Children of men

The population of all countries, especially rich ones, is growing more slowly. It's not good news.

Sep 17, 2021

Imagen — Nurse holding triplets born in 1938. From Argentina’s National Archives via Twitter.

It's not just odd failure to mention Biden tax credit. Within Europe, several, you know, actual democracies have lavish pro-natalist policies, France in particular 1/

Dave Weigel @daveweigel

Interesting speech. Vance praised a policy by Viktor Orban that pays parents who have multiple children - "why can't we do that here?" - but didn't mention the Democrats' child tax credit.

A while ago, Nobel Laureate Paul Krugman posted a Twitter thread about Senate candidate and author J.D. Vance praising the pro-natal policies of Hungarian Prime Minister Viktor Orban. Are those policies actually good - and is their goal of increasing natality a good one? It should be duly noted, and ignored, that Orban’s policies are less “universal childcare” and more “awarding married straight mothers of ethnically pure Hungarian children a Hero Of the Fatherland medal after their fifth live birth”.

A fact to kick off the post: the population of the world’s richest countries is growing more and more slowly, and has been for a while. Some of this is due to population growth in developing countries slowing down (which is good), but the same process didn’t just happen rich nations, it went into overdrive. In fact, some countries have aged so poorly that they will decrease in population altogether, such as Japan, Europe, and paradoxically enough, China.

Mayank Seksaria @MayankSeksaria

Incredible inflection point

Is this good or bad? The movie (and book) “Children of Men” takes the latter position, arguing that the world will become a hellish vortex of strife and poverty, one that’s devoid of hope. To quote from the book:

Man is diminished if he lives without the knowledge of his past; without hope of a future he becomes a beast. We see in every country in the world the loss of that hope, the end of science and invention, except for discoveries which may extend life or add to its comfort and pleasure, the end of our care for the physical world and our planet. What does it matter what turds we leave behind as legacies of our brief disruptive tenancy?

Why nations age

This isn’t a unique phenomenon; it’s, in fact, a common trend throughout history. IThe reason why is simple: the demographic transition. In the distant past, people lived shorter lives, in worse health: dying was more common, especially in the first years of life. With mortality high, particularly child mortality, parents had many more children because they wanted to have as many as possible survive into adulthood. These high mortality rates, plus low life expectancy even for adults (war, disease, and famine were and in some places still are facts of life), meant that the population was stagnant.

Improvements in health and sanitation starting in the 19th century led to declining mortality, particularly of children but across all ages, which expanded lifespans. Since parents had plenty of children even after the new technologies expanded lifespans, populations boomed, only for adults to progressively reduce the number of children they had and for the elderly to replace children more and more as the source of care expenditures.

This correlates heavily with economic development, because not just population was stagnant in the pre-modern era, GDP was too. I’ve written about this period, known as the Malthusian Economy, and the gist of it is that, in premodern societies, economic growth doesn’t create more wealth, just more people. Limited supply of the key productive factor, land, meant that only so much output could be coerced out of the economy at any given time - and because most production was agricultural, productivity increases just resulted in more population that brought down living standards. This cycle was broken by both the development of non-agricultural activities at a significant scale and substantial improvements to productivity.

Regardless of the Malthusian thing, population first booms then slows down and ages. An aging population, with fewer and fewer children by generation, means that the growth rate of the population continues to get slower and slower. This process happened in now developed nations decades ago, is currently well underway in Latin America and Asia, and is burgeoning in Africa; this isn’t a result of cultural norms or of specific technologies, but rather it seems to be a verifiable fact. The momentum of population growth will end given more widespread access to contraception, more female labor participation, and more education to both genders.

Ideas about ideas

Now, here comes the tricky part. As mentioned in the brief Malthus sidenote above, the link between economic growth and population growth is a two-way street. The post I already linked goes into one of those ways: how economic growth changes population growth. Now for the second: how does population growth move the economy?

The first key insight is that economic growth is about ideas. Technology and productivity, the key drivers of economic growth over the long term, can be understood as just a way to combine bundles of labor and goods together to make other goods, ones that people consume to improve their lives. Ideas have increasing returns: if you have ten inputs, the combination of them is 10^10, or 10,000,000,000 (ten billion); but if you have fifteen inputs, the combinations grow to 10^15, or 1,000,000,000,000,000 (one quadrillion). Adding 50% more inputs gets you thousands of times more ideas, even if some are, to quote George Akerlof “like chicken ice cream”. Ideas are also non-rivalrous, meaning that once you come up with them, anyone can use them without having to come up with it again. If you create a recipe for a new sauce, the next person doesn’t have to fiddle around with tomatoes and onion to start from scratch- they can just buy your recipe book and use the recipe. Finally, ideas have spillovers; if you come up with something, and bounce it off me, I can suggest changes or improvements.

From the beginning of human history, all the way millions of years ago, to the present, coming up with new ideas at roughly the right times has moved humanity forward. That’s also why the two-way street is important: if a caveman made a really cool tool, they could only share it with their tribe and neighbouring ones. Same for Greek inventions (how could they reach India or Portugal?), or for Maya wisdom that never made it past the jungle, not to say the ocean. A bigger, more connected population makes sharing ideas easier. If each person has a fixed number of ideas, more people means more ideas. But because ideas are non-rivalrous, have increasing returns, and have spillovers, then more people come up with more and more ideas. So population growth speeds up idea creation, and idea creation speeds up economic growth. This is known as a scale effect, and it’s a key conclusion of endogenous growth models - with which endogenous growth theorists are super uncomfortable, due to the “utter lack of evidence” issue. Scale effects look super dicey until you consider longer timescales (like, eons and not the 50 years we actually have data for) or that knowledge spillover areas are not national borders - Europe is clearly one of them, despite being many tiny countries, and not all of China or India might be one.

Which is why population growth slowing down is really scary. If you have suddenly fewer and fewer ideas, and once the “limits of globalization” are reached, then the number of ideas will plateau - or worse, come down. So far we’ve assuming that the “cost” of inventing/discovering things is constant, but that’s not a fixed law of nature. Maybe scale effects apply too, or research and development (R&D) has increasing returns and/or spillovers as well.

Change in research productivity and in number of researchers for a section of US firms. Most blue bars being on the left means that most firms got *less* effective. From Bloom, Jones, Van Reenen, and Webb (2020).

Problem they don’t and it doesn’t. If you look at plenty of fields (computer science, agriculture, healthcare, AI), the more advances you make in one direction, the less each additional unit of money or each additional researcher adds to the problem - decreasing, not increasing, returns. That’s really bad. The logic here is simple: inventing things is mostly about solving problems with the things. Solving problems requires knowing about the things. We already know a lot of things, so each problems needs much more knowledge to solve - both through more and more insular specialization of professionals, and through larger and larger multidisciplinary teams. Since the time of Adam Smith, we know that assembling huge crews to solve problems makes everything more burdensome and everyone less effective, so you hit the same brick wall eventually.

Older people, older firms, lower growth

Now, knowing what we know about ideas and growth, how does aging and declining population growth play into this?. If global population growth goes way down, then what happens?

Well, because productivity drives growth over the long run, and technology drives productivity, then you can get two sources of per capita GDP growth. The first is the one we’ve had so far: population grows slower than knowledge, so we can get richer and richer because the stock of technology per head gets bigger and each person can “use it” to squeeze more labor from their skills. But there’s, mathematically, another way: population stops growing, or declines, and technology stays the same. In this world, everyone gets richer, but not because they get the same slice of a bigger pie, but because there’s fewer slices of the same pie. If you have an economy that looks a lot like ours, using the concepts above, there’s two “endings”: one where population keeps growing and we throw infinite numbers of bodies at problems, and one where we stop discovering things but the dwindling insights of the human race are about how to reassign the capital we built for a population many times larger. Because growth in population and growth in output feed into each other, you either get one ending or the other - no in between.

One channel through which this can happen is the labor market. A lot of new things are invented by feisty new firms, and these firms need people with knowledge of new things, then you need a growing supply of labor - and of a certain kind: young, with new skills, and eager to learn. If you ever tried teaching your parents to do something on a computer, you know it’s a pain - imagine having the entire economy hinge on it. Because would be entrepreneurs see that there’s less and less people they could hire in the future, they don’t start companies, which means that the average company gets much older and much bigger too. And big old companies aren’t particularly innovative (I’ve written about this too!), so the cycle repeats once more.

Is slower growth even bad?

Economic growth is slowing down, and it’s because of aging, declining productivity, and lower birth rates. But we already said that if the population stops growing, or declines, people will still get richer, since the size of the pie (built on knowledge that wouldn’t be lost) would not decline, or would do so far less than population. We’re already suffering the effects of lower population growth, since economic growth slowed down from 2.25% to 1% in the US.

10 year average of economic growth and TFP. From Matt Clancy’s review of Vollrath’s book.

But here’s a contrarian position, by Dietrich Vollrath: it’s good. Most of that decline (0.8pp) are attributed to declining creation of human capital, mostly due to workers getting older, there being fewer of them, people working far less (over their lifetimes), and stagnant educational attainment. And the decline in productivity was far smaller than the decline in growth - it could only account for some, not all, of the remaining 0.45pp left over. A bunch more of the decline, about 0.25pp, have to do with declines in labor mobility between firms and between places, due to labor market concentration and restrictions on the housing supply.

So that leaves 0.2pp of the slowdown to pure TFP - a residual to be explained by the residual, if you’re feeling punny. This isn’t particularly hard either: it’s because people started spending far less on goods (which are made by the very easily productive manufacturing sector) and far more on services, whose productivity is harder to measure but generally lower. The remaining 0.2pp of the slowdown in growth can be pinned down on low productivity services replacing high productivity goods as the main thing people spend money on. This is caused by because they’re constrained by an actual physical limit (time) and, because they are primarily a people thing, there’s a certain rivalry between services. Plus, services are affected by Baumol’s Cost Disease: less productive sectors of the economy get more expensive over time anyways because, since their main expenditure is wages, they need to compete with the productive sectors for labor - meaning that they have to raise wages as much as the places where wages are actually going up.

Here’s why Vollrath argues this isn’t bad: outside of maybe educational attainment, all those are signs of success. People have fewer kids because they have more power and autonomy over themselves, versus social rules. The population is older because, besides lower mortality, the elderly can live decades longer than they used to - indubitably good. Lower mobility and bad firm demographics are bad, though, but people also have shorter working lives. Finally, goods versus services: it’s not good that services have lower productivity growth, but it is good that we’re spending more of them - a sign that our basic needs (food, clothing, shelter) are increasingly met and we can more up through Maslow’s Hierarchy, so to speak.

In addition, old people save more - the Baby Boomers, one of the largest generations in US history, have behemoth savings as a share of the total. That has actually caused a decline in the natural interest rate, since the boomers save so much, which has made monetary policy less effective and central banking a more hopeless endeavor against phenomena like secular stagnation, i.e. prolonged periods of low growth, high unemployment, and low inflation. But this has a positive counterpart too: borrowing gets cheaper, so certain types of investment should be more plentiful. However, this decline in savings and interest rates has positive effects over the long term, because it increases some kinds of investments, but bad ones over the long term, due to the investments that are increasing aren’t especially “productive” ones.

So, summing up, lower growth is bad (citation needed), then it could be caused by good reasons, and there might be a silver lining of more abundant investment in certain projects. I’m still not convinced that it’s good, though.

Can we do anything?

Of course, if slower and slower population growth would lead to bad outcomes, then the obvious solution is growing the population faster. But let’s circle back on that, because there’s a whole more lot you can do first.

Firstly, and most obviously, you can get as much “juice” from the current population as possible. We already mentioned restrictions to mobility, which are already having enormous costs to the economy - so you can claw back most of those losses, over the medium term, by building more houses. There’s also the issue of market concentration, which you could tackle in ways that promote competition and innovation in the economy.

There’s also stagnant educational attainment: you can both make attending higher education easier and steer students towards fields that have the highest impact on growth, especially since there’s some evidence pointing you can kind of “brainwash” students into pursuing degrees in specific fields. Speaking of, STEM tends to draw disproportionately from one group: men. So building STEM-track careers for women could do wonders - and lift growth as much as 0.3pp per year.

There are also a very “out there” solution: artificial intelligence. If machines as smart as human researchers can be invented, then you don’t actually need to grow the population to continue growing. Because these machines can do at least as much work as a human scientist, then you could just counteract some of the process, or even reverse it altogether. Once the speed at which tasks are automated matches the rate at which ideas get harder to find, then you reach a singularity and growth becomes kind of infinite.

And now the easy, obvious ones: immigration and pro-natal policies. Immigration, as I’ve written before, isn’t just not bad, it’s actually good, and it’s particularly important that immigrants are both more entrepreneurial and disproportionately likelier than natives to become inventors. Since immigrant birth rates converge with those of natives, taking in just one big wave wouldn’t work, you’d need a continuous flow of immigrants to keep the population growing.

Now, onto pro-natal policies, the issue that motivated me to write this post in the first place. The key insight here is that people frequently say that they want more children than they currently have: between 2.2 and 2.3 in the OECD, over the 2.1 required to keep the population stable and significantly higher than in most OECD members. The main reason people don’t do so is finances: housing costs, childcare costs, and other financial burdens posed by children discourage having them. Ergo, if you provide families with a series of economic benefits (maternity and paternity leave, child allowances, etc) then they’ll have as many children as they want, which is a better number than they currently have.

Conclusion

In the end, lower population growth is actually bad and does actually have consequences on the economy. I’m unsure if pro-natal policies actually work, particularly since the most successful examples were also incredibily expensive. But a combination of the other reforms (to housing, competition, etc) and promoting immigration, a case which I’ve made plenty of times, could do the trick. Also keep in mind overpopulation is a fake problem.