Benjamin Friedman on the Future of Productivity Growth in America, the True Meaning of Sustainable Economic Growth, and Keynes’s “Grandchildren”

June 2019. GrowthPolicy’s Devjani Roy interviewed Benjamin Friedman, William Joseph Maier Professor of Political Economy at Harvard University, on the future of productivity growth in America, the true meaning of sustainable economic growth, and J. M. Keynes’s “Grandchildren” essay. | Click here for more interviews like this one.

Links: Benjamin Friedman’s faculty page | Publications | NBER research page | The Moral Consequences of Economic Growth (Penguin Random House, 2006)

Growthpolicy.org. How should we promote economic growth?

Benjamin Friedman: We know very little about how to influence the long-term rate of growth of economic activity.  Presumably such matters as basic research, investment in both physical capital (including infrastructure) and the human capital of the labor force, and the efficiency of the process by which we allocate both investment and labor all matter.  In recent years we’ve been falling behind in every one of those categories.  In principle, therefore, there are plenty of obvious things to do.  But resources and human energy are limited, and so it would be helpful to know, quantitatively, which direction to emphasize in whatever steps public policy takes.  Alas, we don’t have much real evidence to help us make that judgment.

Growthpolicy.org. Where will the jobs of the future come from?

Benjamin Friedman: Recent research suggests that the need for human labor will be primarily of two forms: organizational and managerial roles, and craft-type labor (think carpenters and plumbers).  By contrast, many forms of labor that are somewhere in between—sales jobs, routine clerical jobs, knowledge-related jobs that don’t involve creativity—will lose out to some combination of robots, computers, and artificial intelligence.  The unfortunate consequence will be a further bifurcation of the labor force, with more and more people having either well-paying (and otherwise satisfying) jobs or low-paying and unsatisfying jobs.  The implication for our society and our politics is not good.

Growthpolicy. What should we do about economic inequality?

Benjamin Friedman: I suggest two partial remedies.  For the low end of the income scale, research is clear that there are very large returns to providing increased education, beginning with pre-school, for children whose family situation indicates that they are otherwise unlikely to benefit as they should from the regular education provided.  We also know that doing this effectively is expensive; the forms of early intervention that have been attempted on the cheap don’t work.  But the effort is worthwhile.  It will address economic inequality by improving the prospects of people who are otherwise likely to end up at the left tail of the distribution.  It will also enhance the economy’s overall productivity and growth, by making people who are otherwise likely to have limited productive skills more productive.  And, on the available evidence, it will improve non-economic outcomes as well, such as fewer young people getting in trouble with the law, fewer teen pregnancies, and the like.

To address the high end of the scale, we should reform the governance of publicly held corporations.  I believe that in many cases companies’ shareholders do not support the outsized compensation that the firms’ executives are paid (often even when the companies perform poorly).  It is remarkable how vigorously U.S. corporations have resisted having even a non-binding shareholder vote on executives’ compensation; the shareholders own the firm, and they should be able to say—in a vote that is binding—what they pay the managers who are their employees.  There are many other aspects of corporate governance that need to be changed as well.

Growthpolicy.org. How should we prevent the next financial crisis?

Benjamin Friedman: The Dodd-Frank Act was a good start, but as our principal response to the worst financial crisis in two generations it was very disappointing.  Moreover, some elements—for example, barring the Federal Reserve System from the kind of lending it did in the last crisis except with explicit authorization from the Treasury Secretary—were counter-productive.  The chief improvement that has occurred so far is higher capitalization of U.S. banks.  So far so good, but capital ratios should still be significantly higher.  We also need new procedures for handling bank failures; one good suggestion for this purpose is a new form of bankruptcy (the “Chapter 14” proposal).   I would also favor tighter restrictions on the derivative transactions of financial institutions (not limited to banks), requiring larger down-payments for home mortgages, and other similar measures to reduce risk.

Growthpolicy.org. I’d like you to comment on the direction in which you see our current third industrial revolution headed in the U.S. in terms of future growth in productivity. (Economist Robert J. Gordon has argued, for instance, that America’s “special century” of productivity growth, 1870 to 1970, will never again be replicated.)

Benjamin Friedman: I don’t think we know what pace of productivity growth future technological development will bring.  Bob Gordon makes a persuasive case that the rapid productivity growth the United States achieved from about 1870 to 1970 was a consequence of a unique set of specific advances.  By contrast, Erik Brynjolfsson makes a persuasive case that robots and artificial intelligence are about to deliver a new surge of productivity growth beyond anything we have seen in the past.  I can’t judge who’s right.  I don’t think anyone can at this point.

But either way, we need to bear in mind that while productivity growth can mean more output from the same input of labor, it can also mean less labor input needed to produce the same amount of output.  I fear that over the next generation—say, the working lifetime of our current university students—the second feature will predominate.  The labor market will clear (unless we raise the minimum wage too high), but new technology will degrade the skill requirements for many jobs and over time the wage for those jobs will converge down to the level warranted by the lower skills.  I think this will be the chief economic problem of the next few decades.  And it will happen regardless of whether Gordon or Brynjolfsson proves right about overall productivity.

Growthpolicy.org. We are approaching the ninetieth anniversary of Keynes’s celebrated “Grandchildren” essay, more formally called “Economic Possibilities for our Grandchildren,” a piece of writing abundant in cheerful predictions about the future of consumption, output, and work. As we know, Keynes was right about growth in output per capita but wrong about the shrinking duration of the workweek: “3-hour shifts or a 15-hour workweek.” Indeed, in the past decade we have moved to a “gig economy” in North America and Western Europe, in which workers take on part-time jobs, often several at a time, thereby extending the workweek but as independent contractors bereft of the security net of health insurance or retirement benefits. As someone who has written on Keynes previously, why do you think he was wrong, or what do you believe he missed, about the future of work?

Benjamin Friedman: I’ve written twice about Keynes’s “Grandchildren” essay, and my two papers are different.  Sometime after writing the first one, I changed my mind on the matter. [Editor’s note: The two papers are “Economic Well-Being in a Historical Context” (2008) and “Work and Consumption in an Era of Unbalanced Technological Advance” (2015).]

Keynes was remarkably prescient in anticipating the rapid productivity growth over the century following 1930 (when his essay was published).  He predicted that output per person, in an economy like the U.S. or the U.K., would increase by a factor of between four and eight.  As of now, nearly 90 years into his century of prediction, the U.S. is almost exactly on target for an eight-fold increase.  (Before the 2007-9 financial crisis, we were on track for a nine-fold increase.)  By contrast, Keynes’s prediction that the average workweek would continue to decline, reaching fifteen hours, has clearly proved wrong.  When Keynes wrote, the workweek in the U.S. and the U.K. had been falling rapidly for at least a half-century.  It continued to fall until sometime not long after World War II.  Then it flattened.  It has held pretty steady ever since.

What Keynes failed to anticipate was the return—also not long after World War II—to widening inequality.  Thinking about the choice of how much to work as if all citizens were enjoying an increase in incomes and living standards in pace with the aggregate economy’s per capita output was not the right way to approach the question.  For nearly a half-century now, most of the returns to the economy’s incremental production have accrued narrowly to people at the top of the income scale.  In the middle of the distribution, people have earned little more.  As a result, most people today feel the need to continue working about as much as their grandparents did fifty years ago.

Growthpolicy.org. At Harvard, you teach a very popular course titled “Growth, Technology, Inequality, and Education,” that combines insights from a humanist (Professor James Engell) and an economist (yourself). What does sustainable growth, one that aligns economic progress with the progress of the human condition, mean to you?

Benjamin Friedman: Most of the time, and in most places, economic growth in our usual sense goes along with clear and measurable advances in the human condition.  The evidence is clear that, in most countries of the world, higher incomes and improved material living standards translate fairly directly into longer life expectancy, reduced morbidity, greater literacy, and many other similar benefits (see, for example, Angus Deaton’s The Great Escape.)  Even in countries where most of these benefits have already been realized, sustained improvement over time in the material living standard of the broad bulk of the population leads to non-material improvement in such matters as openness of opportunity, tolerance, generosity to the disadvantaged, and commitment to democratic institutions (see, for example, my The Moral Consequences of Economic Growth.)

What most people mean by “sustainability” is not a matter of whether economic growth brings these benefits—it does—but whether the growth also brings conditions that will eventually prevent it from continuing.  One such possibility is that the growth in per capita output comes with increasing inequality, a concern familiar from Tocqueville’s writings about the very beginnings of American industrialization in the 1830s, and in many settings ever since.  (As I indicated just above, for purposes of my own work I don’t call that “growth.”)  Another familiar possibility is that the growth brings environmental damage that eventually becomes too great to bear.  (This is a particular interest of Professor Engell’s.)  Within my working lifetime as an economist, concerned people have focused on polluted air and water, scarcity of resources such as oil, and change to the earth’s climate.  The first two of those concerns have faded—we now know that advanced economic growth brings reduced pollution, for example—but the global climate issue remains unsolved.

In the course that Professor Engell and I teach at Harvard, we discuss these and other issues surrounding economic growth.  We treat the four issues highlighted in the course’s title as intimately related:  Economic growth per se has important societal consequences.  Technology is a key driver of growth, and it influences the shape the growth takes.  Growth sometimes comes with widening inequality; and if so, the inequality is not only unwelcome in itself, it blocks many of the societal benefits growth would otherwise trigger.  Education is not only a key input to the growth process, as a central element in investment in human capital, but also an important vehicle for addressing inequality.