Stray Thoughts: Nobel Laureate Robert Solow’s Predictions for the Next Century

A leading economist considers such topics as the effects of climate change on economic growth, the rise of income inequality, and the shifting work year.
"We are not good at large-scale redistribution of income, and we do not seem to be getting any better."
By: Robert M. Solow

The questions that prompted Ignacio Palacios-Huerta to reach out to the world’s leading economists were universal: How will this world look when I am not here? Will there be other world wars? Will poverty as we know it today disappear? What kind of life will my great-great-grandchildren live? While not infallible, economists, Palacios-Huerta asserts, “know more about the laws of human interactions and have reflected more deeply and with better methods than any other human beings.” So why not just ask them?

That is the premise for “In 100 Years,” a collection of essays commissioned and edited by Palacios-Huerta for which prominent economists were invited to offer their ideas about what the future might hold. One of those essays, by Robert M. Solow, who won the 1987 Nobel Prize in Economic Sciences for his contributions to theories of economic growth, is featured below. Here, Solow muses on topics including the shorter work year, the effects of global warming on economic growth, and the rise of income inequality.

The Last Hundred Years Are Hard Enough

One hundred years is a very long time, perhaps not on the evolutionary timescale, but certainly on the economic. The conventional estimate of real national income per person in the United States in 1910 puts it at 19.2 percent of its value in 2010. That represents an average annual growth rate of 1.66 percent. There were no national accounts in 1910, so the number itself is doubtful. Besides, it is not clear how to translate that numerical abstraction into a comparison of lived experience or “standard of living.”

This essay is excerpted from the book “In 100 Years: Leading Economists Predict the Future,” edited by Ignacio Palacios-Huerta.

We have a feeling for the 2010 figure of $43,000 per person. We know roughly what it makes possible and what it excludes. The 1910 average of $8,300 per person (“in constant prices”) is much foggier in content, as well as less accurate. Much of the contents of a 2010 standard of living was not available in 1910, was not even thinkable. For that matter, even a 1910 apple was rather a different piece of fruit from the 2010 model going under the same name. How could someone in 1910 have had any grasp of what economic life would be like a century later, even with a good guess at the growth rate of real GDP? The answer to that question was not knowable. I do not mean just the answer was hidden from view, but that it had not been determined. Nevertheless, if Rip van Winkle Jr. had fallen sleep in 1910 and awakened in 2010, he would have been surprised and bewildered, but it would be recognizable human life that he was seeing. The experience of Rip van Winkle III will presumably be similar.

The Next Hundred Are Harder

Those are qualitative difficulties; there are intrinsic quantitative problems as well. Something that starts equal to 1 and grows at an average rate of 1.2 percent a year for 100 years will be equal to 3.30 at the end of the century. If instead it had grown at 1.6 percent a year, it would have ended at 4.89, larger by half. Think of this something as productivity, or total factor productivity, or even income per person. The interval from 1.2 to 1.6 is a plausible range of growth rates. In fact, I could have easily chosen a wider interval with a more dramatic difference in end points. But no one today can seriously argue for one number rather than another in that interval as a forecast for productivity growth over the next 100 years. There may be no harm in guessing, but a mere guess is not an argument. The only justification for making such a forecast today is that you will not live to know how wrong you were. That may not be justification enough.

Nevertheless, I will carry on a bit longer, though I will soon raise further questions. Median family income in the United States is currently a little more than $60,000. Is it possible to imagine that the corresponding figure for 2113 should be four times that, in constant prices? (That is a bit less than a fivefold increase in GDP per person from 1910 to 2010. At this point, averages or even medians seem inadequate, and one wants to know more about distribution, but never mind.) Yes, I suppose it is imaginable. In current American political discourse, perhaps not the acme of sobriety, an annual income of $250,000 is described as the top of the “middle class.” No one seems to laugh. Why would the median family a century from now not be able to achieve and enjoy what passes today for a high middle-class standard of living?

In thinking about all this, one should keep in mind the steady shift of consumer spending toward services and away from material goods. In 1960, 47 percent of all (nominal) consumer expenditure went to services; by 2009 that share had risen to 68 percent. The contrast with 1910 would be even sharper. That shift is likely to continue; one expects education, personal care, tourism, recreation, financial services, and the like to have a high-income elasticity of demand, though health care is complicated and remains to be settled. Even so, the pace of shift to services is uncertain over so long a period, with so large a potential increase in income, and so much possibility for changes in preferences and technology.

Hours of Work

A major uncertainty, with important implications for the pattern of consumption, has to do with the evolution of decisions about work and leisure. During the first half of the twentieth century, the average annual hours of a full-time American worker tended to fall slowly but fairly surely. Then forty or fifty years ago, that tendency seemed to diminish substantially or disappear. Nowadays American (and Japanese and Korean) workers put in many more hours per year than counterparts in the high-income countries of Europe. Average annual hours worked in the United States are about a quarter higher than in France and Germany. There is controversy about the causes of this discrepancy.

I find it easy to believe that the transition to much higher incomes will lead to a shorter work year in the United States, especially if the age of retirement increases.

One view often expressed is that the source of the difference is “cultural.” Americans like to get ahead. In a consumerist society, that means making more money and spending more. Europeans may be less interested in stuff and more interested in leisure. (It is sometimes forgotten that some popular uses of leisure require quite a lot of stuff.) When increasing productivity compels a choice, Americans are generally inclined to choose more goods, Europeans more likely to choose a shorter workday and longer vacations. The other view is that the all-in tax rate on the income from a marginal hour worked is considerably higher in Europe than in the United States. Even without any major differences in preferences or social norms, routine behavior responses would lead Europeans to work fewer hours than Americans.

There is probably truth in both arguments. Personally, I find it easy to believe that the transition to much higher incomes will lead to a shorter work year in the United States, especially if the age of retirement increases. That would at least replicate the difference between 1910 and today. But there is no solid basis for a guess.

But then what about the work year of a unit of real capital, a “machine”? There is no automatic answer. Suppose, just to simplify, that the size of the labor force remains constant. Then a reduction in average hours worked per year is an equal proportional reduction in total hours worked. Suppose we think in terms of given capital intensity (total machine hours/total labor hours). So total machine hours falls too. By definition, the stock of machines is given by total machine hours/average hours worked per machine. One extreme is that average hours worked per machine remains constant (if all machines worked continuously, 24/7). Then the necessary stock of capital falls too. At the other extreme, one can imagine that the work year of a machine falls with the work year of labor (e.g., if “my” machine works only when I am working). Then the necessary stock of capital is unchanged. Anything in between is clearly possible.

Will leisure time activities be especially capital intensive or the opposite? Show me an economist with a strong opinion about these things, and I will show you that oxymoron: a daredevil economist.

I think this is the natural range of possibilities, although there is no logical or physical reason that the work year of a machine should not actually increase, say. But it would seem more likely that increased leisure over the next century should be accompanied by a smaller stock of capital (per worker), smaller gross investment (per worker), and thus a larger share of consumption in GDP. Of course, this tendency will almost certainly be offset by an ongoing increase in capital intensity, even in the service sector. Obviously there are other, totally moot, considerations. Will leisure time activities be especially capital intensive (grandiose hotels, enormous cruise ships) or the opposite (growing marigolds, reading poetry)? Show me an economist with a strong opinion about these things, and I will show you that oxymoron: a daredevil economist.

Climate, Environment, Resources

There is still a potentially larger uncertainty to be reckoned with even if we continue to think only of the already developed world. One hundred years is long enough for the effects of global warming to limit economic growth — perhaps marginally, but perhaps drastically. The model predictions are themselves uncertain; on top of that, it is impossible to know how policy will respond or how those responses will affect measured output and income. As of now, one would have to say: hardly at all. But the policy response might have to change in the course of a century as climate events unfold.

Apart from climate change, other induced environmental effects on air, water, land use, and urban livability may manifest themselves and lead to changes in economic life. The continued and expanded use of nonrenewable natural resources could lead to either effective exhaustion or sharply rising relative prices, either of which would alter growth prospects over a century. Simple extrapolation will hardly do if the goal is to look far ahead. One interesting reason is the shift to services. Casual (excessively casual?) thought suggests that most services substitute human capital for material capital and resource products. (I inserted the word most because I have an appointment with my excellent and all kinds of capital-intensive dentist tomorrow!) If the world were to go Voltairean and choose mainly to cultivate gardens, the input-output table could change a lot. Production would put less strain on the scarce resources base and the waste disposal capacity of the environment.


The current episode (if it is only an episode) of increasing income inequality in the United States dates from the 1970s. The odds that it will continue must depend on its deeper economic or social sources. Is it primarily a by-product of the growth of the financial sector and its proclivity to pay enormous sums to successes and failures alike? Or does it have something to do with an underlying market trend in the relative compensation of labor, human capital, tangible capital, and entrepreneurship? And I have not mentioned other possible influences, like international trade, migration, the decay of labor unionism, or the distribution of educational opportunity. Those are hard questions, and any estimate of the future and the appropriate policy response (which may alter the future) depends on the answer.

I will focus on the second possibility, shifts in the remuneration of labor and capital, not because I am convinced that it is the right one, but because it is what economists are used to thinking about. It is clear that something substantial has been happening. We used to think the proportion in which income is divided between labor and capital as one of the great constants on the economy. Not since 1960, however: There has been a definite tilt against labor income. We can see that from two different vantage points.

We used to think the proportion in which income is divided between labor and capital as one of the great constants on the economy. Not since 1960, however: There has been a definite tilt against labor income.

First, since 1960 the real compensation of labor in the nonfarm business sector has lagged well behind productivity. Output per hour rose by a factor of 2.82; real compensation per hour (which includes benefits) rose only by a factor of 1.94. Second, a quite different set of figures tells us that compensation of labor was 72.1 percent of all nominal personal income in 1960; in 2009, it was 63.7 percent. On either scale, this may seem like a minor change. By the standards of the past, however, it is high drama. If the 1960 proportion had ruled in 2010, roughly $1 trillion of income would have gone to labor that in fact went to other forms of income. (I am not bothering to disentangle business cycle effects from trend effects. Any way you slice it, this is a big deal.)

Neither set of numbers tells us exactly what we want to know. But it is unmistakable that nonlabor income, whether return on tangible capital investment, rewards of entrepreneurship, monopoly profit, or something else, has gained at the expense of labor income (which presumably includes much of the return on human capital). One way or another, this is the outcome of complicated market forces.

Can This Go On?

Should we expect this drift to continue into the future? For that we would need to know more about those “market forces.” Some of the eligible market forces include (1) changes in the ease with which capital can be substituted for labor as economy-wide capital intensity increases, (2) changes in the nature of new technology, (3) changes in the industrial composition of aggregate output, (4) changes in the amount and distribution of monopoly power, not to mention institutional changes like (5) decay of unionism and (6) the balance of political power. “All of the above” is the easy, if unenlightening, answer.

In the context of speculations about the next 100 years, it is interesting to think about the implications if this trend were to continue. If the underlying source is embedded in the composition of aggregate output and the nature of technology, then the shift away from labor income would be hard to reverse. We are not good at large-scale redistribution of income, and we do not seem to be getting any better. So one possibility is a remorseless reduction in the share of income going to human labor, probably accompanied by increasing inequality. (Absolute wages could, of course, continue to rise.) This story is reminiscent of the recurrent bad dream of an economy in which robots do all the production, including the production of robots, with the proles on the outside looking in.

We are not good at large-scale redistribution of income, and we do not seem to be getting any better.

That is pushing the trend pretty far on no evidence. But even a moderate extrapolation would seem to call for a response. That might take the form of a democratization of capital: as wage income shrinks (relative to the total), ordinary people could draw more of their income from capital. The capital would have to come in part from their own saving — for example, funded pensions — and in part from capital accumulated on their behalf by the state, maybe in the form of mutual funds. You realize that this is a fantasy. The reality will be more pedestrian.

The Rest of the World

Everything I have said so far has been about the rich, developed countries. But of course most of the world’s population lives elsewhere, in the poor countries or in the emerging economies. For them, the puzzles take a different form: Will those economies stagnate or, if not, must they, will they, industrialize before deindustrializing, more or less following the historical pattern? Or alternatively, with the example of the rich countries in front of them, available for imitation, will they go from potato farmer to couch potato in a much smaller number of generations? Obviously there is no law of economics about ontogeny recapitulating phylogeny. Nevertheless, it seems clear that some emerging economies will, for at least some time, have a manifest comparative advantage in labor-intensive manufacturing.

There are several related reasons for expecting development to take this form. If development succeeds at all, if more economies “emerge,” the world demand for manufactures and other material goods will certainly grow rapidly. All those people with rising incomes will have to accumulate household goods: houses and their contents, cars or other means of personal transportation, the various public and private goods that accompany at least slightly affluent urbanization, and so on. Satisfying that demand will require investment in productive capacity, much of it presumably local.

When (or if) this happens, it will happen at a twenty-first-century level of technology. Even so, because many of the goods in question are not terribly complex and because a large supply of low-wage semiskilled labor is available, the likeliest outcome is the growth of local manufacturing, construction, and similar sectors in the next cohort of emerging economies, just as happened with its predecessors. These industries will be more sophisticated than those that characterized earlier industrialization. They should provide a natural training ground for the literacy-related and numeracy-related skills that will later, sooner than in the past, lay the basis for the normal shift to services.

Of course, we are all aware of the appearance of tradable-service sectors in countries like India and China. They seem to be fairly small in employment terms, however, and not (or not yet) the signal of a massive shift in comparative advantage. One would not be surprised, however, to see this shift occur faster than it did with the current advanced economies. I suspect that much depends on the speed and efficiency with which the currently poor countries can organize their educational systems and, what is just as important and maybe more difficult, arrange for meritocratic social mobility. I would not know how to quote the odds. This is not to deny that there could be other success stories, like the Bangalore-based IT sector in India, but it is a good guess that this will not be the norm.

The trajectory of the world economy will depend on the speed with which those countries grow and the qualitative nature of their growth. I am referring to such obvious matters as the drain they place on the world’s supply of water and other natural resources, on the care they take with environmental amenities (including their contribution to climate change), and, above all, on their success in speeding up the demographic transition to slower-growing or stationary populations.

How Will It Add Up?

One has to suppose that the successful developers in the world economy will be able to grow faster than the old industrial countries were able to do in the nineteenth and twentieth centuries. The possibility of technological catch-up and the relatively easy availability of a flow of investment capital from the already rich world are a major advantage accelerating the growth of the latecomers. The major uncertainties appear to be political. I could go on at length about governance issues because I know so little about them. It is enough to observe that the passage from national poverty to sustained economic growth requires favorable policy commitments over long periods. Anarchy, violence, cronyism, and corruption are not features of a successful trajectory.

Eventual resource scarcity and environmental stress (especially climate change) are another matter. No one can know with any precision how many people can be supported in the world at a standard of living anything like that of the currently advanced countries. As always, that depends on an implicit race between scarcity and new technology. Demography and technology are forces that are at least partially open to influence by public policy.

I guess this sounds modestly optimistic. Keynes was famously optimistic and worried less about things like excess population and environmental stress in the developing world. He dwelled more on the likelihood that as incomes increased and work time fell, ordinary people would be at a loss to find things to do with their leisure. I do not fully subscribe to Jeremy Bentham’s opinion that pushpin (whatever that may be) is as good as poetry. But that particular worry is fairly low on my list.

Robert M. Solow is an American economist who received the Nobel Prize in Economics in 1987. This essay is excerpted from “In 100 Years: Leading Economists Predict the Future.”

Posted on
The MIT Press is a mission-driven, not-for-profit scholarly publisher. Your support helps make it possible for us to create open publishing models and produce books of superior design quality.