The (price) illusion of declining investment
At the Jackson Hole economic policy conference late last month, David Autor, an economist at the Massachusetts Institute of Technology, presented a paper about job polarization and the role of technology in the creation of jobs in which he argues that technology has not destroyed jobs on net over the past several decades. In the paper, he presents a graph showing that investment in information technology as a share of gross domestic product stagnated in recent years—evidence that technology and capital have not taken jobs from workers.
A deeper dive in the data, however, shows that certain forms of investment may not be stagnating and that the substitution of capital for labor may indeed be a current concern.
Justin Fox at the Harvard Business Review took a look at Autor’s graph about information technology investment and made an important point—the graph used nominal, or non-inflation-adjusted data. That might not seem like a major concern, but using nominal data for measuring investment as a share of GDP doesn’t account for the declining price of IT investment goods over time. To be clear, we’re talking about physical investment, not financial investment. Think of a company buying computers for their workers, not a venture capital firm investing in a new start-up.
As Fox points out, IT investments have simultaneously become cheaper and better due to technological advances. A personal computer today is both cheaper and a better product than a personal computer from 1984. The Bureau of Economic Analysis, the producer of the GDP data, account for these facts when they create inflation-adjusted figures. Fox made a new version of Autor’s IT investment graph that shows inflation-adjusted IT investment as a share of GDP has increased over the years. The nominal amount declined, but given that the price of these investments also dropped considerably the value of these investments increased.
A similar trend is present in the overall investment data. As part of an earlier conversation about trends in investment, Cardiff Garcia at FT Alphaville highlighted research by economists at Citi Research. That research shows that after accounting for the changing prices of investments and depreciation, the share of the economy going to investment hasn’t declined.
These observations have two important implications. The first is that concerns about private underinvestment in the U.S. economy might be overblown. Accounting for price differences reveals that investments have been relatively constant. Our economy might be in need of even more investment, but companies appear not to be pulling back on it.
The second implication goes back to Autor’s point about the substitution of capital for labor. The nominal trends might not be indicative of increasing substitution, but looking at the trends in prices we can see that investment goods are becoming cheaper and cheaper compared to GDP and consumption goods.
Research by economists Loukas Karabarbounis and Brent Neiman, both of the University of Chicago, show that the declining price of investment goods is an important factor for why the share of income going to labor has been on the decline recently. Capital may not be poised to take a large swath of jobs away from workers now, but the evidence shows that it is already reaping a larger reward. A future where capital increasingly substitutes for labor is a very real threat.