The Productivity Slowdown- Exploring A Role For Demand
Almost a decade on from the financial crisis, U.S. economic growth continues to disappoint. While the labor market has recovered to something like full employment, growth in output has remained weak. The main culprit appears to be anemic growth in productivity, which is the technological efficiency with which we employ capital and labor. In the decade up to 2007, this “total factor” productivity improved at an average rate of 1.59% a year. Since the crisis however, it has slowed to a paltry 0.58%.
A seemingly small difference like this has enormous practical consequences. If productivity had continued on its pre-crisis trend, the economy would be around $1.7 trillion dollars larger today, or roughly $5000 annually for every person in America. Naturally, this has not gone unnoticed by economists, who have proposed several explanations. I’ll outline two main arguments here, which fall short on key dimensions. In doing so, I’ll highlight the need for more research on a third potentially important factor- that the weak economy itself might have contributed to a decrease in the rate of technological progress.
The first strand of thought questions whether there has been a slowdown at all. Perhaps we are simply not measuring productivity well enough in the age of digital technology. After all, when a person binge-watches Netflix all weekend, the contribution to measured GDP is the monthly $9.99 subscription fee. Previously, two videos from a video store might have cost a comparable amount, but were a universe away in terms of the variety of content available and the convenience of accessing it. The same holds for music streaming services and communication, and goes doubly so for free products like Facebook.
Entry of new, higher-quality products like this is a notoriously difficult issue- in most cases, such quality improvements are simply ignored in the national accounts. It seems undeniable that by doing so we understate the growth in real incomes, with several studies concluding that the magnitude of the bias is large[1]. However, it is also true this has been an issue for quite some time, being a central finding of the Boskin Commission on the CPI in 1995. What is unclear is whether this bias has increased with the advent of the tech economy, and how it explains the slowdown in sectors outside tech (areas as diverse as manufacturing, wholesale trade and transportation). Moreover, why did it become such a problem in 06-07?
Another explanation for the productivity slowdown concerns the supply-side of the economy. Perhaps the extraordinary growth experienced in the late 1990s and early 2000’s was a one-off event, associated with a boom in information technology which has now run its course. This is the view of several prominent economists, including the economic historian Robert Gordon and San Francisco Fed economist John Fernald. As evidence, they point to the fact that the slowdown actually predates the crisis, beginning in 2006, or even earlier on some measures, and was particularly acute in IT-intensive sectors. Our technology gifts exhausted, we have now returned to the “norm” of the low-growth period that held sway from 1973 to the mid-nineties.
However, this cannot be the whole explanation, for several reasons. First, it cannot explain the equally poor productivity performance of European economies, which generally lag behind the U.S. in adoption of ICT technologies. One would think there should haven been technological gains left to be exploited for these economies to move closer to the U.S., yet they suffered very similar downturns in productivity growth during the crisis and afterwards. Secondly, a truly supply-constrained economy should generate an increase in inflation with interest rates so low, especially now that the labor market has normalized. This is what we observed when the oil shocks of the 1970’s lead to a period of constrained supply, weak productivity[2] and sustained price growth. To date, however, inflation remains below the Fed’s target.
Lastly, it ignores the link between business investment and R&D and future growth, both of which were high at the eve of the crisis, but have since suffered severe contractions. At the risk of oversimplifying, productivity growth arises when firms invest in new technologies and products to reinvent their businesses and ways of working. They do this, by and large, in the expectation of a return on their investment. When that return is impacted by a weak economy, they invest less.
This should lead us to consider the role of weak aggregate demand in contributing to the productivity slowdown. Consumers impacted by the financial crisis spent several years paying down debt accumulated during the boom, leading to a period of weak consumption growth. In turn, this may have affected businesses’ willingness to invest. This idea ties in to the “secular stagnation” agenda advanced by Larry Summers. With interest rates near zero (and unable to go any lower), there may have been no equilibrating force that can get consumers and businesses spending again.
Research on demand side explanations for weak productivity are beginning to emerge. Diego Comin at Dartmouth and coauthors have emphasized that the speed of diffusion of new technologies fell significantly in the Great Recession, as firms stopped investing in upgrading their capabilities to the state of the art. My own research attempts to gather evidence from the experiences of businesses in different U.S. cities, some of which were harder hit by the housing downturn that others. However, much more research is needed before we have a clear view on how much the crisis contributed to our current predicament.
The implications for policy in disentangling this issue could not be clearer. If the slowdown truly reflects an exogenous wave of innovation that has passed (the technology shocks of old Real Business Cycle theory), then there is little we can or should do about it. We should simply adjust our expectations, and look forward to a diminished future. However, if the link between the depth of the crisis and the slowdown is borne out by future research, then the importance of taming the business cycle and preventing future crises is magnified. Not only would we worry that deep recessions cause irreparable harm on impact, they would also leave scars that could persist for a generation.
[1] One study by Yale’s William Nordhaus focused only on the amount of light households could buy throughout history. He found that the improvement from Babylonian lamps to compact fluorescent bulbs allows us to purchase around 350,000 times more light with a day’s work than our ancient forbears. A more comprehensive study by Broda and Weinstein on supermarket purchases found we overstate inflation (and therefore understate income growth) by around 0.8% a year. This compounds drastically over time.
[2] Nordhaus found in a separate study that up to two-thirds of the previous productivity slowdown came from industries impacted by the oil shocks, such as pipelines, motor vehicles and auto repair.