Every month, The Big Question video series brings together a panel of faculty experts for an in-depth discussion. In this edited excerpt from January’s episode, Chicago Booth’s Steven J. Davis, William H. Abbott Professor of International Business and Economics, and Chad Syverson, J. Baum Harris Professor of Economics, are joined by Robert Gordon and Joel Mokyr of Northwestern University in addressing the influence of innovation on the US economy. The discussion was hosted by Hal Weitzman, Booth’s executive director for intellectual capital.
What headwinds is the US economy facing in the coming decades?
Gordon: When we talk about economic growth, we’re talking about growth in the standard of living—growth in output per person. That can be divided into growth in output per hour and growth in hours worked per person. Most of our difficulties come in the latter. Hours per person rose in the 1960s–1980s as women entered the labor force. That has now turned around, partly because of the retirement of the baby-boom generation. That alone takes a full percentage point off of our growth compared to the 1980s and 1990s.
But you’re forecasting that US growth could fall to as low as 0.2 percent.
Gordon: Most of it has already happened. Growth in per capita income in the decade since 2004 is only 0.6 percent, after having gone to 2.1 percent from 1890 to 2007. We have slowing population growth. We have slowing growth in the labor force as baby boomers retire and young people don’t work as much. We have an educational system that has problems from bottom to top. The top 1 percent get a far greater share of the small growth we have than the bottom 99 percent do. And, gradually, we’re building up a lot of debt that we’re going to somehow have to pay off. Those are the headwinds.
Are they inevitable?
Gordon: We can do a lot about some of them and nothing about others. Eventually, we have to pay off the debt, and that involves rearranging the way we spend money on transfer payments and changing our tax system. If our population growth is slowing, we can increase it by welcoming immigrants. We can turn the dial on our own population growth because so many people still want to come to the United States. We can do a lot about education, particularly preschool education. We can change our system of financing higher education so we don’t wind up with $1.5 trillion in debt.
Syverson: Bob talked about the slow growth since 2004. Part of that is cyclical. We have seen drops in hours worked per person. A little bit of that has been retirement of the baby-boom generation, but a lot has been people dropping out of the labor force because the opportunities haven’t been good enough for them to work. Some of that will be reversed. The growth that we saw in hours per person in the decades prior to that was not going to be sustained. Women coming into the labor force: that wasn’t going to continue paying dividends forever. Population growth more broadly is something that’s going to be very slow to react and hard to turn around, although immigration policy can be affected very quickly, if we want it to.
Mokyr: We should also talk about tailwinds. From 1870 to 2000, all industrialized economies were lifted by this fantastic tailwind of technological change. My attitude is, “We ain’t seen nothing yet.” I can’t exactly say where it’s going. But there are a number of factors that make me optimistic. First, science is driven by technology and vice versa. Technology provides science with tools. New technology feeds back into our scientific understanding of nature and eventually that will allow technology to change through an indirect feedback loop. Secondly, we have a system in which incentives to innovate are better than ever. We have tenure, Nobel prizes, and presidential awards to reward scientists. Society has been extremely good in setting an agenda for where knowledge is going to go.
Davis: As Joel said, it’s hard to predict the pace of technological innovation. But there are things we can do that affect the conditions under which technological advance is more or less likely. One is our stance toward the immigration of highly skilled, highly ambition people. The US makes it very hard for ambitious and talented people to come in and push the frontier here. On the education front, there is both good news and bad news. The US has a mediocre school system compared to much of the world. We have a lot of room for improvement. That won’t generate the immediate returns that changing our immigration policy would. But over the long term, it’s vitally important that we do that if we want to reverse some of the tailwinds.
Mokyr: I would add to that federal funding for basic research. That money has been traditionally extremely well spent. Much of the innovation since World War II has been driven by federal funding. We are cutting off our oxygen supply by not spending that money.
Will technology unleash big, paradigm-shifting innovation?
Gordon: A part of the economy where we’ve seen a lot of innovation is in manufacturing and wholesaling, in the form of robots taking over for people. This is nothing new. The first industrial robot was introduced by General Motors in 1961, so this is not a quantum leap into some unknown future. This is a continuation. Robots are becoming cheaper, smaller, and more effective, and will continue to produce above-average productivity growth in manufacturing. I like to walk around every day, and I look at people and say, “Is their job going to be eliminated within the next 30 years?” And I find it very hard to locate anyone whose job is going to be replaced. Here we are, 50 years after the introduction of the robot, and in our grocery stores the goods are still put on the shelves by human beings. The great majority of the wholesale sector is trucks delivering to stores where the truck driver also pushes the stuff through the store and arranges it on the shelves. In offices, we had a revolution from 1970 to 2000. We went from paper-driven typewriters to everybody connected to the web. By 2005, everybody had a flat screen. But what people are doing in offices today is the same as what they were doing 10 years ago.
What will that mean for productivity?
Gordon: There’ll be a slow, incremental improvement in productivity just as there has been over the last 10 years. We’ve got a precedent of the last 40 or 50 years of relatively slow productivity growth, and we’re going to see that continue. We’re not going to see some great miracle of inventions.
Syverson: Predicting the future is very difficult. It’s fun, but is there any indication that we’re good at it? Not so much. Historically, we know with the last diffusion of general purpose technologies—the electric motor and the internal combustion engine around the end of the 19th century, into the early 20th century—it’s clear in the data that the productivity gains that those yielded came in waves. Some came over the course of a decade, and then productivity growth slowed again for another 10–15 years, and then it picked up yet again after that. I’m not sure anyone at the time could have predicted either the first slowdown or the second pickup. But we do know that this stuff does not have to just only come once. Technologies that are already having effects or have had effects, like information technologies: it could be that there’s a second wave of those productivity gains coming.
Mokyr: The view of technological change as machines coming in and replacing people is an overly narrow view. [Technology] really transforms not only how many workers are needed for a job, but also exactly what these workers do in the environment in which they live. In the past 15 years we’ve seen the rise of telecommuting. It’s growing, and it will continue to grow. That doesn’t mean fewer workers; but it will mean workers may not have to spend an hour driving to work, sit in a cubicle for eight hours, then spend another hour driving home. Instead, the work may come in twice a week for a few hours and much of it may get done from a living room. Is that a gain in productivity? No, because we don’t count commuting as work. Is it an improvement in economic welfare? Of course. In that sense, the impact of technology is much larger than is measured by standard productivity indicators.
Davis: There are other ways in which technological innovation does not show up very well in the measured productivity statistics. For example, extensions of life or improvements in the quality of life may affect productivity while working, but also increase the value and the amount of leisure time over the course of a lifetime tremendously. Another area is the value of entertainment services. If you go to a poor country, you’ll see somebody spend 5 or 10 percent of their annual income on a color TV. You have to ask, how much surplus is the middle-income person in the US, with their 150 cable channels, getting from their TV? It’s enormous. In principle, you can try to capture that in our standard productivity statistics. In practice, it’s extremely difficult to do.
Gordon: What are going to be these great benefits in the future—in the next 20 years compared to the last 20? Talk about telecommuting. All the tools we needed for telecommuting were already in place 20 years ago. So why are we still waiting here for everybody to telecommute? Because there is socialization at the office. People are afraid of being left out of the loop, left out of the power structure. There are fundamental reasons why this isn’t happening. A lot of these dreams about the future are just that, dreams. I’m interested in specifics.
But what about the time-lag argument? When you’re in the middle of the process, how can you say how the process will play out?
Gordon: With all respect to Chad, there was one big wave. The two great inventions, electricity and the internal combustion engine, happened within 10 weeks of each other in 1879. It took a good two to three decades to figure out how to use them. Paul David, the economic historian, has made this point with an analogy between the electric motor and the computer. Both of those took about three decades to come to fruition, and then the productivity growth took off. The same thing is true of the internal combustion engine. Here, we had horse-drawn carriages, and then we had this powerful engine. It took about two decades to figure out how to get the power from the engine to the wheels without tearing apart the vehicle. But once that happened, it took another three decades for the number of motor vehicles in the economy to take off to a high enough level to make a dent in productivity. If we look at the productivity statistics, there’s one big wave. Productivity growth, total factor productivity, took off in 1920. The growth rate between 1920 and 1970 was three times as high as it’s been since then. So our great years of productivity growth are in the past.
Syverson: If you look at the labor productivity growth, you see exactly the same pattern repeated with the IT revolution. There’s a 25-year lag or so between invention of the technology and the first measurable productivity growth from it. That occurs over a period of about a decade, and then there is a slowdown again. That happened with the electric motor and the internal combustion engine. If you want to extend the window, put another 20 years after that, and call that one big wave, that’s fine. But let’s give ourselves 20 more years after today before we declare the end of the IT revolution.
Mokyr: There was a 20- or 30-year lag for the electric engine. For the steam engine, it was more like a century. So why not give us 20 or 25 years for telecommuting to take place? I actually think we are going to look at the growth of that very quickly. We may not see the disappearance of the office. But people don’t have to go every day, and they don’t have to show up all at the same time. Think of the impact that that will have on, for instance, mothers trying to raise children, when a child is sick and she can say, “I’m going to work from home today.” That kind of flexibility didn’t exist in the factories. It’s a huge boon to human welfare and to the way work is organized.
Davis: It’s fun to talk about these prognostications, but it’s also fundamentally hard. There are other things we talked about that we know do have measurable effects on productivity growth, like improvements in our schooling system, like bringing in more highly talented immigrants. Those are things that are under our control. If we reform those institutions and those policies, there will be productivity gains. In terms of schooling, the US is like some advanced middle-income economy looking to the US for technological leadership and saying, “If we can just get to the frontier, we’ll grow a lot.”