Are 40-Year-Old Women About to Start Having Tons of Kids?

What it would take for "completed cohort fertility" to hold steady for Gen Z

I have neglected this space for too long—though I hope you all have been checking out my weekly column on interesting research and data over at City Journal’s page.

Today I have a quick data exercise inspired by a Twitter back-and-forth I had with Nathaniel Bechhofer, reproducing and extending some AI-generated charts we exchanged.

The issue is fertility. The standard narrative is that things were going great for America in the mid-2000s—with total fertility right around the replacement rate of 2.1, having recovered from its lows in the immediate post-Pill period—before everything fell off a cliff with the Great Recession and never bounced back:

Not so fast, insist some skeptics!

The total fertility rate (TFR), they note, is kind of a weird statistic. To calculate it, you start with a single year’s birth data, and then add up the fertility rates of women at each age to come up with a total. It’s the total fertility of a hypothetical woman who lived her entire life inside that year, having the average number of kids (well, fraction of a kid) at each age.

TFR can especially be thrown off if women delay fertility. If, on New Year’s Day 2030, all twentysomethings simultaneously decide they’re not going to have any kids until they’re 30, TFR will dip down during the years they’re delaying, even if they go on to have the same number of kids in total. In other words, the hypothetical woman trapped in 2030 won’t have any kids in her 20s, but she also won’t have the births the real women delayed into their 30s (which don’t happen in 2030).

The alternative is to look at “completed cohort fertility,” the total number of kids women have had by the time they reach the end of the usual childbearing years. Right now, we older Millennials are the ones hitting that point (I’m 1984), and our completed fertility looks just fine, so what’s all the worry about?

I’m intensely skeptical of this line of argument, based largely on what happens when you break things down by age. As you can see here, (A) older Millennials spent their 20s in years when twentysomething fertility was doing okay (basically the 2000s, when either the Great Recession hadn’t happened yet or the big decline was only starting), so you wouldn’t expect us to be hit that hard; and (B) since then, fertility has been declining in the prime childbearing years way, way more than it’s been increasing for older women. (Data from Table 2 here.)

As Bechhofer pointed out to me, it can also be useful to map these data to cohorts, which is pretty easy to approximate from the same data. For example, 22 is the midpoint of the 20-24 age group, and people born in 1990 turned 22 in 2012, so I treat the 2012 rate for 20-24-year-olds as the 1990 cohort’s rate in that age range.

Your mileage may vary, I suppose, but to me this shows that while there was a healthy mix of increases and decreases for cohorts from the mid-’70s up to about 1985—i.e., these cohorts each lived through a mix of higher- and lower-fertility periods at different ages—things look brutal for Gen Z. Twentysomething (and, fortunately, teen) fertility is way down. Even the 30-34 group stopped rising with those born in the mid-’80s! We’re left hoping for a fertility explosion among those 35-44, or perhaps even above that, where women have historically had extremely few kids.

When you sum up the numbers vertically for each cohort, multiply by five (as these are five-year bins), and divide by 1,000 (as these are per 1,000 women), you can approximate the completed fertility for the cohort, which came to a bit above 2.1 for us Oregon Trail fans. As an exercise, I asked the question: If we start with the fertility rates we know, and assume flat lines after that for those below age 35 (a conservative assumption because these are actually trending down), what has to happen with the two older age groups to hold completed fertility steady at the value of the 1981 cohort, the last the CDC table provides full data for?

Here are the results, starting with the 1978 cohort. When both the 35-39 and the 40-44 bins are missing, I had them increase over the previous year by the same proportion as needed to keep completed fertility even.

My intemperate hot take is that you’re delusional if you think anything like this is going to happen. TFR is indeed a quirky measure and might exaggerate the final extent of the problem, but it’s not hallucinating the problem entirely. Fertility is going down and 40-year-old women are not about to start popping out tons of kids to stop it.

My more considered take is that while I do think it highly unlikely, two possibilities are worth stressing:

• If infertility treatments improve drastically and there are enough older couples who don’t mind the thought of having teenagers in their 60s, that could totally change the math.

• Gen Z is highly susceptible to social contagions. Maybe they’ll all decide they have baby fever at the same time. I’m not even half joking when I say this.

Help us, Taylor Swift!