Projections Show That Increasing the Birth Rate (As if You Could) Is the Wrong Way to Get More Workers

Some actual populations projections below, after a little commentary [and a new set of comparisons added at the end]

There is no evidence that any proposed policy would actually raise US birth rates. Some massively expensive (and highly valuable) policies might make a little difference — like subsidizing childcare, healthcare, housing, and so on. But my reading of the evidence from many countries (here’s the thread of posts) is that at any remotely likely scale such policies would do little more than slow the ongoing decline.

On the other hand, if you actually could increase birth rates, you would produce extra very-expensive children who wouldn’t be able to get to work in the AI dungeons for many years. I did a little simulation of this last year, which I described in my essay in The Hill:

The Urban Institute estimates that U.S. governments (at all levels) spend an average of $17,660 per child per year. At that rate, if we somehow devised a policy to increase birth rates for one year by 10 percent, it would cost some $115 billion to raise those 360,000 additional children up to age 18. That’s before we start spending on the higher education needed to make them highly productive workers — and does not even include whatever spending the fertility-promoting policy itself requires. And it’s just one year of extra babies, who won’t be available to take care of their grandparents for a quarter century or more.

Projections

Demographic projections are very complicated and tricky to perform, and I am not expert in them. Fortunately, very simple projections can be effective for explaining how more complex things work. That’s what I’m doing here.

I will start with the 2023 United States population and what we know happened that year in terms of population, births, migration, and deaths. Then I’ll show where it is leading. That sets us up to alter a few parameters and show what difference it would make.

Based on numbers reported by the Census Bureau and National Center for Health Statistics for 2023, there were 337 million people in the USA. That population had a total fertility rate of 1.61 — meaning, if nothing changed, the women turning 15 that year would have 1.61 children each, on average. In terms of immigration, we gained about 850,000 people (subtracting the few who moved out), about half of whom were in the ages 18-32. Life expectancy was 76 for men and 81 for women. Those are summary numbers, averaged over every age and sex combination. If we plug the full set of numbers into the Census Bureau’s Demographic Analysis and Population Projection System (the free DAPPS app), we can see what would happen through, say, the year 2100 if nothing changes. I’ll start with what we can call that the status quo model, then add some variants.

In the figure below I show three relevant targets for our current debates about birth rates: (a) The total population, (b) the number of working-age people (25-64) per retirement-age person (65+), and (c) the percentage of the total population that is in the working-ages of 25-64.

In this status quo scenario, the population peaks at 349 million in 2041. The number of working-age people per retiree falls rapidly from 2.9 to 2.4 as the Baby Boomers retire in the next 15 years or so, and then continues to decline to 2.0 by 2085. This is what a lot of people are worried about with regard to old-age support — workers per retiree is falling, so we’ll need to spend a bigger share of income on care for them. But the last panel shows that the overall percentage of the population in working ages doesn’t change very much, staying between 50% and 52% until the 2070s. That is because the falling number of children is canceling out the rising number of old people, so our total working age share is staying pretty constant. We have an aging investment challenge, but our low birth rate is helping to keep it from being worse.

Now let’s add two more scenarios, which we can call Ridiculous Nativism (banning immigration immediately) and Miraculous Pronatalism (increasing the total fertility rate 10% by 2030). Stopping immigration is extreme but not as extreme as it seemed a few years ago — lots of MAGAs want this now. And boosting the total fertility rate by 10% would be a virtually unprecedented pronatalist success but might be possible somehow.

What happens to our three measures now?

With regard to total population, the immigration effect is huge. No immigration would reduce the US population by 88 million, or 29%, by 2100. In contrast, boosting the total fertility rate to 1.77 by 2030 would end up increasing the population by just 12%, or 38 million. Note that in this simple projection new immigrants have the same birth rates as the existing population, and they keep arriving at the same ages they did in 2023, which was pretty young, so they are adding a lot of births as well as adult bodies.

What happens to our economic scenarios? The middle panel shows the working-age to retiree-age ratio. Without immigrants, this falls from 2.9 all the way down to 1.7, and doesn’t stop falling till 2085. It eventually starts to rise a little. Clearly, cutting off immigration would make it much harder to pay for retirement — choking off a vital supply of labor. But what good would increasing birth rates do? It would have no absolutely no effect until after 2050, when those extra babies would start reaching age 25. Eventually, by 2100 the extra births would increase that ratio by .17 as the babies became workers. That’s much smaller than the effect of cutting off immigration, which reduces the ratio by .29 in 2100. So to address the number of workers per retiree, having more babies is quite a slog! It takes decades, costs a fortune, and still does less than immigration (which is cheaper than free, productivity-wise).

The last panel helps explain why boosting birth rates would be problematic, at least for this century. Adding more babies reduces the share of working-age people in the population until 2100 and beyond! It compounds the dependency problem instead of fixing it. Cutting off immigration would be even worse, though, reducing the working-age share by 1.6 points at the end of the century.

My main point is for readers who think that increasing births is essential, or even advantageous, for addressing the fiscal challenges of our aging society. It won’t help. As a solution it takes too long, and throws up huge expenses as it goes — children are expensive, too. It might be good for other reasons, but not for increasing the share of workers in the population.

Note that these are projections, not predictions. What actually happens is determined by many other things. I use projections like this to show not what will happen, but the shape of the future implied by the present — and by proposed alternatives. That’s a powerful tool, as long as you keep it in perspective.

Next day addendum: A different way to look at scenarios

I thought of a more intuitive way to describe and compare the immigration and fertility boosting effects.

Here are three projections. In the baseline (the gray dotted lines), I hold the 2023 population age and sex distribution, fertility rates, and mortality rates fixed — with no immigration. Then, compare two possible approaches — maintaining immigration or increasing fertility. In the second projection (blue lines), I increase the age-specific fertility rates by 10% between 2023 and 2030 and then hold them fixed (at a total fertility rate of 1.765). In the third (brown lines), I hold fertility at the current level (1.61) and let the 2023 immigration level continue every year (about 850,000 net migrants).

Current immigration levels generate much greater population than increasing fertility 10 percent would — 308 million versus 252 million by the end of the century, a difference of 56 million. The increased fertility scenario population peaks in 2032 at 339 million, while the immigration scenario rises to 349 million and peaks eight years later, in 2040. The immigration scenario has a higher ratio of working-age to retirement-age population ratio, especially through the middle of the century, before the extra babies start entering working ages. But even at the end of the century the immigrants are generating more potential workers per retiree, even at current fertility rates. Finally the immigration scenario has the highest population share age 25-64 throughout the century, because it has a steady stream of working-age adults, while the fertility-increase scenario pours extra dependents into the population, increasing the burden on working-age adults.

You could imagine a bigger fertility increase and get closer to the immigration numbers, which I don’t find plausible — falling fertility is more likely, based on the experience elsewhere. But no matter how much you boost fertility you will always have to raise those kids before they can get to work. And of course you could increase the immigration numbers, too, instead of keeping them at 2023 levels. But I think these projections give us a reasonable sense of the scale, type, and timing of impact these approaches imply.

This is work in progress. I haven’t figured out how to share the work product for this, and I might have made any number of errors. Open to all feedback, and happy to share materials.

—

Previously Published on familyinequality with Creative Commons License

***