Martin Feldstein
George F. Baker Professor of Economics, Harvard University, and President Emeritus, National Bureau for Economic Research
The conventional wisdom is that labor productivity (that is, nonfarm business output per hour worked by employees in the nonfarm business sector) has grown only very slowly over the long term and that the rate of productivity growth has recently experienced a surprisingly sharp decline. Both aspects need to be reconsidered. The official measure of the growth rate of productivity over the past sixty years (ending in the second quarter of 2017) has been just 2.0 percent. I believe this is an underestimate and may be a substantial underestimate. As I have explained elsewhere (Journal of Economic Perspectives, Spring 2017), the growth rate of real output is underestimated because the official statistics do not accurately reflect the contribution to real output of quality improvements and of new products. The underestimate of real output growth translates directly into an underestimate of the growth of productivity. While it is impossible to know by how much the true growth rate of real output has been underestimated, I believe it could be as much as 2 percent a year. If that is true, it would double the long-term growth of productivity. The official productivity growth rate has declined to just 1.2 percent in the most recent decade (to the second quarter of 2017) from 2.8 percent in the previous decade. But this just brings the average productivity growth over the past twenty years back to the exact 2.0 percent average of the previous forty years. The anomaly worth studying may therefore be the sharp rise in productivity growth from 1997 to 2007. Researchers who are concerned about the low productivity growth in the past decade should instead be asking why the decade from 1997 to 2007 was an outlier with the productivity growth rate rising sharply from the rate during the previous forty years.
Robert J. Gordon
Stanley G. Harris Professor in the Social Sciences, Northwestern University
Many recent discussions of the U.S. productivity growth slowdown focus on the sharp slowdown from the rapid 2.5 percent annual pace briefly achieved between 1996 and 2004 to the 1.1 percent that the economy has registered since 2004. I prefer to take a longer view and emphasize the two stages of the slowdown, the first step downward occurring between the 2.8 percent registered over the five epochal decades from 1920 to 1970 to 1.8 percent from 1970 to 2004, followed by the second step downward to 1.1 percent since 2004. And things are getting worse, not better, as the rate between mid-2010 and mid-2017 was only 0.6 percent (all these numbers refer to the total economy, not the private economy). The United States is not alone in experiencing this slowdown, as Western Europe and Japan have experienced even more severe decelerations from productivity growth rates well above that of the United States prior to 1996 to an even slower pace than the United States over the past decade. This common experience of slowdown suggests a common explanation, that innovation today— while many-faceted and ongoing—does not have as great an impact on productivity growth as earlier generations of inventions. The devastating loss of power and shortages of fuel after the recent hurricane in Puerto Rico stand as reminders of how fundamental were the invention of electricity and the internal combustion engine to the functioning of the modern economy. The post-1970 digital revolution made possible by personal computers turned out to have less impact on productivity growth than earlier inventions, and the transition those computers achieved in business methods from typewriters, paper files, and mechanical calculators to our current world of wordprocessing and spreadsheet software, broadband, web access, and search engines, was largely completed by 2006. Smartphones are ubiquitous but are mainly used by consumers for social networks and photography rather than as tools to boost business productivity growth. Some critics claim that productivity growth is currently understated because of mismeasurement, due to the upward bias in price indexes that fails to take account of the benefits of quality change and new products. But this has always been true, and so is not a source of slower productivity growth now than earlier in the twenty-first century, for which a case can be made that the benefits of the transition from horses to motor vehicles, or from the scrub board to the automatic washing machine, or of the invention of elevators, subways, and commercial air travel, caused even more measurement bias than the arrival of smartphones in the past decade. Similarly, unmeasured consumer benefits from medical advances must be assessed in comparison with the conquest of infectious diseases and infant mortality that allowed life expectancy to grow twice as fast in the first half of the twentieth century as in the last half. Surely productivity growth will revive soon from the unprecedented 0.6 percent growth rate of the last seven years. As the unemployment rate inches down toward 4.0 percent, labor shortages are emerging that will revive investment in labor-saving devices. But the productivity growth heyday of the mid-twentieth century, with growth rates of close to 3 percent for five straight decades, is nowhere in sight.
Jeffrey A. Frankel
Harpel Professor of Capital Formation and Growth, Harvard University’s Kennedy School
There are a variety of explanations for the declining growth rates in productivity and GDP that have been observed in recent years. The most prominent explanations involve technology. On the one hand, Robert Gordon (2016) has argued persuasively that we should not expect information and communications technology and other technological innovations of recent years to have as big an economic payoff as electricity, the automobile, and other technological revolutions of the past. On the other hand, Martin Feldstein (2017) has argued persuasively that productivity growth is higher than we realize because government statistics “grossly understate the value of improvements in the quality of existing goods and services” and “don’t even try to measure the full contribution” of new goods and services, and that these measurement errors are probably becoming more important over time. Not much attention has been given to another possibility: while information and communications technology and other technological developments bring many heralded benefits, they have some less-heralded negative sideeffects that may contribute to the slowdown in productivity and growth. At the risk of being thought a Luddite, I offer a partial list. n The advantages of each new incarnation of computer software or hardware are partially offset by the hours that everyone has to spend learning how to use the latest iteration. n Employees spend part of each workday on nonwork emails, social media, internet videos, and videogames. Even work-related emails can interfere with productivity because of excessive interruption of concentration. n Addictive videogames may undermine job skills and hours worked for some of the young. A recent study by Mark Aguiar and co-authors finds recreational computer activities partly explain a decline in labor supply by men ages twenty-one to thirty. n I will try to forebear expanding into things that merely undermine quality of life without showing up in the productivity statistics. (Have you stopped answering your phone due to the proliferation of robocalls? And how about the dangers of texting while driving?) But spam, viruses, and security breaches impose big costs on businesses as well as on households. n Those are just negative side effects of information technology. A list of other technological innovations with obvious downsides would include opiates, advanced weaponry, and more. To be clear, I am not suggesting that the net effects of recent technological advances are negative. But some innovations have negative side effects, including for productivity, and they seem often to be ignored.