International Monetary Fund

Please send us your feedback

    IMFSurvey Magazine: IMF Survey Interview

    Subscribe Share This
    A robot developed in Japan to care for the sick. But in most of the care sector, robots still cannot substitute for human beings (photo: Gellie/Corbis)

    A robot developed in Japan to care for the sick. But in most of the care sector, robots still cannot substitute for human beings (photo: Gellie/Corbis)


    Demand for Tech-savvy Workers Fuels Inequality

    IMF Survey

    February 22, 2013

    • Rising income inequality more marked in advanced economies
    • New technology demands higher skills from workers
    • Policymakers urged to balance opportunities for all employees

    Over the past several decades, advanced economies have seen a striking rise in inequality. In the United States, the top 0.1 percent of households receive more than 10 percent of national income—more than double the figure of 30 years ago.

    In an interview with IMF Survey, David Autor, Professor of Economics at the Massachusetts Institute of Technology, argues that this trend is driven by rapidly developing technology, which has made highly educated workers more valuable within the workforce.

    IMF Survey: Why does inequality matter?

    Autor: For a long time economists were saying that equality did not matter that much. Many people say you need some inequality to provide incentives, and that if you do not have incentives, people will not work hard.

    And how much is too much inequality? Who knows? I think that view has started to change over the last decade. United States Supreme Court Justice Potter Stewart once famously said [paraphrasing], “I don't know what pornography is, but I know it when I see it.” Many economists are now saying: we don’t know what too much inequality is, but it looks like we might have it.

    I think a reason to be concerned—the reason I am concerned—is because too much inequality, although it creates great incentives for working hard and investing, can create a society that is ultimately more dynastic, meaning that the income of the household that a person is born into has a great effect on his or her life chances.

    Thus, inequality affects not just material consumption but also, for example, access to good schools, because getting a child all the way through a good college requires almost two decades of investment, and not all households can do that. I do not have strong feelings about inequality of income as long as people are not destitute, but I worry very much that rising inequality of income will reduce economic mobility in the next generation.

    IMF Survey: While economists have been concerned about inequality, the subject has also been in the popular imagination, and in the news, so this is obviously something that people are thinking about now. Can you quantify how much inequality has really risen?

    Autor: Starting in the early 1980s, economists observed that the return on education, that is, inequality between college and noncollege workers, rose a lot. Over the last 25 years, the college/high school wage differential has roughly doubled, so that is one common measure of inequality. In the U.S., that corresponds to a rise in incomes of college workers, and a real decline in the earnings of noncollege workers, especially noncollege males. It is not simply that some groups are doing well and others are doing very well, but that some groups are doing pretty well and some groups are doing pretty badly. That is one reason to be concerned.

    Another factor capturing the popular imagination is the work of Emmanuel Saez at the University of California at Berkeley and Thomas Piketty of the Paris School of Economics. They looked at the concentration of top earnings using tax records and saw that the change in the share of national income accruing to the top 10 percent or 1 percent of households has just risen astronomically since the late 1970s, and rather continuously.

    So at this point the top 10 percent of households account for more than 50 percent of all national income, and the top 1 percent of households account for about 23 percent of national income. This is more than a doubling of what it was 30 years earlier. In many advanced countries there has been rising income concentration, but the U.S. is, by far and away, the extreme.

    IMF Survey: Why is it that the return on education is rising, in other words, the value of getting an education is higher in terms of increased wages and such? Why has that been changing?

    Autor: There are a few factors. Demand for skilled labor has been rising for as far as we can measure it—at least 90 years, probably more—and it is not hard to understand why, because society has become increasingly dependent on engineers, scientists, and on people who can read, write and communicate.

    Many of the labor-intensive tasks of 100 years ago—in farming, in repetitive production activities, even in labor-intensive artisanal activities—have now been supplanted by technologies. Whereas the importance of creativity, analytical capability, creating new technologies, and solving problems have all become more essential to things we cannot automate, the things that we require humans to do.

    What has changed—actually the biggest factor that is not often talked about—is that there is a slowdown in the supply of highly educated labor. So what spurred the very rapid rise in return on education in the 1980s was this deceleration of supply. It is not that there were fewer college graduates, but that there were relatively less of them when the demand was rising rather steadily.

    IMF Survey: Related to this is the idea of “employment polarization,” where people at the bottom are earning less while people at the top are earning more. Can you explain in detail what is meant by polarization?

    Autor: The term “employment polarization” refers to the sort of simultaneous growth of high-wage, high-education, high-skill occupations, and low-wage, low-education, low-skilled conventional occupations at the expense of the middle. When many people think of technological progress they think it is all onward and upward—the top is rising, the middle is rising, the bottom is falling and there is an increasing demand for skills everywhere.

    So, the thinking is that people would be doing much less “drudge work” and more office work, and more creative thinking for a living. But, in fact a lot of the drudge work is still necessary. For example, if you think about driving vehicles, or cleaning, or jobs in food service or personal care—these are all things that require flexibility, adaptability, sightedness, language and visual recognition—skills that have proven extremely hard for computer science. They have proven to be much harder to automate than anyone had expected.

    We do not have domestic robots. We have great computer chess players. We have Google to search for information. Service jobs have stayed around and grown, and are very difficult to trade internationally because they often require in-person interactions. So, these jobs are not at the moment very susceptible to technological advance or to trade. Further, because the skills used are generic, they tend not to be high-wage jobs—that is, because there are a lot of people who can potentially do them.

    On the other hand, manufacturing employment has declined substantially for two reasons: the first is automation. There is a lot less labor used in manufacturing because firms increasingly prefer sophisticated machinery to hands-on assembly.

    The second is trade. Growing trade penetration with developing countries, particularly China, has reduced employment in a lot of labor-intensive U.S. manufacturing, textiles, and footwear assembly lines. At the same time, computerization has substantially reduced employment in clerical and administrative support occupations.

    These production and office support occupations have typically provided good, middle-waged jobs for people who have a high school or maybe a college education. They are not top-flight executive jobs, but they often have some employment stability, with skills that are specific to an employer, or techniques that are specific to a firm.

    Polarization is of concern to people and rightly so, because there is an elite strata of jobs that are great to have. However if you are not in them, there may be fewer rungs in the ladder in between the top and the bottom.

    IMF Survey: These changes are driven by very big, unstoppable forces such as growing international trade and technology. What is the proper policy response?

    Autor: There are very few people who are willing to say we should try to stop technology—you could try. More people would say we should try to stop trade. I do not support that. I would suggest that in order to realize the benefits from trade and technology, it is important to help insure some of the losers against the losses, by, for example, providing job training, or some level of income security, or an assurance of ongoing health care coverage.

    Arguably, a country where the social safety net is more expansive is one where people would be less fearful about adaptation because the personal cost they bear might be smaller.

    Should we be smashing technology? No. Should we recognize that there are people whose livelihoods are harmed by technological advances? Yes, we should recognize that. For most people, their wealth is tied up in the value of their human capital, and that is their most valuable asset. If, all of a sudden, your human capital is no longer scarce—in other words, whatever an individual could do manually before can now be done for 25 cents with an iPhone app—that may be good on a global scale, but it is not good for an individual’s earning potential.

    It is not a mistake to recognize that technological change can create winners and losers. It is a mistake to fight technological change, so it would be better to find some way to channel this change into prosperity.