Alan Jacobs has been musing lately about “limits to growth” (here’s his latest post). I’m going to join him. Let’s see whether this gets us anywhere.

First of all, what is economic growth, properly calculated? If the economy is growing, that means the productive output of the society has increased. Because in aggregate labor cannot be saved, this is equivalent to saying that the consumption of goods and services by the society has increased.

Measuring that increase is tricky. First of all, there’s the issue of quality. If I buy a computer every five years, is my consumption flat or increasing? It depends on whether the “quality” of that computer is identical. If my computer today is faster than my computer five years ago, most of us would agree that quality has increased. But what if most of that increase in speed has been “eaten up” by less-efficient software that is able to do all sorts of things I don’t need. In that case, do I really experience the rise in quality? If not, then in what sense has my consumption increased? These questions get even tougher when you start talking about the experience of dealing with automated software rather than human customer service personnel, or of bagging your own groceries, or of more water-efficient toilets. Are these changes an increase in quality? A decline? A case could be made for either.

Then there’s the question of measuring externalities. Suppose I build a dam which generates a lot of electricity but drowns a beautiful canyon. The new electricity is easy to measure, as is the direct cost of building the dam. What is the value of the canyon that was drowned? Do we measure it in terms of its then-current value in terms of leisure consumption? What if consumption patterns change, and that value goes way up after it is gone? How do we capture that loss properly? This is not even to speak of changes to the productive capacity of the planet caused by loss of habitat, etc. (I’m not even to address the possibility that measuring value in terms other than human consumption is worth trying to do – because that would take us outside of the realm of talking about economic growth and how to measure it.)

Recognizing all these measurement difficulties, we’re still left with, broadly speaking, four possible sources of economic growth:

  • Increase in population. More people producing and consuming = more output. There may be absolute Malthusian limits to this equation.
  • Increase in average labor hours. If, on average, people spend more time in productive activity, you get higher output. If you treat idleness as a form of consumption, then this reduces to a sub-instance of higher productivity.
  • Increase in resource inputs. If we discover a massive new source of oil, then the price of oil falls and it gets cheaper to make anything that is made with oil. Which means we can make more stuff. Again, this can be treated as a sub-instance of higher productivity.
  • Increase in productivity. If you get more output per labor hour, and all else is held equal, you get higher productivity. If we don’t call working more hours an example of higher productivity, and we don’t call the addition of new resource inputs an example of higher productivity, then what’s left to drive higher productivity is innovation in the process of production or in the product being produced. Note that you can get higher productivity not only by producing more stuff per hour, but also by producing the same amount of stuff per hour while producing “better” stuff.

As you can see from the above, not all economic growth is always good. Is population growth always good? Julian Simon notwithstanding, I think the clear answer is “no.” Sometimes it’s good, and sometimes it’s bad – and sometimes it’s good in the short term but bad in the long term, and vice versa. One source of wealth is human innovation, and arguably greater populations mean more innovators. But not necessarily – you’d need to see whether population growth imposes costs that retard innovation to a greater degree than the sheer number of minds increases it. You’d also need to see what segments of the population are growing more rapidly. And then, of course, once resources are exploited beyond a certain degree, population growth just means sharing those resources across more people – a decline in per-capita wealth. All this is true even before you reach true Malthusian limits where the growth in population leads to a catastrophic decline in aggregate wellbeing, and not just a decline in average individual wellbeing.

An increase in average labor hours is also not always good. This is a situation where measuring the value of leisure hours becomes vital. We are very prone to valuing more highly those goods and activities that are easier to measure, and this valuation process can affect our expressed preferences. So we should be careful about making cavalier assumptions about the relationship between market value and what one might call “educated” preference.

But, if we get measurement reasonably right, I’m inclined to say that an increase in resource inputs or an increase in productivity due to innovation is, on balance, pretty much always positive. When we decide these things are bad, it’s usually because we think we aren’t measuring externalities properly – the loss of that beautiful canyon, for example.

How important are these kinds of growth? Well, it matters who we’re talking about. If we’re talking about an impoverished Bangladeshi, the answer is: really, really important. If she could go from her current labor and consumption patterns to something more closely approximating the labor and consumption patterns of, say, a middle-class Turk, that would be a huge step forward. If that transformation happened on a large scale, it would represent a huge increase in economic growth. And that would be true even accounting for the wreck of the beautiful canyon.

Could such transformations be achieved through redistribution rather than growth? Not really – or, rather, what is really implied by arguments that they could is that redistribution of nominal wealth would lead to large increases in economic growth. A one-time transfer of wealth from, say, America to Bangladesh would not mean a convergence of consumption patterns unless that transfer enabled the people of Bangladesh to become substantially more productive. After all, a convergence in consumption patterns would mean much higher consumption in Bangladesh, and only modestly lower consumption in America, and therefore would require an increase in the productive capacity of the planet as a whole. So arguments about redistribution are really arguments about how to achieve growth, not about whether growth is necessary.

For most of the world’s population, economic growth is vitally necessary in order to escape profound absolute evils, even as this growth is going to create some new evils. Moreover, the main argument that such growth isn’t achievable is neo-Malthusian – an argument that absolute resource constraints or environmental consequences will make it impossible.

Some of the pessimists in the article about Robert Gordon that Alan Jacobs cites, are coming from that neo-Malthusian place, and that’s a difficult argument to engage with because too many variables are too comprehensively unknown. What isn’t hard to argue is that, politically, selling zero-growth to Bangladesh is a nonstarter. So we’re going to find out the hard way whether those Malthusian limits are near to hand, or further away, and in the meantime we’ll have to do what we can to limit the costs of growth – by restraining population growth, most fundamentally.

For the wealthier portion of humanity, however, limits to growth more and more are revealed as limits to measurement.

Take a look at Japan. Since the late 1980s, Japan has experienced virtually no growth as generally measured. This is partly a consequence of demographic change – Japan’s population is aging rapidly. But anyone who visited Japan in 1985 and again twenty-five years later would have instantly recognized huge improvements in living standards. How could that possibly be? It’s not because Japan is massively indebted – Japan is still a massive net foreign creditor; its debt is owed to itself (though this is just starting to change). It’s a measurement problem.

Measurement problems can cut both ways. Just to pick a random example, that jet skis suddenly became much more affordable. That would represent an instantaneous jump in productivity, and hence in aggregate economic growth. Lots more people would buy jet skis – and the utility of a jet ski would rapidly decline as the “best” places to jet ski got hopelessly crowded. But measuring this decline in value of the jet ski experience would never get properly measured, and factored into what looked initially like an instantaneous jump in growth.

On the other hand, massive increases in productivity in the manufacturing sector are, due to Baumol’s Cost Disease, a meaningful driver of higher prices in the service sector. Concert tickets become vastly more expensive – which may be an acceptable cost of productivity increases elsewhere. But so does the cost education, and health care, even if quality doesn’t rise. For those whose incomes aren’t growing fast enough, that may mean an absolute decline in the quality of education and health care (as well as an absolute decline in concert attendance). Which most of us would understand as a decline in quality of life that isn’t really offset by the increasing affordability of ever-better televisions. But it’s not clear whether this is properly measured either, because we have poor metrics for measuring quality in some of these fundamental services.

If you’re in Bangladesh, none of this matters that much. Lots of people are working extremely unproductively at the level of subsistence. The prospects for improving quality of life by applying already well-comprehended innovations are very substantial. The question is not whether economic growth is desirable but whether we know how to engineer it effectively (and whether we’re going to hit those hard Malthusian limits soon).

But if you’re living in Manhattan or Tokyo, the question isn’t “do we still need to grow in wealth as a society” but “would we know what real growth would look like if we saw it?” A lot of what people mean when they talk about living without growth could probably be re-cast as redefining how we measure growth, to more accurately account for externalities and more accurate measure experiential quality.

In the developed world, my instinct is that properly measuring growth will require properly accounting for improvements in environmental quality and in the quality of both essential and inessential services. The first of these should be self-explanatory. As to the second, as manufacturing employs fewer and fewer people, and as we reach the limits of how much stuff we can actually use, we’re going to run up against limits of how much additional growth productivity increases in that sector can achieve. Growth will then need to come from improvements in services – which are more likely to be measured in terms of higher quality than in terms of quantity, because services take time. And time is the one absolutely finite resource.

As for the prospects of a third industrial revolution – or a fourth, depending on whether you agree that the information revolution deserves the hype – here’s a hypothetical. Posit the development of two technologies: cheap, reliable fusion and cheap, reliable batteries. Neither is implausible to imagine within the next century. That combination wouldn’t quite get us to the universe of Star Trek, but it would be a huge revolution, essentially removing practical limits to energy generation. But what would the practical consequence of such a revolution be? I would argue that the most important consequences would be, first, that it would make it possible to repair much of the environmental damage caused by the second industrial revolution (and the continued advance thereof), and, second, that it would make the extra-terrestrial expansion of humanity plausible for the first time. Both would be momentous civilizational achievements. But would either register as economic growth as we currently measure it?

Jacobs ends by worrying that the challenge of our time is going to be preparing the next generation for relative poverty, and giving them the tools to survive under those conditions. I’m never going to argue against being prepared for the worst, but I don’t think it’s a good idea to give in to the assumption that we must expect relative impoverishment for the rising generation even if the pessimists are right about the prospects for revolutionary levels of growth in the first world are dim. What I think is true is that we have to grapple with the increasing importance of questions of distribution in developed societies.

Even if you project continued productivity improvement in information processing and manufacturing out into the future, which is a prerequisite to projecting continued economic growth, you eventually wind up in a world where much of what we think of as “work” involves very few people. A small group of people will have large entrepreneurial/managerial responsibilities with respect to this substantially-automated activity. And the vast majority of people will be engaged in services of one sort or another that are not readily subject to productivity improvement through automation. (This is basically the world that the optimistic Erik Brynjolfsson, mentioned in the article Jacobs linked to, expects to see.)

But it will only work if there is adequate remuneration for that service-sector employment! If there is not – if the entrepreneurial/managerial elite focus primarily on their own relative power, to the impoverishment of the society at large – then it’s easy to imagine an alternative world in which increasing portions of society are rendered (from the perspective of that elite) economically superfluous. (Note that this is different from what Marx expected; his impoverished proletariat was far from economically superfluous – they were producing all the actual value!)

I’m not a dystopian by any means; the right way to bet on most social trends is on the side of self-correction, and that’s especially true for social trends that appear to lead to dystopian conclusions. But if we want a world in which Alan Jacobs’s sons can live a comfortable middle class life even if they are high school English teachers or electricians, then we need to grapple with the economic forces that are going to make that outcome harder and harder. And that’s a good deal of what politics is for.