Home > income inequality > Technology didn’t kill middle class jobs, public policy did

Technology didn’t kill middle class jobs, public policy did

from Dean Baker

A widely held view in elite circles is that the rapid rise in inequality in the United States over the last three decades is an unfortunate side-effect of technological progress. In this story, technology has had the effect of eliminating tens of millions of middle wage jobs for factor workers, bookkeepers, and similar occupations.

These were jobs where people with limited education used to be able to raise a family with a middle class standard of living. However computers, and now robots and other technological innovations are rapidly reducing the need for such work. As a result, the remaining jobs in these sectors are likely to pay less and many people who would have otherwise worked at middle wage jobs must instead crowd into the lower paying sector of the labor market.

This story is comforting to elites because it means that inequality is something that happened, not something they did. They won out because they had the skills and intelligence to succeed in a dynamic economy, whereas the huge mass of workers that are falling behind did not. In this story the best we can do for the left behinds is empathy and education. We can increase their opportunities to upgrade their skills in the hope that more of them may be able to join the winners.

That’s a nice story, but the evidence doesn’t support it.  My colleagues Larry Mishel, John Schmitt, and Heidi Sheirholz, just published a paper showing that the pattern of job growth in the data doesn’t fit this picture at all.  This paper touches on a wide variety of issues related to technology and wage inequality, but first and foremost it shows that the story of the hollowing out of the middle does not fit the data for the 2000s at all.

Since 2000 the increase in employment has occurred almost entirely in low-wage occupations. There has been a decline in relative employment for both workers in middle wage and high wage occupations. If this occupational shift story explained trends in wages we should expect to see sharply rising wages for retail clerks, custodians and other workers employed in low-paying occupations.

Of course we see the opposite. Workers in these occupations continued to lose ground in the 2000s as they did in the prior two decades. Their wages barely kept pace with inflation over the last three decades.

The paper makes an impressive case that technology is not the main explanation for the rise in inequality that we have been seeing. In fact, even M.I.T. economics professor David Autor, the leading proponent of the occupational shift story concedes this point. He was quoted in an NYT column as saying of the view that technology explains inequality:

“It can suck all the air out of the conversation, … All economists should be pushing back against this simplistic view.”

Given the evidence compiled by Mishel et al, it would be difficult to maintain that technology has been the main culprit in the upward redistribution of income that we have seen.

It is not difficult to identify other potential culprits. Trade would certainly rank high on the list. A trade policy that quite deliberately puts factory workers in direct competition with low-paid workers in the developing world, while protecting doctors and other highly paid professionals, would be expected to redistribute income from the former to the latter.

The weakening of unions is likely also an important factor. The private sector unionization rate in the United States has shrunk from over 20 percent in the 1970s to less than 7 percent at present. In the same vein, the deregulation of major industries like airlines, telecommunications, and trucking has been another factor putting downward pressure on wages. The higher unemployment rates we have seen, not just in the last five years but in the last 35 years, compared with the early post-war decades, has also weakened the bargaining power of workers at the middle and bottom of the pay ladder.

We have also seen big changes that contributed to growth of income at the top. The highlights in this category would be deregulation in the financial sector and the changes in corporate governance that pretty much allow top management to write their own paychecks.

The big difference between the items listed above and the technology/occupation story is that this is a list of items that involve policy changes. If this list (which could be extended) explains the growth in inequality over the last three decades then it means that inequality was a result of policy. It was not something that just happened; it was something that we did or was done to us.

That presents a very different policy agenda for addressing inequality. No one would quarrel with the idea that our children should get a better education, but if a lack of skills was not the cause of inequality then more skills will not be the solution. Rather we might look at an agenda that would rein in finance and CEO pay, restore the strength of labor unions, and include a more balanced trade policy.

This agenda wouldn’t just mean empathy from those on top, but also lead to them losing some of the gains from the last three decades. Therefore we are likely to hear stories about technology destroying middle wage jobs for some time longer, even if the evidence doesn’t back up the stories.

  1. Podargus
    November 30, 2013 at 7:22 pm

    Globalization,free trade, deregulation,privatization,union bashing,immigration – wind those abominations back and introduce a steep rate of progressive taxation and you would have made a start on building a decent and sustainable society.
    But you won’t see any of that coming from the administration of the feckless fool who is masquerading as your President.

  2. BC
    November 30, 2013 at 7:37 pm

    Amen, Podargus.

    US real wage and salary disbursements per capita (same as in 1998-99), including the top 10% who receive 45% of US income:


    US real private average hourly earnings per capita for the bottom 90% working-class (non-supervisory and goods-producing employment) employees (lowest in 50 years and back to near the lows of 2008):


    US real wage and salary disbursements per capita and US full-time, private employment per capita (same as 1987):


    The US has not created a net new full-time private sector job per capita in 26 years (the fall off of the US oil production plateau and the onset of unprecedented US supranational firms’ investment abroad, coincident with deindustrialization, financialization, and feminization of the US economy).

    Real wages and salaries per capita are at the same level as 14-15 years ago.

    Real working-class wages per capita have fallen 6% since 2000, 10% since 1990, and nearly 40% since 1973 (far worse in purchasing power terms after higher payroll taxes).

    For a growing share of the bottom 90% of US households, it no longer pays to work after taxes and debt service (“rentier taxes”). That is, it costs more to work to be able to afford the costs of obtaining and retaining paid employment after taxes, debt service, and price increases than the net for subsistence after all costs of working in order to work.

    In this sense, the US does not have an unemployment or underemployment problem per se but costly OVEREMPLOYMENT and UNDERPAYMENT of wages and salaries after taxes and debt service. That is, we have too many people working for low pay and wasting scarce resources and increasing costs of housing, energy, medical serices, child care, gov’t, etc., just to maintain the wasteful hierarchical system of concentrating upward flows to the top 0.1-1% and their global imperial military and “trade” regime.

    The study of labor economics in the US has all but disappeared from the economics discipline, which is a bloody da@n shame and tragedy for the bottom 90%+ of US households and families.

  3. November 30, 2013 at 8:36 pm

    In a letter to the Editor of the Economist a week ago, I wrote:
    in your article “Labor pains” you pointed out that the share of wages in national income has fallen after being nearly constant for decades after the second world war. During the post-war decades the middle class prospered because of the full-employment policies started by Franklin Roosevelt and continued by both Democratic and Republican presidents and Conservative and Labour governments in Britain.

    In this period the growth of union power, enshrined in legislation and policies, pursued the sharing of monopoly rents and profits of corporations with their workers. By the 1970s, however, the seeds were sown for the beginning of the end of middle-class prosperity. The anti-union policies of Ronald Reagan and Margaret Thatcher made it socially and politically popular to see unions as the villains in the economy. This was quickly supplemented by firms outsourcing to foreign countries where an hour’s worth of labour was paid a much lower real wage.

    But now, a new threat is growing that will further hollow out the middle class and make even more significant differences in the distribution between the top 1-2% and the rest of society. This threat is automation. You correctly indicate that policymakers should think about broadening capital ownership as a way of boosting income to workers and restoring a prosperous middle class.

    For a creative approach to restoring middle-class prosperity, I recommend the work of Professor Robert Ashford in the forthcoming issue of the Journal of Post Keynesian Economics called “Beyond Austerity and Stimulus: Democratising Capital Acquisition With the Earnings of Capital As a Means of Sustainable Growth”. Professor Ashford proposes a capital-ownership broadening policy that big companies adopt to produce enhanced earnings for their employees, customers, and other poor and middle class people; enhanced corporate profit and growth; reduced need for welfare dependence; and enhanced sovereign creditworthiness.

    Paul Davidson

    • November 30, 2013 at 10:05 pm

      “Democratizing Capital Acquisition” sounds suspiciously like Louis O. Kelso’s ESOP panacea from the 1960s and 1970s, an outgrowth of Kelso’s and Mortimer Adler’s “Capitalist Manifesto.” Kelso was indeed an intriguing character and his ideas were refreshingly unconventional, albeit not entirely coherent or convincing.

      This notion of “democratizing capital” is based on a fallacy of misplaced concreteness. “Capital” is a relationship, not a thing (a machine or a supply of raw materials), and it is a manifestly unequal relationship. Democratizing capital thus makes about as much sense as democratizing slavery. Only half of the population, minus one, can ever be “above average” in anything — wealth, income, I.Q, height or whatever.

      • November 30, 2013 at 10:19 pm

        Actually, I should have been more direct than “sounds suspiciously like” Kelso. Robert Ashford is a proponent of Kelso’s “binary economics.” Here’s a little anecdote from Kelso’s reminiscence about the time he pitched his ESOP idea to Louisiana Senator Russell Long, son of Huey and tireless champion of tax breaks for the oil and gas industry:

        “The senator was very patient, asked many questions,” Kelso recalls. “The waiters were standing around waiting for us to get out. It must have been around midnight. Then the senator said to me, ‘Are you saying that using financing techniques based on a two-factor (labor and things) economic theory can make haves out of the have-nots without taking it away from the haves?’ I said, ‘Senator, you put me to shame. I take three hours to explain something and you cover it in a sentence.’ ‘That’s the kind of populism I can buy,’ he said.”

        Nothing says “con game” like “making haves out of the have nots without taking it away from the haves”!

  4. BFWR
    November 30, 2013 at 8:54 pm

    The wise thing to do is embrace BOTH technological innovation AND the unemployment it effects with a universal dividend and compensated retail discount. That way you get out in front of BOTH the half brained economic thinking AND the true source of the problem…the lack of symmetry in finance generally and the lack of a countervailing form of credit issuance specifically in CONSUMER finance. The BOTH/AND perspective saves a lot of intellectual back and forth and back and forth and back and forth. And when as a species we are facing numerous looming crises….that might be advantageous.

  5. BC
    November 30, 2013 at 9:53 pm

    Paul, most employment in the US, for example, is by small firms (including most employment “growth” resulting from the growth of “new” small firms) with little surplus capital in excess of production, including labor compensation. Most of the firms’ capital is in the form of structures, equipment, and non-home business equity owned by the proprietors.

    Large, equity-, capital- and cash-rich supranational firms that can afford to offer up equity shares in lieu of, or as a supplement to, wage and salary compensation hire only a small share of the total US workforce at $450,000 in revenue/employee, far in excess of the revenue/employee of 80-90% of firms.

    But, yes, accelerating automation of a growing majority share of paid employment in the service sector will further devastate the after-tax purchasing power of the lower- and middle-income working class and eventually the professional middle class next 9% below the top 1%, those who remain dependent upon the successful functioning of the real economy for the bottom 90%+.

    If only the largest 25-100 to 300 firms can afford equity sharing as compensation to employees, including executives, that leaves 80-85% of the population without any access to income and purchasing power from equity share and no means to change the situation.

    The other issue is that historically equity returns after price changes and currency effects yield on average over the very long term no more than the average dividend payout (before fees or taxes), which has been ~4% but just 2% since the 1990s bubble market valuations that persist today. That 2% return (again, before fees and taxes) from current high valuations is THE BEST return equities will produce over the next 10 to as many as 20 years, given the precedent of returns following such high valuations as we have today. The 3- to 4-year returns along the way, however, are likely to be negative.

  6. kiwichick
    December 1, 2013 at 12:46 am


    we don’t need as many people

    introduce policy’s to slow population growth

    most developed countries are overpopulated

    USA , for example has an estimated sustainable population of 150 million compared to the
    current population of 315 million? , and is increasing that population at 0.8%?? / year or

    250,000 extra people / year or 20,000 / month

    • BC
      December 1, 2013 at 1:48 am

      kiwichick, you’ve pointed to the proverbial Great White Elephant in the room whose trunk and tail are wrapped around our collective throats, heavy foot depressing our skulls, and weight risking collapsing the building, yet most of us blame the shortage of oxygen and the headache we’re experiencing on the lack of defict spending and “pro-growth policies” by politicians.

      Or perhaps there is a new anti-elephant-gravity technology someone will soon discover?

      Yeah, more debt, lower wages, increased immigration, higher stock and real estate prices, more wealth and income concentration to the top 0.1-1%, and “Drill, baby, drill!!!!” more $100 oil will permit us to ignore the oversized pachyderm crushing our collective frontal lobes into Aussie Vegemite.

      And we’re alledged to have a self-selected dominion over the elephants and fellow non-human creatures; however, we’re more like the caretakers of the elephants’ graveyard, and the absentee variety of our own in good time.

    • December 2, 2013 at 8:15 am

      US birth rates have been at or below what they say is replacement level – 2.1% – since 1976 except for one year. So the US population growth – which is expected to be 100 million in the next 60 years – is coming from immigration. Just as an experiment, go around town and tell people you are for stopping immigration. People don’t just react negatively to that suggestion, they get mad. Immigration is a sacred cow and no one pushes it more than the elite and the CEOs. They make money off of population growth and hundreds of applicants per job makes them drool like Mr. Burns.

  7. charlie
    December 2, 2013 at 1:31 am

    population, and many variants, pop growth, location, concentration are the ultimate economic problem for the earth. We need some sort of mystical superhuman intervention, soon, if there is to be a ‘humane’ solution. It looks increasingly like the lemmings and the cliff solution has the highest probability. Economists could only help by shouting the truth in unison … economics as currently taught is inadmissible. What is the solution?

    • BFWR
      December 2, 2013 at 3:20 am

      Transcendence of the economic and money systems with a policy of a universal dividend and a general retail discount to consumers which is rebated back to participating merchants.

      Followed by an intense but ethically sensitive and so morally tolerant acculturation process that stressed the cardinal virtues of prudence, that is balance. Systemic transformation first! Self determined restraint and perhaps reform….later!

    • December 2, 2013 at 10:22 am

      Unfortunately population growth is the one thing – probably the only thing – that all parties, religions, young and old, etc. agree on – it is either not a problem or not a problem to be dealt with. The elite in America not only don’t see American population growth as a problem – they see it as part of the solution.

  8. December 2, 2013 at 10:20 am

    “If this occupational shift story explained trends in wages we should expect to see sharply rising wages for retail clerks, custodians and other workers employed in low-paying occupations.”

    I don’t follow this statement. We have an ever rising population. Over a hundred thousand workers into the labor market each month legally and tens of thousands illegally. Why should we be seeing a rise in wages for workers?

  9. A.J. Sutter
    December 2, 2013 at 1:41 pm

    If I understand this paper correctly, it seems to be targeted to refuting certain academic explanations of wage inequality – not to refuting the notion that automation may be contributing to the increase in unemployment (in the broad, colloquial sense that includes both those people who are looking for work and those who have given up doing so). For one thing, unemployed people aren’t participants in the occupational wage distribution. Would it be premature to exculpate the robots for for contributing to unemployment, or have I misunderstood the subject matter of the paper?

  10. December 2, 2013 at 7:49 pm

    This is reassuring to elites as a result of it implies that difference are some things that happened, not one thing they did. They won out as a result of that they had the abilities and intelligence to reach a dynamic economy, whereas the large mass of employees that square measure falling behind didn’t. during this story the most effective we are able to do for the left behinds is fellow feeling and education.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.