18 December 2012

Jaws of the Snake

Writing in the New York Times last week, Erik Brynjolfsson and Andrew McAfee argue that “a wonderful ride” began to unravel in the late 1990s when employment growth became “decoupled” from productivity growth.

Their contention is that something fundamental has changed in the economy over the past decade, illustrated in the following graph by the increasing gap between gains in productivity and employment, described ominously as the “jaws of the snake.”

A closer look at the data shows, however, that the “jaws of the snake” have been open for more than 30 years. More fundamentally, rather than a “great decoupling” between trends in employment and productivity, it is clear that productivity and employment have been diverging for a very long time as the composition of the economy has changed.
And here is how it concludes:
So when we look into the “jaws of the snake” it is not the perverse consequences of rapid technological change on the economy, an influence which Brynjolfsson and McAfee implies that “we also need to start preparing for a technology-fueled economy that’s ever-more productive, but that just might not need a great deal of human labor.”

To the contrary, an analysis with more appropriate data suggests that the rise of the machines is not the main reason for our current economic challenges, and a world where labor is less needed is not yet upon us.
For the meat in the sandwich, head on over to the BTI, and feel welcome to come back here and comment and critique.


  1. McAfee goes back 30 years to the introduction of the PC to the U.S. workplace in his narrative elsewhere:
    Declining Labor Share

  2. Nice article.

    I personally find little to fault in your reasoning. However, there is a noticeable lack of an alternative reason for growing unemployment/divergence of TFP vs. employment.

    Do you have a personal view?

  3. I do not understand the logic coupling technological change with total employment. Money is a medium of exchange, period. There is nothing else you can do with it, other than give it to another human, usually, though not always, in return for some product or service. If
    technology releases Alice from the need to give a dollar to Bob, all that means is that she will be looking to find an excuse to give it to Carla. (She can put it in the bank, but that just transfers the responsibility for finding the right exchange opportunity to that bank; it doesn't change the general picture.) It doesn't matter how smart machines get. Even if you imagine a world in which machines are
    better than humans at everything, humans will still be hiring other humans do something, because they will have nothing else to do with
    their money.

    One can imagine worlds where this is not true. These would be worlds in which people do not care about money, burning it in the back yard whenever they get some, or worlds in which machines are fully economically autonomous, with their own interests, which they advance in competition with humans I can imagine a world in which everyone devotes all their income to philanthropy. But none of these are at hand.

  4. Roger,

    I don't really see how using TFP leads to a fundamentally different conclusion, other than to change the time frame. It does more clearly suggest, however, that computers and mobile media have fundamentally changed how we create wealth, just as McAfee says. I have argued this multiple times on this blog, pointing out that the widening wealth inequality has litle correlation with government policy and everything to do with the information revolution. And this is likely to continue for a long time.

    All in all, it's not a bad thing. Employment will continue to rise slowly but steadily, in line with population growth, while we all become fabulously wealthy (no I'm not kidding).

  5. -3- Fred Hapgood:

    You seem to have an assumption that the money we get comes from productive employment. That's obviously not the case. Fortunately, we're productive and wealthy enough that many people are able to avoid this fate. Sadly, many choose or are forced to do so.

  6. MattL --

    I'm not following your point. The argument has to do with the relationship of machines to jobs. My claim is that the total percentage of the population that is employed will not change much over time, regardless of how smart machines get. (As is testified by the last several hundred years.) Do you differ? Why?

  7. -6- Fred,

    You seem to be saying that the graphs are not correct. Let's start by agreeing that it has changed. Maybe that's within your tolerance for "not much," but it has changed.

    But to answer your question: Higher productivity means that a single person's output can earn more now than it used to. In fact, much more than is strictly required to simply "get by."

    That surplus can be used in several ways. It can be spent on more luxury, etc, which is in line with your argument. And to a large extent, it does. But some of it is also used in other ways, such as donations to charity or as taxes which are subsequently redistributed (NB: I'm not saying that's all taxes do!).

    Some of that redistribution will allow some people to avoid having to work in order to get by. That's not a value judgment, but an observation. Our surplus from improved productivity is a lot bigger than it used to be, and more people are able to get by without working and by relying on redistribution of the wealth of those who are productive.

    I would add that machines getting more capable is more of a quantitative than qualitative difference. So, it's not that there's something magical about machines, but that we've simply become a lot more productive in part due to those machines.

    Since we've already observed this, it's not unreasonable to presume that the trend may continue into the future. Which is not to say that it must, of course.

    If you still disagree, you have to explain how the decreased participation in the labor pool in the US (and increased disability rolls) don't show that the total percentage of the population can change.

  8. #7 Mattl

    ==Higher productivity means that a single person's output can earn more now than it used to. In fact, much more than is strictly required to simply "get by."==

    You miss the point that higher productivity leads to a DEcrease not INcrease in purchasing power.

    Your point would be valid in a system where the productivity gains were not appropriated by the owners of capital, i.e. where they passed on to wages.

  9. -8- Reiner Grundmann,

    I never said that owners of capital do not benefit from increased productivity. And if it didn't, then what's the point of owning and growing capital?

    But human capital is no different. Here was a recent example of increased purchasing power gleaned from an old Sears catalog.

    I don't know what system you live under, but I don't recognize it.

  10. Actually, there is a revolution that started in 1980 - that had little to do with computers: the financialization of the US economy.

    FIRE - Finance, Insurance, Real estate, and Education all started to get rolling, and all of these sectors are primarily 'services'.

  11. When production per man-hour goes up, fewer employees produce more goods and services. But the growing cadre of unemployed cannot afford to consume, and demand for non-essentials goes down. There are fewer people to consume what is being produced. It would appear that Malthus’s model of the necessary exhaustion of non-renewable resources, as populations grow exponentially, is finally leading to technological unemployment and a market that cannot be repaired with either increased demand or supply.

    Possible means to resolve this problem include: an end to population growth; shorter work-weeks; greater employment in human services; more research in the arts and sciences; increased welfare-like entitlements; and generally, a switch from – using capital growth and GDP – to quality of life (survival, convenience and comfort) as the main motivator for steering economies.

  12. Hi Roger,

    Your article concludes, "To the contrary, an analysis with more appropriate data suggests that the rise of the machines is not the main reason for our current economic challenges, and a world where labor is less needed is not yet upon us."

    However, labor is less needed in the U.S., isn't it? Isn't the labor participation rate falling significantly?


    If the labor rate is falling significantly, and the economy is not shrinking, isn't the U.S. a place where labor is less needed?

  13. There are many reasons for lower employment.

    Globalisation and its acceleration under carbon trading.

    Workers being forced to work longer hours to suit employers (assuming we are counting heads).

    Higher standards. In the 1970s, there were masses of low level jobs for alcoholics and other misfits.

    There were unions to protect jobs in the past.

    The idea that welfare level jobs like Walmart and the social care jobs we have here replace well paid employment is a nasty lie and undermines the argument.

    The political solution has been to make the working classes disappear from economics and demonise them as welfare cheats.

    There is also the fact that, particularly in the United States, a political system, designed from day one to exclude citizens from power, has consolidated into a highly efficient and arguably criminal public/private hybrid monster (see the entire US financial system) to suck wealth and power from the mouths of the poor and into the bellies of the rich.