“Let me put this in perspective. For the total economy, productivity growth was 2.7% from 1920 to 1970, 1.6% from 1970 to 1994, 2.3% from 1994 to 2004 during what we call the dotcom era, and just 1.0% from 2004 to the second quarter of 2015.1 So the productivity growth of the last 11 years was not only slower than in the dotcom era, but even slower than in the so-called slowdown period beginning in the early 1970s.”
That’s from Robert Gordon, a professor of economics at Northwestern University and it comes courtesy of Goldman who has taken a close look at declining labor productivity in the US.
As the Vampire Squid notes, falling productivity is something of a paradox. That is, better technology and advances in efficiency should by all rights have increased productivity but apparently, a number of factors are intervening to short circuit the system. Here's Goldman's full interview with Gordon.
The reason for the slowdown after 1970 is straightforward: we simply exhausted the productivity benefits of prior innovations. In the late 19th century, hugely important “general purpose” technologies, like electricity and the internal combustion engine, were invented. Then there were major developments in entertainment and communication in the form of the telephone, telegraph, radio, motion pictures and television. We made major breakthroughs in health. And we vastly improved working conditions. All of that came together between 1920 and 1970. The last three spin-offs of the great inventions— interstate highways, commercial air travel, and air conditioning in most businesses—were also largely complete by 1970.So at that point we had run through the productivity payoffs.
Allison Nathan: Why has productivity growth stalled?
Robert Gordon: Let me put this in perspective. For the total economy, productivity growth was 2.7% from 1920 to 1970, 1.6% from 1970 to 1994, 2.3% from 1994 to 2004 during what we call the dotcom era, and just 1.0% from 2004 to the second quarter of 2015.1 So the productivity growth of the last 11 years was not only slower than in the dotcom era, but even slower than in the so-called slowdown period beginning in the early 1970s. The reason for the slowdown after 1970 is straightforward: we simply exhausted the productivity benefits of prior innovations. In the late 19th century, hugely important “general purpose” technologies, like electricity and the internal combustion engine, were invented. Then there were major developments in entertainment and communication in the form of the telephone, telegraph, radio, motion pictures and television. We made major breakthroughs in health. And we vastly improved working conditions. All of that came together between 1920 and 1970. The last three spin-offs of the great inventions— interstate highways, commercial air travel, and air conditioning in most businesses—were also largely complete by 1970. So at that point we had run through the productivity payoffs. We have also now run through the payoffs of the digital revolution that followed. Between 1980 and 2005 there was a total transformation of business practices from paper and filing cabinets to flat screens and search engines. But that transition is over. And the temporary revival of productivity during the dotcom era was uniquely concentrated in a very short span, with remarkably few gains in productivity growth since. We’re using software and computers now that are very similar to the ones we used ten years ago. So it is no surprise that productivity growth has been slower over this decade.
Allison Nathan: Are the productivity statistics simply failing to account for the impact of new technologies?
Robert Gordon: Many consumer benefits are clearly missing from the GDP statistics. But GDP has always suffered from this fault. For example, GDP completely failed to capture the transition from the horse to the motorcar and the enormous benefits that resulted from an environment free of horse manure droppings in the streets. If anything, I think a case could be made that what productivity statistics failed to capture 1 Note from GS Research: The figures cited here are for the overall economy; corresponding numbers for the US nonfarm business sector (the conventional measure) tend to run about 0.4 pp higher. in the first 50 years of the 20th century was larger and more important than what is missing now. At that time, we left out the benefits of conquering infant mortality; of going from the 60-hour work week to the 40-hour work week; of the new ability to travel with a car. In any case, what we’re seeing now is more of the same: a general failure to translate new inventions into GDP, and therefore into productivity measures.
Allison Nathan: Should we be measuring productivity differently?
Robert Gordon: I think it’s impossible to quantify the benefits of new inventions. Economists have done experimental work on specific inventions like tractors, and it is possible to come up with ballpark estimates. But quantifying those improvements has always been difficult. And the hypothetical measurement of the benefits of more recent inventions like smartphones and tablets is probably more difficult than most.
Allison Nathan: Could we be experiencing delays in seeing the effects of new technologies on productivity?
Robert Gordon: Yes, we could be seeing some of this dynamic. For example, the rollout of electronic medical records has been very slow even though we have had the necessary technology for a good 15 years. But the real delay happened in the early 2000s. Despite the sharp drop in the stock market and a tremendous collapse in high-tech investment from 2000 to 2003, productivity growth was very rapid throughout the whole decade from 1994 to 2004, reflecting the delay in learning how to make full use of the internet, which was first introduced in the early 1990s. My favorite example is the introduction of airport check-in kiosks, which took place between 2001 and 2005 using technology that had been invented a decade earlier.
Allison Nathan: You argue that recent technological developments don’t hold a candle to the breakthroughs of the past. Are the world’s best innovations truly behind us?
Robert Gordon: In my view, the inventions of the century from 1870 to 1970 utterly changed human life in a way that now is taken for granted. When you consider the immense progress in getting rid of disease, filth, manure; the advances in health with antibiotics and treatments for heart disease and cancer; the liberation of women from the chores of doing laundry with a scrub board; the transition away from steel workers working 12 hours a day, six days a week, there really is no comparison with the inventions taking place today. Smartphones and social networks are entertainment and not basic to human life. But “best” is subjective. Some people may think it is more important to have a social network than indoor plumbing.
Allison Nathan: Some would say that the productivity contributions of past inventions, particularly during the industrial revolution, did not properly account for environmental or other costs. What are your thoughts? Robert
Gordon: More than overstating productivity growth during the industrial revolution, I think we have understated the growth of productivity from 1970 to the turn of the 21st century when we had major improvements in air and water quality mandated by legislation. We have incorporated part of this clean-up into productivity statistics in a very subtle way by accounting for emissions control devices on auto engines. But most of the improvements in the environment are missing from GDP. That being said, the costs of current technology are probably lower than the costs of past industrialization, so these types of omissions are likely less prevalent today. Allison Nathan: Are there any areas of innovation that hold substantial promise in your view? Robert Gordon: Most of the excitement is centered on artificial intelligence and robots. Robots are nothing new. The first industrial robot was introduced by General Motors in 1961. Since then, robots have steadily replaced human labor in manufacturing, and they continue to create more rapid productivity growth in the manufacturing sector than in most of the service sector. Another place where robots are gradually appearing is warehousing. But they don’t fetch individual items and bring them to a station for packing; they simply pick up an entire tier of shelves and bring it to a person who selects the right item and manually packs it. Developments in robotics have so far been unable to duplicate the actions of the human hand, even for many tasks that human beings do intuitively. So the gradual arrival of robots in the economy is very slow. As far as artificial intelligence, computer technology has already steadily replaced human jobs. Think of the disappearing travel agent and reservation clerk, or, more recently, the legal associate. So there is a lot of excitement about technological change, but it is taking place at a very measured pace, especially to the extent that it is replacing human labor.
Allison Nathan: Will these innovations be sufficient to boost productivity?
Robert Gordon: Not meaningfully. I expect productivity growth over the next quarter-century of 1.2%, slightly above the 1.0% growth rate of the last 11 years but still below the 1.4% rate over the past 45 years if you take out the dotcom decade, which was an unusual period that I don’t think will be repeated. That difference of 0.2% is the contribution of slower innovation compared to history. Keep in mind that this slowdown already occurred in the last ten years. So I am basically predicting more of the same, not some new arrival of stagnation.
Allison Nathan: How important is the pace of productivity to your overall outlook for US economic growth?
Robert Gordon: It's absolutely central. By definition, growth in real GDP is equal to growth in productivity plus growth in hours of work. The growth in hours of work is limited by population growth and growth in the number of hours that each member of the population works. The latter is going to be shrinking over the next 25 years due to the retirement of the baby boomers. So while US population growth should be about 0.8% per year, we can only expect growth in hours of work of 0.4%, much lower than what we observed in the latter part of the 20th century. Adding that to the 1.2% I expect for productivity growth, my projection for growth in real GDP is 1.6% a year. This is just the same as the last 11 years, but it is only half of the 3.2% growth rate we experienced from 1970 to 2004. Allison Nathan: You seem skeptical of technological tailwinds and more focused on economic headwinds. Which headwinds concern you the most?
Robert Gordon: I see four main headwinds to economic growth. The first is rising inequality. Our winner-take-all society provides very high payoffs to the top rock stars, CEOs, lawyers, and so forth. And at the bottom, we have machines gradually but steadily replacing workers, and an erosion of manufacturing jobs from globalization and trade. So the gap between the very top and the mass of people in the middle and the bottom continues to widen inexorably. The second headwind is the end of the great expansion of education that brought Americans from completing only an elementary school education in 1900 to a great majority having a high school education by around 1970. There has been a gradual increase in the share of young people going to college, but the United States has fallen from its previous position of leadership in global education and now ranks about 16th among nations in the percentage of its young people completing a four-year college degree program. The third headwind is the demographic shift I mentioned of baby boom retirement pushing down overall hours worked. And the final headwind, also related to aging, involves federal government expenditures on Social Security and Medicare increasing faster than the shrinking workforce’s ability to provide the tax revenue to finance these benefits. This will eventually necessitate tax increases and/or benefit reductions, which will cause people’s after-tax disposable income to grow even more slowly than their pre-tax income. Allison Nathan: Does your outlook owe more to a measured pace of innovation or to these headwinds?
Robert Gordon: Quantitatively, the headwinds are more important. That said, there is a whole list of policies that would help address them, from a more progressive tax system and increased spending on pre-school education to massive immigration reform. And many of those proposals also deal with productivity by raising the quality of human capital.
Allison Nathan: You are often described as a “technopessimist.” Is that a fair characterization?
Robert Gordon: I would certainly classify myself as a technopessimist. But, if you think about it, the terms techno-optimist and techno-pessimist belie the meaning of the words optimism and pessimism. Techno-“optimists” are predicting a future of massive technological unemployment with a quarter or half of the labor force unable to find jobs. Under the hood of their optimism, they are deeply pessimistic about the future of work. I think that technological change is proceeding slowly, just as it has over the past decade, which should allow us to keep our unemployment relatively low. So under the hood of my technopessimism, I'm very optimistic about the future of work. Where I see the real problem is not in finding a job for everybody, but in finding good jobs for people, and in dealing with the inevitable rise of inequality.