Review of DeLong's "Slouching Towards Utopia"

Posted by Dietrich Vollrath on October 11, 2022 · 21 mins read

Brad Delong recently published his long-gestating economic history of the long 20th century, Slouching Towards Utopia. If you are reading this blog, you are almost certainly aware of that, and I would guess the probability that you’ve read it (or are reading it) are around 50/50. It has garnered well-deserved high praise from a variety of people. To avoid just repeating what everyone else is saying, let me simply say that it is a necessary book to read if you are at all interested in the subject of economic growth. It is informative and thought-provoking. If you are one of those 50% who have not read it, go do so now.

One of the things DeLong has been asking online is what he got “big time wrong”. While I don’t think there is anything that is big time wrong, I do think there is a feature of the 1870-2010 period that does not get sufficient attention. I doubt it is because DeLong doesn’t appreciate the relevance, but at some point in a 500-odd page book, choices must be made. Somewhere in the multiverse is a reality where DeLong has bludgeoned his editor to death and produced a 1500 page tome that includes a few chapters on what I’ll expand on below. One way to consider this post is something a supplement for the STU volume that manifested itself in this particular strand of the multiverse.

DeLong makes clear that it wasn’t until around 1870 that productivity growth finally grew large enough to outpace the increased population growth that had always previously occurred in response to higher living standards. Productivity growth outpacing the population growth response made possible historically unprecendented growth in living standards after 1870. DeLong attributes that increased pace of productivity growth to some combination of the research lab, the corporation, and globalization.

I’m sure we can quibble about the precise natures of timing or the importance of those three factors in productivity growth. But what I’m more interested in is that very relevant and ultimately very necessary change in how population growth responded to living standards that occurred around the same time.

What happened around 1870 was not just that productivity growth was able to outpace population growth, but that population growth started to fall in response to increase in productivity. This was a fundamental flip in the typical Malthusian relationship that had held up until (roughly) that point in time. Prior to this inflection point, productivity growth had increased before, but every time had induced a response of higher population growth rates, eroding any gains in per-person living standards. After this inflection point, when productivity growth increased it induced lower population growth rates, accelerating the gains in per-person living standards. Without this flip in the population response, the productivity growth associated with the research lab, corporations, and globalization would not have generated the sustained growth in living standards that form a core part of STU.

It’s not that DeLong ignores the population response. Throughout the book he points to the demonstrable decline in fertility rates, using the vivid description that prior to 1870 the typical woman might spend 20 years of her adult life either pregnant or breastfeeding, while by the mid-20th century it might be only four years. What I’m nitpicking here is that this decline in fertility deserves equal billing with the research lab, corporation, and globalization. Without the flip to an “anti-Malthusian” negative relationship of fertility and productivity, you don’t even slouch towards utopia, you stagnate.

Consider that with material needs taken care of, human populations are capable of a shocking rate of growth. A common example to trot out here is the Hutterites. Between 1880 and 1950 - so exactly during the period when productivity growth had accelerated demonstrably - their population increased at a rate of over 4 percent per year. Each woman had approximately 10 children, and given that they did not marry until their early 20s, its is conceivable that they could reach a growth rate of 4.5 percent or more, with each woman having 12 to 14 kids.

The point here is not that this is an ideal to aspire to. The point is that it was entirely possible for human populations to respond to the productivity growth inherent in the research lab, corporations, and globalization by doing what they had done throughout history. It was plausible to respond to increased abundance by having more kids - a lot more kids. DeLong’s rough index of productivity growth is that it grew by about 2.1 percent per year from 1870 to 2010. One does not need to imagine that everyone was a Hutterite to think that human populations could sustain 2.1 percent growth or more over long periods of time, thus offsetting the productivity growth in a most Malthusian way.

It was not enough that productivity growth accelerated around 1870. It was an acceleration in productivity combined with a decline in population growth that put us on a slow trudge towards utopia. The decline in population growth was entirely a-historical at that point. Without this anti-Malthusian response the research lab, corporation, and globalization cannot provide sustained growth. But it is also the case that without the research lab, corporation, and globalization the anti-Malthusian response would not have occurred. My case is not that the demographic response is primary or more important, but that it is a necessary complement to those productivity boosters.

Making kids cost more

My preferred structure to think about the population response to productivity is that it follows the following rough formula

\[\text{Pop growth} = \textit{Parameter} \times \frac{\text{Income}}{\text{Price per child}}\]

This is really an expression for fertility, and I could add something explicit accounting for mortality rates, but this ultimately gets the right intuition. The price per child is a measure of the material costs to parents of getting a child “up to speed” with life. Basic food, clothing, and shelter, yes. Along with enough cultural and technical knowledge to care for themselves (and often for the parent in old age). In other words, the price of a child is the price of getting them to be a functioning adult.

In a Malthusian setting, in general what we get is that productivity growth increased income, but did not have a significant effect on the price per child. Hence an increase in income increased population growth, and ultimately any productivity growth was eaten away (literally, to some extent) by congestion and crowding and diminishing returns to resources. At that point ncomes fell back to the point that population growth was just keeping pace with productivity growth. When productivity growth is positive then this allows population growth to be net positive (births just outpacing deaths), implying that income is stagnant but above some biological subsistence level, and that the size of the population goes up every year.

Holding the price per child constant, the faster is productivity growth in this Malthusian setting, the higher is that stagnant income level. This gives a way of understanding how living standards could creep up over time even as everyone remained tied down by the Malthusian logic. Before 1870, J.S. Mill and others were all underwhelmed by the living standards that rapid (relative to history at that point) technological change had wrought. Each was convinced that the Malthusian logic held and perhaps doomed the common man to stagnant living standards even as science and technology remade the world.

After 1870, though, we get something new. Productivity growth did increase in speed, raising incomes even faster than before. This, by itself, would not change the overall logic. It would have allowed a somewhat higher - but ultimately still stagnant - level of income per person. Humans could have procreated fast enough to make this happen.

The key change is that productivity growth, starting around 1870, reached down and profoundly raised the price of bringing a child to adulthood. I have in mind a few forces here.

  1. Technological change led to demand for education. To riff on DeLong’s characterization, around 1870 we had innovated away from using our human backs (with new energy sources) and fingers (with machines) to do things, and were innovating ways to replace some of our cognitive skills. Working with those new technologies required more abstract skills, which in turn led to more demand for education. More education per child raises the price child. Alternatively, you could argue (as we in higher education often do) that education enables you to deal more easily with rapid technological changes. The acceleration of productivity growth in 1870 would have raised the return to education in that sense, raising the cost of a child.

  2. Urbanization. People had been moving to cities for centuries, of course, but corporations, globalization, and research labs accelerated that move. Urban areas are often argued to have higher child-rearing costs. That may be direct costs (relative price of food or shelter/space may be high compared to rural areas) or indirect costs (they may not have as much use as household workers compared to rural areas). Either way, by changing the location of economic activity it raised the cost of children.

  3. Health and nutrition improved. Child mortality changed in a fundamental way, and for any given number of desired adult children, you could have fewer pregnancies, good for moms. But beyond that, it meant that investments in those children - education, skills, migration - were likely to pay off. That raised the desired investment per child, raising the price of each child.

  4. Beckerian time costs. Kids take time. As incomes levels increased due to productivity growth, this increases the opportunity cost of time for parents, perhaps limiting their willingness to have kids. More practically, it probably works to increase marriage age as people acquire more resources/skills/experience prior to starting a family.

There should be like 1000 citations attached to these points. But I’m trying to finish this off today so my apologies to all the people whose work I didn’t explicitly cite (except, I guess, Gary Becker).

Regardless, I think a crucial aspect of the story from 1870 forward is not just the acceleration of productivity growth, but the fundamental effect of this productivity growth on the price of kids. This generated the anti-Malthusian response which was necessary to the sustained increase in living standards behind STU.

Let’s really have some fun

In the book, DeLong sets up Hayek and Polanyi (at some point forced into an awkward reconciliation by Keynes) as the intellectual poles of the story. Hayek and Polanyi, and to a great extent Keynes, were writing and thinking about economic growth well after this flip to an anti-Malthusian population response has occurred. While they all would certainly recognize Malthus’ argument and understand the logic, what I think is notable is that they all generally took the anti-Malthusian population response of their contemporary period as a given.

Moreover, I think that their positions are in some sense dependent on the anti-Malthusian population response. Take Hayek: the market giveth, the market taketh away, blessed be the market. This became a viable theory only when the market had in fact delivered a series of demonstrably good outcomes. Hayek could look backwards and point at definitive improvements in living standards and attribute them (often, but not always, correctly) to markets. With the anti-Malthusian response, the productivity gains after 1870 does not just allow for higher living standards, but noticeably higher living standards.

There were markets in existence well before 1870, and well before the Industrial Revolution. There is a whole other post I have some intention of writing about the a-historical nature of claims that people existed in some kind of pre-market or non-market economy prior to like 1800 (I’m looking at you, Karl Polanyi). Regardless, prior to about 1870 it would be hard to make the Hayekian claim about markets because they were not delivering the massive growth in the pie that let you elide distributional questions associated with pie-slicing. Markets at that point were just another way of slicing a pie that appeared to be somewhat constant in size.

Hayek could even make the case (Beware, completely untested hypothesis follows) that the market was the reason for the flip to anti-Malthusian population responses. Once markets put a more explicit price on children (education costs, etc..) that was when people flipped their behavior, ridded themselves of the Malthusian response, and delivered the boost to growth in living standards observed after 1870. Filtering family decisions through the market was exactly how we were able to slouch much closer to utopia. Without making the costs of children explicit through markets, we’d be stuck with continued Malthusian stagnation.

Okay, that’s Hayek. What about Polanyi?

I read Polanyi as being about identity. We are very protective of our identities, and those identities often coalesce around occupations (farmer, doctor, weaver, etc.) and place (the family farm, a block or neighborhood, a region). In many cases the pieces of those identities overlap; they didn’t call them the Pittsburgh Steelers by accident.

The danger that markets pose is that they want to treat people and places as interchangable and substitutable. The “labor market” wants to treat each individual as an identity-less lump of labor effort. The “land market” wants to treat each place as an identity-less location. The market will find the most efficient allocation of that identity-less labor across a bunch of identity-less places.

Those people and locations will, according to Polanyi, refuse at some point to be treated in such a manner and will protect those identities. Steelworkers in Pittsburgh defined themselves to a great degree by those two characteristics, and their families and children defined themselves by those characteristics even if they were not (a) steelworkers or (b) living in Pittsburgh. It is important to people that those identities are maintained even if the market is demanding that steel work be done in Mumbai. Those steelworkers will act in ways - organizing politically to put up trade barriers, for example - to save those identities.

What does population growth have to do with this? (Beware, completely untested hypothesis follows) When population growth remains high it is (more) plausible to maintain identities in the face of market changes. I don’t need all my children to farm our ancestral land so long as some of my children farm the ancestral land. I don’t need all my children to become steelworkers so long as some of my children carry on the family trade. When productivity goes up, the Malthusian response generates more kids, allowing some to remain in place and in roles that they’d traditionally be expected to occupy. It also allows some set of kids to move off to new locations and try new jobs. With the Malthusian population response it is plausible to generate a slow transition of occupations and locations. Places decline and occupations disappear, but they decline and disappear over generations, not years. Steelworkers get to retire as steelworkers in the homes they raised their families in in Pittsburgh. With a generational transition, old identities can be retained while new identities are built in new jobs and new places.

When we flip over to the anti-Malthusian response, however, the disruptions caused to identities by productivity growth translate into smaller families. Both my kids move off the farm, or both are unable to get jobs as steelworkers. Even worse, the steelworker may have move to Atlanta (seriously, the Falcons?) to find a new job as …. very much not a steelworker. The anti-Malthusian response causes living standards to rise rapidly, but also causes the disruptions of that to happen in years, not generations.

As Polanyi speculated, threatened identities will try to defend themselves. This might be through unions or tariffs or some other attempt to slow down and moderate the effects of the pure market; all gumming up the market. In the worst cases that defense ends up turning into things like anti-immigrant attitudes, racism, or as Polanyi suggests, fascism. These are all things dominated by the concept of maintaining some kind of identity. Those ugly versions of identity defense often explicitly deal with questions of fertility and population growth, by the way. They get hung up on things like race replacement or racial purity and almost invariably promote large families as some kind of ideal. Population growth is the visible confirmation your identity is thriving.

Both poles of the ideological spectrum that STU uses to interpret the long 20th century are, in this framing, ideologies that exist because of the anti-Malthusian population resposne. Or maybe exist is a strong word. Some sort of Hayek and some sort of Polanyi were probably bound to occur. But without the anti-Malthusian population response, there would have been less for them to argue about. The flip to family size declining with prosperity generated conditions where their ideas were more salient, to the point that DeLong could usefully and capably view much of the long 20th century through their eyes.

Take this whole section for what it is, pure speculation. You should go read Slouching Toward Utopia and learn something about how the economy evolved since 1870. While you are doing that, keep in the back of your head the population responses required to generate the economic growth observed.