If Automation is Destroying So Many Jobs, How Come Labor Productivity Sucks?

productivity

by James A. Bacon

What accounts for persistently sluggish United States employment growth in the sixth year of an economic expansion? Some blame it on the Obama administration’s economic policies, others on broad economic trends such as robotics, artificial intelligence and the automation of jobs. The latter explanation has broad intuitive appeal because we can all see what ATM machines have done to bank teller jobs, and it doesn’t take much imagination to foresee the impact of, say, self-driving vehicles on the truck-driving professions a decade from now. Technology is obliterating entire occupations. While new jobs such as the manufacture of ATM machines or the writing of driverless vehicle software may arise from the new technology, according to this line of thinking, fewer jobs will be gained than lost.

People have made the same argument since Ned Ludd went around England smashing textile machines that were putting spinners and weavers out of work. Yet somehow, the most dire of forecasts never panned out. That’s true, acknowledge advocates of the automation hypothesis, but things are different this time. Technological change is accelerating and the economy is displacing workers faster than the economy can create new jobs.

That hypothesis is certainly one that I take seriously. But I’m not yet persuaded. Here’s why. If accelerating automation is behind the decimation of jobs, we would expect to see two things: (1) an increase in the number of jobs destroyed (as opposed to a failure to create new jobs), and (2) an increase in productivity, reflecting the fact that it takes fewer people (and more robots/AI) to produce a given unit of economic output.

The first matter is difficult to answer given the spread of part-time and contingent employment. We may be creating more jobs, but if they are part-time jobs or contract gigs, that’s not the same as full-time jobs. It’s all very hard to measure, so we can’t draw firm conclusions. But we can address the second issue. Productivity, rather than shooting through the roof, as one might expect if technology were automating millions of jobs, is lagging severely. From today’s Wall Street Journal:

U.S. worker productivity fell in the opening months of 2015, extending a poor track record since the recession and underscoring longer-term risks to American workers’ wages.

The productivity of nonfarm workers, measured as the output of goods and services per hour worked, decreased at a 1.9% seasonally adjusted annual rate in the first quarter from the previous period, the Labor Department said Wednesday.

That marked the second consecutive quarter productivity has declined, something that has happened only three times in the past quarter century.

Look at the chart above. Productivity performance in this economic recovery lags that of the three previous business cycles. If automation is destroying millions of jobs and increasing economic output per unit of labor, it’s not being reflected in the productivity numbers. It’s all but impossible to make the case that automation is accelerating.

Here’s another way to look at the issue. If robot/AI automation is decimating more jobs than in the past, one would expect to see the effects in other countries such as Japan and the Euro zone where job growth has been even more listless than in the United States. Surely productivity must be soaring in those regions of the world. But it’s not. Also from today’s Wall Street Journal:

productivity2
Productivity growth in the United States, as anemic as it is, exceeds that of Japan, the United Kingdom and the Eurozone. At least productivity has risen measurably here in the U.S. since the Great Recession; it has flat-lined in the other advanced economies. Unless you want to argue that U.S. businesses are slower to embrace new technology than the businesses of other countries are — and maybe there’s a case for that, but I haven’t seen it — then it would appear that factors other than accelerating automation would best explain the great productivity stagnation both here, in Japan and in Europe.

What might those other factors be? My hypothesis: The rise of the rent-seeking state in which the political class allocates an increasing share of society’s resources to entitlements, the protection of special interests and the commandeering of resources for environmental goals. That rent-seeking process has increased in the United States — I warned about it in “Boomergeddon” — but is even more advanced in Japan, the U.K. and the Eurozone. If we have an employment problems, our policies, not technology, is primarily to blame.