Will US productivity stagnate or flourish? Maybe neither.
Stagnationists tend to underestimate the growth potential of information technology, while the techno-utopians overestimate it.
To increase living standards, we must increase productivity, which is a simple measure of economic output per unit of input. Unfortunately, US productivity performance has been utterly dismal.
The Bureau of Labor Statistics recently reported that the country’s productivity declined at an annualized rate of 1 percent in the first quarter of the year, further prolonging an eight-year rut that already counts as the worst we have experienced since the government started measuring productivity in 1947. Our productivity growth in this period has averaged little more than 1 percent per year. No wonder wages have hardly budged.
The need for stronger productivity growth is only going to become more urgent in the years ahead as baby boomers retire and the share of the population above 65 increases from 15 percent this year to 22 percent in 2040. Without strong productivity growth, either those retirees will have to accept smaller entitlements, or workers will have to pay higher taxes to preserve the status quo. Conversely, every tenth of a percentage point increase in productivity would add $50 billion to the federal budget.
So the central economic question of our time is whether productivity will continue to lag or will rebound. Yet as with so many important questions, economists can’t agree on what the answer is likely to be.
Much like the story of the blind men and the elephant, economists ponder the same evidence but come to different conclusions. On one side, there is a stagnationist camp, led by such luminaries as Robert Gordon and Tyler Cowen, who assert that the US economy has harvested all the low-hanging fruit there is to be found for the purpose of increasing productivity, so we are in for a long period of doldrums. On the other side is a techno-utopian camp, which includes MIT’s Erik Brynjolfsson, among others, who argue we are on the brink of the next industrial revolution, which will be driven by amazing new technologies that will power a productivity boom.
Both camps are likely wrong, but for different reasons.
Stagnationists make a number of errors, most importantly dismissing today’s innovations in information and communications technology as inconsequential because they don’t change the material world the way prior eras of innovation did. It is certainly true that many of today’s most important innovations are not based on changes in atoms (such as physical advances like steel bridges and cars), but in bits (such as digital tools and systems), but the stagnationists underestimate the extent to which these new types of innovations allow economic activities to be conducted much more productively.
In fact, it’s very likely that, 50 years hence, America will look pretty much like it does today in physical terms. Unlike the Jetsons’ world, cars will still drive on the streets, buildings will appear much as they do now, and we will be wearing clothes made of fabric. But the cars will be self-driving, buildings smart, and clothing connected to the Internet.
Moreover, there have been stagnationists throughout history, and they have always been ready to proclaim the end of innovation. For example, in 1899, at the end of an awful economic decade, Charles H. Duell, commissioner of the US Office of Patents, is said to have declared, “Everything that can be invented has been invented.” Just as such pessimists have been wrong in the past, they are likely to be wrong now. As noted economist Joseph Schumpeter stated, “There is no reason to expect slackening of the rate of output through exhaustion of technological possibilities.”
If the stagnationists are too pessimistic, the techno-utopians are too optimistic, breathlessly proclaiming the emergence of a new industrial revolution. Klaus Schwab, head of the World Economic Forum, reflects this view when he writes that we are in the midst of a “Fourth Industrial Revolution” in which “major technological innovations are on the brink of fueling momentous change.”
For their part, techno-utopians make two key mistakes. First, they assume that the current pace of IT-enabled change will continue or even accelerate. But this overlooks the fact that the rate of Moore’s Law – named after Intel co-founder Gordon Moore, who famously predicted the speed of computer processing would double every 18 to 24 months – has actually slowed by half in the last 12 years compared to the prior three decades. Meanwhile, transistor costs have been increasing for the first time. And the pace of IT advancement could slow even more as silicon-based IT systems are likely nearing their physical limits. Even Gordon Moore has said Moore’s Law is petering out.
Second, if the stagnationists think everything is about atoms, techno-utopians assume everything is about “bits.” Notwithstanding the tech entrepreneur Marc Andreessen’s insightful comment that “software is eating the world,” so far it’s only been the appetizer and not the main course. Despite IT advances that boost productivity in an array of information-based functions (such as check processing and information retrieval), a large part of the economy still involves people interacting (think of nursing homes, police, firefighters, and psychologists) or performing physical tasks that are difficult to automate (such as building construction, janitorial services, or haircuts) and there is no sign of Moore’s Law emerging in these physical-world industries.
So if future US productivity growth is unlikely to be either too hot or too cold, will it be just right? (Just right in this case would be somewhere around 2.4 percent per year, which was the growth rate we achieved from 1995 to 2008.)
Getting to “just right” anytime soon might be a stretch. Some of the technologies touted as powering the next wave of productivity growth, such as cloud computing and advanced broadband, are evolutionary extensions of existing technology rather than radical breakthroughs. Others, like robotics, artificial intelligence, nanotechnology, and biotechnology, which constitute truly powerful and transformative technologies, are likely at least two decades away from where the PC and Internet were in 1995 – that is, reasonably powerful, easy to use, and affordable. When these technologies do come into wide use, we might reasonably expect, based on the emergence of past transformative technologies, a bump up in productivity to perhaps 3 or 3.5 percent a year, if all goes well. Nice, but hardly the kind of revolutionary step forward the techno-optimists promise.
But this doesn’t mean the stagnationists are right to argue we are consigned to a decade or two of slow productivity. We are not powerless to do anything while we wait for scientists and engineers to make robots really effective, for example. Rather, the next president needs to put forward national productivity strategy that is robust and coherent. The Information Technology and Innovation Foundation estimates that such a strategy, if effectively crafted and implemented, could boost US productivity by at least one percentage point a year. My next column will outline what such a strategy should look like.
Robert D. Atkinson (@RobAtkinsonITIF) is president of the Information Technology and Innovation Foundation.