On average, over the last 145 years, the economy of the United States, measured by GDP per capita, has grown 2% per year. Over the last 45 years, however, US per capita growth has been markedly slower, averaging 1.62% for the period 1970-2014. In the wake of the 2008 financial crisis and the slow recovery from the subsequent recession, investment manager Mohammed El-Erian dubbed this low growth experience “the new normal”.
During the same post-2008 period, economist Robert J. Gordon has been on the conference circuit1 with the prediction that US economic growth is unlikely to return in the near future to that earlier 2% rate. Opposing Gordon (often on the same stage) have been techno-optimists such as economic historian Joel Mokyr, his Northwestern colleague, and the MIT business school pair of Erik Brynjolfsson and Andrew McAfee, all of whom see a much brighter future. Now, Gordon has released a massive and impressive book, The Rise and Fall of American Growth: The U.S. Standard of Living since the Civil War, designed to solidify and document his case and present it to the public.2
Basically, Gordon’s opponents have been saying that, just as innovation has been the main driver of past growth, so it will be the cure for our current problems; using “brilliant technologies” that are reaching or an “inflection point”, we can and will innovate ourselves out of the current slump. To only slightly misquote Mokyr, “The unimaginable is plausible.”
Gordon’s response has been twofold. First, he has pointed to six “headwinds” particular to this stage in US economic history, obstacles which any growth-enhancing innovations must overcome. Second, while he admits that recent (and future) technological innovations may be impressive, he insists they have not had the same impact as the those of the earlier, first age of US industrialization and economic expansion (roughly 1870-1970). Not only did the early-stage inventions like electrification have greater impact (so far) than the computer, but that impact was intensified and extended by the basically unrepeatable combination of the Great Depression and the Second World War.
Let’s hold off on the innovation question for a moment and look at what Gordon means by “headwinds”. You can think of them as trends or situations that create obstacles or limit possibilities. My version reads as follows:
- Steadily increasing inequality.
- Decreasing educational attainment.
- Changing demographics and reduced workforce participation.
- Reduced options for investment due to high personal and debt and the need to fund existing entitlements.
- Social breakdowns among the lower-earning groups.
- External constraints due to globalization and environmental stress.
Note that the headwinds do not have to be considered in isolation. Stanford economist Charles I. Jones, using the work of David Autor of MIT, has produced a graphic (Figure 1) that powerfully combines headwinds #1 and #2, and shows how lagging educational attainment is translating into increased inequality
Figure 1. Less Education = More Inequality
Source: Jones, “The Facts of Economic Growth”, 2015.
The thick blue line on Jones’ graph shows that the proportion of hours worked by college grads in the US is increasing, but the dotted blue line, which I added, underscores the fact this fraction is not increasing as fast as it did, peak-to-peak, between 1960 and 1980. And since 1980 college-educated workers have gone from earning 50% more on average than workers with no college to earning 100% more. So, fewer are catching up even as the consequences of falling behind have gotten bigger. One needs only to look at the current election cycle to confirm that this is no longer a silent crisis. The angry, populist, nativist, anti-intellectual, backcountry rebellion we see today has little to do with the Occupy! movements of a few years past that claimed to represent the “99%”. The new battle is within the 99%.
There is no argument between Gordon and his opponents over the facts represented by this graph. Indeed, Brynjolfsson and McAfee, for example, devote several chapters of The Second Machine Age to the “left behind” phenomenon. Gordon’s point is that headwinds like these represent a deduction against the growth benefits of future innovations. What caught my attention is that Jones, in a paper written with John Fernald of the San Francisco Fed, actually attempts to work this out. The attempt is significant in that Jones is both a distinguished growth economist and a long-time student of the ways in which ideas (read “innovation”) work their way into increased GDP.
Fernald and Jones calculate that for the period 1970-2007, educational attainment (or “human capital”, as growth economists call it) accounted for roughly 20% of that period’s 2% average per capita GDP growth. If educational attainment is slowing, just as if population growth (another 20% contribution) is slowing, then the 20% contribution of human capital to growth will get smaller and will have to made up by something else if 2% overall growth is to be maintained. This is precisely what Gordon means by “headwind”.
Can “innovation” itself make up the difference? The growth model used by Fernald and Jones already figures that innovation was been responsible for 60% of the 2% per year growth experienced between 1970 and 2007. The future contribution, they say, depends upon the future shape of the “idea production function”. [In growth theory, just as the whole economy is represented by a “production function” that turns inputs like capital, labor, and innovation into GDP – so there can be similar production functions that define the growth paths of the inputs.] As Figure 2 shows, there are several paths that the idea production function could take.
Figure 2. Fernald and Jones, “The Future of US Economic Growth”, 2014.
As Fernald and Jones explain, the recent slowing US growth path may reflect a “fishing out” of technologies currently in play, inadequately balanced by new efforts at R&D. This might continue. Or, we could reach an “inflection point” (just like the techno-optimists are predicting) where a new idea, or waves of new ideas, expand the possibilities. The point is, they don’t know. One could make assumptions that cause an production function to display explosive future growth, but today, since ideas account for most growth (60%) in the model used by Fernald and Jones, and since US growth is slowing, one can only conclude that the production of ideas is diminishing.
How can this be, given the rapid technological change we see everywhere around us? It may be only a lag in effect. Computers, after all, were widely used for over 30 years before they produced the famous spurt of growth seen in 1994-2004. It may be – and this is much discussed – that recent new technologies produce benefits of a different sort than are counted in GDP; smartphones may make us happier and more connected, but they don’t make us wealthier or produce many great new jobs. The counter argument here is that GDP statistics have always missed these kind of benefits; they didn’t measure how peoples’ lives changed with electrification and indoor plumbing.
Gordon allows that technological change is happening, but he says that it is narrowing. The concentration of innovation in computers, communications, and entertainment has reduced technological change to a succession of gadgets. If I buy a bigger TV every few Super Bowl Sundays, what does that really mean?
But I think the problem goes back to the headwinds. I can’t say whether innovation is “narrowing” or not, but our society is. Technology is changing the definition of who can really participate in and benefit from our economy. And, as Figure 1 shows, we are not producing enough people who can make the grade. If inequality is both the outcome of a technology-driven economy and a limiting factor on its success, then it may be that increased growth, if we ever get it, will come as much from social as from traditional engineering.
1 See the references for Gordon’s recent papers. Note also an online talk. A remarkably similar take from a different political perspective can be found in Tyler Cowen’s The Great Stagnation and his more recent Average is Over, also in the references.
2 A number of reviews of the book and responses to Gordon’s earlier papers can be found in the references.
Autor, David, “Skills, education, and the rise of earnings inequality among the ‘other 99 percent’”, Science, May 23, 2014. Online at
Brynjolfsson, Erik and Andrew McAfee, The Second Machine Age: Work, Progress and Prosperity in a Time of Brilliant Technologies, Holt: New York, 2014.
Cowen, Tyler, Average is Over: Powering America Beyond The Age Of The Great Stagnation, Dutton: New York, 2013.
Cowen, Tyler, The Great Stagnation, Dutton: How America Ate All the Low-Hanging Fruit of Modern History, Got Sick, and Will (Eventually) Feel Better, Dutton: New York, 2010.
The Economist, “The American Dream: R.I.P.?”, September 21, 2013.
The Economist, “Why economic growth soared in America in the early 20th century, and why it won’t be soaring again any time soon”, January 9, 2016.
El-Erian, Mohammed, “Navigating the New Normal in Industrial Countries”, Per Jacobsson Foundation, 2010.
Fernald, John G., and Jones, Charles I., “The Future of U.S. Economic Growth”, American Economic Review: Papers & Proceedings 2014, 104(5): 44–49
Glaser, Edward, “Those Were The Days”, Wall Street Journal, January 15, 2016.
Gordon Robert J, “Is US Economic Growth over? Faltering Innovation confronts the six Headwinds”, NBER Working paper series, 18315, August 2012.
Gordon, Robert J, “Why Innovation won’t Save Us”, Wall Street Journal, December 21, 2012.
Gordon, Robert J., “The Rise and Fall of American Growth: The U.S. Standard of Living since the Civil War”, Princeton, Princeton University Press, 2016.
Gordon, Robert J, “US Economic Growth is Over: The Short Run meets the Long Run”, in Think Tank 20: Growth, Convergence and Income Distribution: The Road from the Brisbane G-20 Summit, Washington, D.C., Brookings Institution, November 2014.
Jones, Charles I., “The Facts of Economic Growth”, 2015, forthcoming in John Taylor and Harold Uhlig, Handbook of Macroeconomics, Vol. 2.
Krugman, Paul, “Review: The Rise and Fall of American Growth”, New York Times Magazine, January 31, 2016.
Mokyr, Joel, “Is Technological progress a Thing of the Past?”, post on the Vox blog of the Centre for Economic Policy Research,
Mokyr, Joel, “What Today’s Economic Gloomsayers Are Missing: Science is enabling invention like never before and in ways that will improve life but isn’t captured by GDP statistics”, Wall Street Journal, August 8, 2014.
Porter, Eduardo, “America’s Best Days May Be Behind It”, New York Times, January 19, 2016.