This article appears in the July 21, 2014 issue of National Review.
In the aftermath of the Great Recession, observers have searched for a cause for the persistent limping pace of economic growth. Demand-side theories have formed the basis of the common assertion that the lingering effects of the financial crisis are to blame for our slow growth, but a striking new paper argues that another culprit may be to blame: the computer.
It’s not that computers are bad, of course, but that the impact of the computer on growth appears to have been much less impressive than expected. Economists thought the computer’s impact on the economy would have three main stages. For illustrative purposes, imagine a machinetool shop. In the pre-computer stage, a customer might ask the shop to make a part for an engine. The shop owner would dig through manuals, find the specifications of the part, and then use his machine tools to craft it. When the computer arrived, the manuals could be organized on a computer—the first stage. Instead of spending time digging through manuals, the owner could pull the specifications up right away, significantly increasing the number of parts he could craft in a day—the second stage. And the final stage has been reached today: The computer not only knows the specifications, but runs the machines as well.
Economists expected the growth effects from this multistage process to be great, and enduring. But the new paper, by John Fernald and the San Francisco Fed, suggests a dramatically different picture. The economy received a huge jolt from computers between 1995 and 2003, owing to large improvements in technology that accompanied the widespread introduction of computers—but the jolt tapered off suddenly, so much so that a decline in the growth rate of productivity is visible even before the financial crisis.
Growth can come to an economy if its workers become more talented (“labor quality”), if it accumulates capital that can then be used to increase production by its workers (“capital deepening”), or if it experiences technological progress that makes it capable of using its capital and labor more efficiently (which appears as growth in “total-factor productivity”). The nearby chart shows Fernald’s estimates of how these three factors have contributed to growth in labor productivity in the United States.
It is apparent in the chart that there have been two periods of extremely heightened technological progress in the U.S.—between 1948 and 1973, and between 1995 and 2003. This most recent boost was largely due to a boom in information technologies, in both the industries that produce them and those that use them. The slowdown of productivity improvements in these sectors is a large force behind the decrease in productivity growth in the economy that preceded the Great Recession. Fernald documents that the industries that sped up the most slowed down the most, and that there was no geographic variation in the slowdown that might have been related to the real-estate crash or other cyclical factors. Thus, technology must be the culprit.
With the productivity boom behind us, truly healthy growth will have to come from policies that drive capital and labor inputs higher. A corporate tax cut, for example, could well induce a machine shop to purchase a large number of new computer-driven machines. If it did, output would go up, even if tomorrow’s machines were no more productive than today’s.