Will Evolution Undo the Demographic Transition?

Let me run a crazy idea by you. Judging by the conversations I’ve had over the past week, it may help if you’ve had a couple of glasses of wine.

As you may already know, developed countries tend to go through a “demographic transition,” meaning a huge drop in fertility. There are various theories as to what exactly drives this phenomenon, most of which have their critics: Birth control and abortion allow us to satisfy our sexual desires without creating kids to take care of; modern medicine makes the children we do have less likely to die before adulthood, reducing the need to have more; kids no longer do very much work around the farm; women have fewer kids when they get more education and join the workforce; general abundance gives us many ways to amuse ourselves if we decide not to start a family at all; etc. 

But, bottom line, most people have fewer kids than they would have had hundreds of years ago, and some don’t have any.

Demographers generally assume that this process plays out and then stops. Fertility plummets, and then it stays low. But what if this is completely wrong?

The idea here is that changes to our environment have transformed what it takes to pass our genes to the next generation. Today’s environment presumably selects people who consciously want to have kids, for example—a high sex drive can no longer do the trick by itself—and people who don’t stop after one or two even if there are lots of fun or rewarding things they could be doing instead of chasing whiny toddlers. These pressures are different from the ones we were designed to face, many of the old pressures are gone, and there’s every sign that our genetics affect how “well” (in Darwinian terms) we handle the new situation.

Put bluntly, “have more kids” genes exist—whether they work by increasing maternal and paternal instincts, by boosting religiosity, by decreasing educational attainment, or through any number of other possible mechanisms—and they will spread going forward. A similar process could play out among cultures, via the growth of groups that succeed in pressuring their followers to multiply. Over the long run, fertility rates will bounce back. And the long run may come sooner than you think.

Sound bonkers? Well, a new paper by Jason Collins and Lionel Page in Evolution & Human Behavior spells out the math, starting with U.N. population projections and factoring in some basic principles and findings from quantitative genetics. (Blog post summary here.) It’s not proof that this will happen, as it’s permeated with strong assumptions. But it demonstrates that the idea is highly plausible.

The backbone of the paper is the “Breeder’s Equation”: the formula spelling out how much a trait will change, per generation, when it’s selected for. The more strongly genetic a trait is, and the stronger the selection pressures, the bigger the response will be. Specifically, the “narrow-sense heritability” (h) is squared and multiplied by the “selection differential” (S), giving us the response to selection (R).

How do we run this calculation in the case of fertility? The heritability is easy enough to estimate, on the basis of twin studies as well as more modern genetic techniques, with estimates ranging from about 0.2 to 0.4 (on a scale from 0 to 1)—so the authors go with 0.3. The selection differential, meanwhile, is “the difference between the mean fertility of the total population and the mean weighted fertility of the parents (equal to the mean number of children that each child has in their family).” Basically, when a big share of kids are in a small share of families, the selection differential is high. Unfortunately, the authors assume that the distribution of family size follows a certain statistical pattern based on each country’s overall fertility, rather than using the actual distributions, though there’s research showing this pattern (Poisson) is a good fit for the data.

In the U.N.’s models, fertility oscillates but never continuously rises once the demographic transition is over in a given country. When Collins and Page add the Breeder’s Equation to the mix, a lot changes:

[T]here are no substantial differences between the model forecasts for Africa, where most countries are yet to enter the low fertility state following the demographic transition. For Asia, there is a clear difference, with population remaining stable over the second half of the century rather than entering into decline. Europe sees the most dramatic turnaround with a predicted return to population growth instead of prolonged population decline. Our median estimates suggest a European population at around 800 million at the end of the century, near its current level, compared to a base model forecast of 660 million in 2100. North America's population also sees an increased rate of growth with a population estimated to be 150 million higher than in the base model in 2100, at 650 million.

As the authors concede, a major limitation here is that ignoring heritability may not be the only problem with the U.N.’s models. The simple fact that the demographic transition exists shows that environmental changes can happen far more quickly than evolutionary changes—and if new environmental changes push fertility down even further, they could cancel out or even outweigh the bounce-back brought by other forces. In the long run, we’ll probably be outsourcing pregnancy to robotic wombs and uploading our brains to the cloud, after all.

Till then, whether we’ll see a population rebound is an open question. That the modern world will change humanity? That seems guaranteed.

Robert VerBruggen is a deputy managing editor of National Review.