off by a factor of 3) but also for deflation/velocity. Only, velocity is buried in the totaled inflation+deflation+velocity ‘inflation=rise in prices’ measure in real life. We have no idea what actually happened.

    Women work very very different jobs than men. It’s not at all clear that their entry in fact made more stuff. More busywork, for sure, (HR departments) but then it’s hardly like men disdained well-paid busywork.

    • peterdjones says:

      Why’s everyone so down on HR? They’re Stakhanovites compared to Marketing….

      • Alrenous says:

        Marketing produces something: lies.

        Though to be fair I think HR internally functions as Government immunity. It’s mainly the outer, obvious-to-employees shell that’s primarily waste and graft. Even then, the blame for much of this is misplaced and properly belongs on regulators.

        Nevertheless, in an Ancap or Amish utopia, HR would very obviously not exist. Marketing would still exist, even though we’d probably be better off without it.

  12. You are all looking at it the wrong way. When you feel something about GDP doesn’t add up, the first order of business is to look at the accounting.

    Between 1950 and 1990, average hours worked per worker fell slightly. The employment to population ratio (EPR) rose by less than 8 points. This is not a massive change over the course of 40 years, so obviously it would be barely visible in a time series of aggregate GDP.

    But why would the EPR rise so little given women’s entry into the labour market ?

    (1) The female EPR nearly doubled from ~28% to ~53%, but the male EPR fell from <78% to ~65%. The fall in the male EPR happened across all age categories to some extent, but was most dramatic in the 65+ category. Really dramatic. It was also pretty substantial in the 55-64 category.

    (2) The male average working hours fell after 1950. It took place almost entirely in the 65+ age group, as well as in the 15-24 bracket. So you have both fewer old people working and the old people who continued working, working less.

    I don’t think there is any mystery. Eyeballing the data I’m satisfied that it all adds up.

    As women entered the labour force, men supplied less labour. At the household level, this is probably the income effect : as more wives were earning, more husbands could afford to earn less.

    At the same time, more generous retirement options allowed older people to retire earlier and/or work less.

    More people also went to college.

    All the gorey details are contained in this PDF of a Minneapolis Fed paper :

    http://pseudoerasmus.files.wordpress.com/2014/06/labforce.pdf

    It’s not analytic, it’s straightforward presentation of data.

    • Scott Alexander says:

      Thank you. This seems like the most plausible theory so far.

    • Neil says:

      There weren’t that many old people in the first place. Just changing numbers in the workforce isn’t actually a more plausible story and especially not a reason for not just holding steady but absolute GDP growth given population growth is not in proportion.

      So I get the argument that more women working is overridden by noise on larger scales except the factors you mention are also irrelevant noise that wouldn’t make a difference looking at the big picture.

      Mass labor contributes almost nothing in a relative sense to the GDP of modern industrialized societies.

      One can already see that only about 30 million Americans contribute at least three quarters of present day GDP. There is further lopsidedness at higher resolutions into that and it probably gets more extreme if we had the capability to go beyond nominal figures to truly accurate accounts real production.

      Contribution of individual workers practically follows a lognormal distribution. This trend is also reflected by income and consumption over the decades, to much chagrin. Where it can be argued that societal policy can change those outcomes there is no way around the argument that it is actually the case in production nor as many reasonable ways we can expect to change that trend in the future.

      Technology, trade, general efficiency in energy and infrastructure and capital and so on have produced most of the GDP gains since the second half of the twentieth century. All these things are so much bigger such that women entering the workforce is a much, much smaller thing and had minimal effects on the most productive parts of the economy in a raw production sense.

    • Michael Edward Vassar says:

      But total number of hours worked went way up. Also, it had previously been falling. Also, the elderly were generally low value workers. Also, college is supposed to increase productivity!

      • Randy M says:

        I’m not sure about the elderly being low-value workers, it certainly varies by industry but on the net it could go either way (up to certain definitions of elderly, of course, which I hope are well abover 60).

        The college point is an excellent one, though, for those who believe it to be much more than zero-sum signaling. Shouldn’t GDP have jumped over the last 30 years with the big increase in college enrollment?

      • But measured GDP nearly quadrupled in real terms between 1950 and 1990. Aggregate working hours less than doubled. The difference is productivity growth.

        Now, some have said GDP is fictitious, or whatever. It really doesn’t matter. It’s only a question of whether the tautologies are consistently applied in the measurements across the decades. We only know what is productivity growth (at the economy-wide level) because we count up GDP and (separately) count up working hours, and divide the one by the other.

  13. Samuel Skinner says:

    Male contribution to the work force declined during the same period women were entering it. This gives an overall drop of 10%. Other sources show different levels of a drop, but all show a reduction in men’s participation.

    http://money.cnn.com/2013/06/19/news/economy/men-workforce/index.html

    Also the 1930s was abnormally low in terms of female participation due to the Great Depression as was the 1950s.

  14. zac says:

    Most GDP growth comes from technological growth, and labor isn’t the limiting factor in technological growth.

  15. Josh says:

    I think the answer here is really simple and it actually disproves a lot of bullshit economic thinking.

    The economy is entirely demand driven.

    What does that mean in this context? Even though women were entering the work force they were still “demanding” essentially the same amount of goods (food clothing etc.) before and after they began working. Therefore since there was a negligible increase in demand, there was a negligible increase in GDP. End of story.

    • Michael Edward Vassar says:

      Not a crazy claim. Basically, when Scott pointed out that he wished that people didn’t show that they totally don’t understand Econ 101, I wanted to point out the people who do understand Econ 101 don’t understand Econ 201, so they still don’t understand basics such as one party’s savings necessarily being another party’s debt or a prediction of another party’s future consumption.

      • Josh says:

        Right, in a manner of thinking GDP is basically a measure of aggregate consumption. Thus it should be no surprise that who generates the funds necessary to effectuate that consumption is rather irrelevant.

    • ElTighre says:

      Can we get some fleshing-out of this idea? I’m well on board with the idea that aggregate demand is highly important at times like the present, but I’m not used to the idea that the supply-side is actually not a real thing.

    • How does this fit with the idea from the Two Income Trap post that employed women demand more from the money economy? (Family needs a second car, more prepared food bought because of less home cooking, women need a work wardrobe…)

  16. Ashley Yakeley says:

    That second chart is “adjusted for inflation”. How is that adjustment done?

  17. Eli says:

    total number of hours worked stayed about the same due to men entering the work force later, retiring earlier, and working shorter weeks

    No, there’s a simpler explanation than a sex divide. What’s been happening is that the poor get cut back to short hours and zero-hour contracts, while those remaining among the middle and professional classes get their hours lengthened.

    Just an example.

    Class struggle, mothafucka, do you acknowledge it?

  18. Konkvistador says:

    You do notice that if model 4 is correct, this is a point in favor of basic income right?

  19. Konkvistador says:

    “But so far I’ve never heard the conspiracy theory that women never actually entered the workforce, that all the working women you see around you are animatronic robots or carefully crafted stage illusions.”

    Actually this is equivalent to the idea that women are mostly employed in makework.

    • Sarah says:

      I’ve thought about this, but on reflection it doesn’t make sense.

      If women were mostly employed in make-work, that might reduce “real” economic growth (a population of bureaucrats shuffling papers doesn’t produce new technologies or productivity improvements) but it would still increase GDP. Even the most pointless bureaucratic job is paid in dollars and counts towards GDP.

      If you gradually added women who were mostly doing pointless things to the workforce, that wouldn’t keep GDP from growing unless something was ALSO happening to the male workforce (fewer hours worked or lower productivity.)

      [I suppose there’s also the uber-sexist theory that not only do women do pointless jobs, but that this makes it more difficult for formerly productive men to do useful things. But this seems really fanciful compared to straightforward reductions in the male labor force as in pseudoerasmus’ comment.]

      • lmm says:

        >If women were mostly employed in make-work, that might reduce “real” economic growth (a population of bureaucrats shuffling papers doesn’t produce new technologies or productivity improvements) but it would still increase GDP. Even the most pointless bureaucratic job is paid in dollars and counts towards GDP.

        If IBM pays someone $40,000 to do nothing, that person gets $40,000, IBM shareholders get $40,000 less, and the GDP remains exactly the same. No?

        • Sarah says:

          Not instantly, unless I’m misunderstanding.
          Does the share price really drop immediately? Wouldn’t it take time for the market to learn that the new job wasn’t useful?

      • Steve Johnson says:

        “I suppose there’s also the uber-sexist theory that not only do women do pointless jobs, but that this makes it more difficult for formerly productive men to do useful things. But this seems really fanciful compared to straightforward reductions in the male labor force as in pseudoerasmus’ comment.”

        Is it really that fanciful? If you grant that women have entered the labor force doing make-work you have to ask what that make-work actually is. As it turns out – it’s mostly bringing pointless jobs in bringing progressive ideology to corporations – which almost certainly would work by reducing productivity. At the very minimum think of a productive employee going to diversity training – the time he spends there being lectured is time he’s not working (the content could be a slide show of the diversity trainer’s summer vacation as far as that goes) – maximally they introduce cultural changes in the workplace that are harmful to productivity (progressives talk constantly about that sort of stuff but they spin it differently – they talk about “changing work culture” constantly).

        All of this matches up very well with what we see in the only remaining dynamic sector of the economy – technology. What’s stopping large tech companies from creating all the products that start ups do? In theory they could just take the formerly productive employees, segregate them into teams and let them create things. The fact that they can’t do that is really strong evidence that the state mandated bureaucrats (mostly women) are actually actively interfering with productive work and not just passively draining the firms.

        • Nornagest says:

          You can’t get anywhere near ~30% of the present workforce by looking at ideologically charged jobs. Properly aggregating that category would be godawfully hard, but as regards the example you gave, Google is turning up values on the order of half a percent of revenue for typical HR departments. Payroll as a whole is usually around 20-30%, and there’s more to HR than compliance with ideology/bureaucratic requirements, so this case is negligible.

          You might be able to break single-digit percentages by looking at growth in the public sector and in those parts of the private sector that interface with it, but putting that all down to direct propagation of ideology is absurd. That said, regulatory burden probably does have something to do with stagnation in certain sectors — I just don’t think it has much to do with the main progressive talking points at this level.

        • Sarah says:

          http://stats.bls.gov/mlr/1997/04/art2full.pdf

          This breaks down where women entered the workforce between the 70’s and 90’s.

          The biggest jump seems to be in managerial work, where the population of women almost doubled. The changes are much smaller in working-class jobs.

          Managerial work, btw, does *not* include secretaries (who have been mostly female for most of the 20th century). We’re looking at marketing managers, HR managers, purchasing managers, and so on.

          The stereotype of “in the 70’s, women went into HR” is not too far wrong.

          There are more people in managerial jobs than there were in the 70’s. By a lot. And a lot of those new managers are women.

          Is adding more managers bad for growth? Maybe. I’m going to hesitate before assuming that the people running large corporations are all idiots.

          But I don’t see how hiring a bunch more people would completely fail to show up in GDP numbers unless
          a.) people left the workforce
          b.) prices went up by enough to push the GDP deflator up

        • http://www.dol.gov/wb/stats/Occupations.htm

          The proportion of government-mandated managers doesn’t seem very high.

  20. Avery says:

    Forgive me my confusion, but what is causing the confusion? I see two graphs with upward sloping lines–both the % of women in the workforce and GDP went up at the same time.

    • ckp says:

      Because it’s a near-perfect exponential function that you could theoretically have extrapolated from before women started entering the workforce en masse.

      • Jack says:

        I think Avery’s point is that the graph of female workforce participation also shows an exponential trend – there isn’t an obvious jump. So it’s plausible that the increase in working women was baked into the GDP growth trend. But in that case we have to explain why the GDP trend continued after the fraction of women in the workforce leveled off.

        Aside – does the ‘percent of women who work’ graph look suspiciously smooth to anyone else? How the hell do they get data on this from 1790, anyway?

        • Paul Torek says:

          Did the GDP trend continue after the fraction of women in the workforce leveled off? I don’t think so; the exponential GDP trend is definitely lower for the past few decades. There might be a few years where women were mostly-in and GDP was still rising fast, but that could be a blip. See hypothesis 3 by Athrelon.

  21. Army1987 says:

    Maybe the increase in the number of women in the workforce compensated for the increase of men in college and retirement.

  22. Pingback: Outside in - Involvements with reality » Blog Archive » Chaos Patch (#16)

  23. Devin F says:

    First, GDP is a useless number. I wrote a detailed tear-down on my own blog. It is fundamentally impossible to determine in an objective, quantitative way how many times better a 2014 computer is than a 2004 computer. Do you go by processing power? Or by time it takes a person to complete a task? What task? Writing a Word document? Editing a picture in photoshop? The GDP calculations is filled with complicated techniques to create answers to these impossible questions. The techniques look plausible – but if you examine them with rigorous logic you will see the numbers they produce are entirely subjective and arbitrary. In reality, the economists will just keep adding up adjustments until the GDP number overall fits with people’s intuitions. The end result is meaningless.

    “Women contributed through unpaid labor in the home, and their paid labor substituted for that but didn’t add to it. But as far as I know GDP only counts paid-for goods. Not only should women’s labor in the home not have counted, but GDP should overcount the benefits of putting women to work because paid daycare and so on appear as valuable new services.”

    That’s not quite how it works.

    The GDP is calculated by taking total nominal income (or expenses, they should be the same, but income is easier to think about) and then adjusting by a price index (the GDP deflator). Total nominal income is determined by the classic econ equation – quantity of money times the velocity of money. In other words, quantity of labor has zilch to do with nominal GDP.

    So the impact of women entering the work force would need to be seen by the prices in the GDP deflator being lower than they otherwise would. But when women moved from providing childcare for free, to entering the workforce and paying for childcare, demand for childcare went up faster than the supply went up (changes to regulations and young folk living further away from grandparents also has driven child care prices up). Prices for childcare actually have gone up, and so the impact on the GDP deflator would be toward negatively affecting “growth.”

    “Somehow in total contradiction to usual economic theory, all gains made by women came out of the pockets of men, leaving the same growth as would have happened anyway.”

    That’s because standard GDP theory views economic output as a unitary concept, and does not think about what output really means on the ground.

    Are the women who entered the workforce making cars? Building houses? Inventing new gas fracking processes? Designing solar cells? Those productive professions are still massively male dominated. Most women went into fields like marketing or sales, which are zero-sum. Many others went into healthcare, HR, administration, and education. Those fields have all seen greatly rising costs in the last forty years. If more labor enters a free-market sector, costs should go down (and GDP deflator will go down). If more labor enters a bureaucratic sector, costs may go up, as the workers politically agitate for more make-work in order to employ everyone.

    This is another way the GDP stats are useless. A solid portion of female entry into the workforce went toward increasing the quantity of healthcare produced. But was this real output, that made us all better off? Do we have better health care due to more spending on healthcare labor? Or was much of the spending driven by bureaucratic inefficiency, of expenses rising to meet whatever was budgeted (last time I went to an emergency room for a simple bump in the head, I had to tell my story to no less than eight different nurses and assistants)? The quantity of schooling has likewise increased since women entered the workforce. Is this real output? Or is it just bureaucratic make work? There is no way to tell mathematically, you can only use your own judgement.

    • Michael Edward Vassar says:

      Seconded, but it’s important to note that Scott is pointing out that GDP is a useless number with a totally false value, which is also true and is a VERY important point for would-be EAs etc.

    • Multiheaded says:

      But was this real output, that made us all better off? Do we have better health care due to more spending on healthcare labor? Or was much of the spending driven by bureaucratic inefficiency, of expenses rising to meet whatever was budgeted (last time I went to an emergency room for a simple bump in the head, I had to tell my story to no less than eight different nurses and assistants)? The quantity of schooling has likewise increased since women entered the workforce. Is this real output? Or is it just bureaucratic make work?

      As far as I, being almost infinitely ignorant on the matter, have heard, healthcare in Britain has gotten better (in large part due to the quantity of publicly funded services fixing their permanent undersupply to poor people) *and* more bureaucratic since the creation of the NHS. Likewise, some services like railways have been complained about after the Thatcher privatizations.

      Would any British readers care to comment?

      (Also, insert the requisite insults here.)

      • soru says:

        The UK NHS runs on health-oriented bureaucracy; there are targets for how many people die of this, or suffer from that, and management attempts to change those figures.

        That’s different from a money-oriented bureaucracy, where the same kind of managers make the same kinds of decisions, but with an extra constraint that everything must have an individualised dollar cost that can be charged to a specific person (without getting sued). That’s a much more constrained optimisation problem, so naturally it costs more and delivers less.

        As a result, the US health care industry is able to make a more credible threat of failing to deliver care, and so is more successful at capturing a larger amount of GDP than systems with other priorities. Adding workers, female or not, would presumably decrease its effectiveness at making that threat, and so reduce the amount it is able to charge.

        This is in contrast with situations where a money-oriented bureaucracy typically does better than a goal-oriented one. In, say, food retail, the nature of the business means there is no real ability to deliver a credible threat of not being able to get any food.

        So people actually have to want the product, not fear not getting it. Given that, adding more workers not only helps produce a better result, but will increase, not decrease, the gdp share captured by that sector.

        Perhaps overall the two cases balance out and so there is no net effect of labour on GDP?

        • Multiheaded says:

          That’s a really fucking impressive argument for social democracy! Whoa! Thank you!

        • Sarah says:

          The reason, I presume, why food providers can’t credibly threaten to starve the population is that the food industry is decentralized.

          Does medicine *have* to be centralized?

    • ADifferentAnonymous says:

      This inspires the fascinating hypothesis that many facets of liberalism reduce productivity in such a way as to mostly cancel out the positive effects of women working. That is, an otherwise-politically-neutral increase in female employment (Rosie the Riveter?) would increase productivity, but in the postwar era liberalism is a common cause of female employment and a bunch of anti-growth crap that pretty much washes out.

  24. Randy M says:

    “This seems a lot like the question “do immigrants entering the workforce drive down native-born wages?” to which economists tend to answer “no” with more or fewer caveats. But the economists’ explanation for the immigrant effect is that in addition to producing, immigrants also consume, increasing demand. ”

    I wonder if the economists consider that in many cases the immigrants were consuming before, if the destination country had had net export balance?

    • Paul Torek says:

      And what about the labor/capital balance? Seems to me that if the ratio shifts to more labor per capital, the capitalist take rises somewhat at the relative expense of workers (at least until capital increases to some equilibrium level).

    • pwyll says:

      I think Borjas claims that immigrants entering the workforce typically drive down wages for low-skill work, while raising wages for higher-skill work.

  25. roystgnr says:

    Could you expand the phrase “income inequality explanation“? By itself the phrase “income inequality” refers to observations, not explanations, and I wasn’t aware we’d puzzled out all the explanations for it yet. (in fact, most of the people talking about it don’t even seem to clearly understand the observations yet…)

    But the economists’ explanation for the immigrant effect is that in addition to producing, immigrants also consume, increasing demand.

    Economists who think that demand is absolute rather than relative should have picked a line of work they were better suited for. I demand a fleet of interstellar spaceships, but alas, my ability to afford that demand is insufficient to affect the nuclear rocket industry. You can’t talk about increased demand for X without expressing it in terms of willingness to trade Y.

    A priori we’d expect the effect of immigration to be the same as the effect of improving technology: higher production overall, but with the gains going to capital (including human capital) and with net losses for unskilled labor. Since that seems to be roughly what we see I’m not sure how far we need to stretch to postulate and explain a different effect.

  26. David Foster says:

    “Percent of women who work” probably did NOT include women working on family-owned farms…not an insignificant number prior to 1950 or so…and value of farm produce was certainly included in the GDP calculations.

  27. Army1987 says:

    <rant>Why the f— is the y axis of the GDP plot not logarithmic? Who the hell thinks the difference between $2M and $4M is no more important than the difference between $12M and $14M?</rant>

  28. johnwbh says:

    Possible natural experiment would be to look at other countries where major changes to gender workforce participation happened at different points in their history and economic development, so we don’t get the same confounders there were in post war USA.

  29. JME says:

    One issue I have here is that it effectively treats “trendline growth” as a sort of acausal, inevitable process. Yes, the economy has grown fairly steadily, but all of that growth has reasons! If you were looking at the effect of, say, containerized shipping on economic growth, you could look at its introduction (in the 50s) and say “where is its effect on trendline growth?” Well, the answer is that trendline growth is the aggregation of many improvements in technologies or systems of social organization. The introduction of containerized shipping — along with many other advances — is why the trend is generally upward in the first place.

    The same logic might apply — crudely — to women entering the workforce. Economic growth requires change, not just keeping on and waiting for the trendline to move up. Some of the changes might be technological, some social. It is possible that the entry of women into the work force was a reason why the upward trend continued.

    (Note: I don’t want to overstate the degree to which economists actually understand economic growth or total factor productivity. There are lots of mysteries involved, and I don’t mean to definitively say that without women’s entry into the workforce growth would have dipped below the previous trend. I just think that economic growth must have some causes, and women’s entry could be one.)

  30. But so far I’ve never heard the conspiracy theory that women never actually entered the workforce, that all the working women you see around you are animatronic robots or carefully crafted stage illusions.

    Sure you have. Neoreactionaries all the time say that most women in the workforce are doing makework created for women, for example HR and government employment, rather than the real productive work that men do.

    • Not to mention that women were working before, and the equivalent value of a man’s wages was worth both of their labor. Putting them in the ‘workforce’ merely rationalizes that value, which must in that case cut the rationalized value of the man’s wages, though if both man & woman are in the workforce the total should be the same.

      – In the old system, bachelor men and lazy wives were partial free-riders.

  31. Phil Goetz says:

    GDP is adjusted for inflation. If women entering the workforce leads to families having more disposable income leads to inflation, the effect would be invisible in GDP. Check the inflation rates versus women in the workforce. The US had its longest period of high inflation in history from 1968-1983.

  32. RJMeyers says:

    Assuming that women are roughly as productive as men, and noticing that their workforce participation rate is both an exponential function (as is GDP growth) and has its most dramatic rise from the 1940s – present (as does GDP growth), it makes sense that when you add them into the economy you would still end up with a relatively smooth exponential fit for GDP.

    I would only expect a noticeable discontinuity in GDP if their participation exponentially shot up very early (sometime in 1790s – 1900, when GDP was still in “liftoff”) or if their entry into the workforce was much more dramatic, almost a step function. As it is, when you account for tech and knowledge being the major part of GDP, women entering the workforce are a minor, time-aligned, directionally-aligned (increase, not a decrease) exponential contribution to a much more powerful exponential trend.

  33. Pingback: GDP is ageist but not sexist | Pseudoerasmus