Egg Freezing When Young Isn't A Good Investment

The Math Doesn't Work

Many people, such as Ruxandra Teslo , argue that reproductive technology can meaningfully boost fertility in the future, but that the reason future reprotech will do this when past reprotech didn’t is because egg freezing at younger ages increases IVF success.

I am skeptical of this argument, and in this post, I’m going to convince you that you should be skeptical of it too, basically because the reprotech advocates are not incorporating the actual costs into their calculations.

When we actually do the math on success rates at various ages and real costs at various ages, egg freezing early in life simply doesn’t make a lot of economic sense except for a subset of women with very high confidence that they will not want kids until late in life.

---

A History of Older Fertility

Before we go on, let’s briefly review the history of older fertility. The graph below shows TFR35+, i.e. the amount of the TFR calculation occurring from women 35+ over time, in the United States and some other countries for which we have long-term data.

What you can see is that between 1850-1880, a decline in later-life fertility kicked off basically everywhere. This is what, in demographic parlance, we tend to call “stopping behavior.” People started to show evidence of a behavior where they intentionally stopped having babies having reached some number. You can also, as early as 1850, see a pretty clear distinction between the Anglo countries and the Nordic countries. I’m not totally sure what explains that difference, but it’s definitely interesting!

But what you should notice is that late-in-life fertility is… not high today.

The reason for this is not that women today lost the ability to reproduce. It’s very simply stopping behavior. Modern families tend to want smaller family sizes than you’d get if you had nonstop noncontracepting sex throughout a typical marriage. They often hit their goal before older ages, and so “stop.”

But okay, let’s steelman, and ask: what if we look only at married childless women? These are women who likely would be interested in having more children, and make up a huge share of IVF users. So we want to ask: for married childless women 35+, i.e. women likely to be open to having kids, what has happened to their MASFRs?

As you can see, they remain below Baby-Boom highs, and far below the levels observed in 1880-1920. Now, this isn’t a perfect measure; for that we want fertility rates for noncontracepting married women. How have those changed over time?

That’s a pain in the butt to calculate in lots of periods so we’ll look at it in just two periods: a survey in 1955 and a survey in 2022-2023:

As you can see, in 2022-2023, the fertility rates of noncontracepting married women ages 35+ are already higher than they were in 1955, i.e. the Baby Boom. But we are nowhere near the pre-contraception levels even for noncontracepting married women!

Why does this matter? Well, it tells us two things:

1. To the extent that the 1955 GAF tells us the limits of late-life fertility before reprotech but after the emergence of widespread family-size-targeting, American women today have already blasted straight through biological “limits” in that social regime, likely due to IVF.

2. But it’s blindingly obvious that even among noncontracepting women ages 35+ that stopping behavior remains overwhelmingly decisive. Married women 35+ not using contraception today must either be having less sex, or be much less healthy, or using a non-contraceptive method to avoid conception (rhythm, withdrawal, nonreproductive sex), or else all 19th century demographic data must be hilariously wrong, and I don’t think it’s that last one. The latent health indicator is interesting: life expectancy at 35 is WAY higher than it was in 1870, but also obesity is higher today, and positive selection on robust health was stronger in the past. So maybe the modern health environment yields less fecund women at ages 35-50? But that change happened 100% between women born 1825 and women born in 1900. I suspect “less sex” and “avoiding conception through other means” are the big drivers here.

Okay, this all matters for understanding reprotech because it helps us figure out where we stand: we are nowhere near the biological limits if every married noncontracepting woman ages 35-50 had low general morbidity (esp. obesity) and was having frequent penile-vaginal sex with male ejaculation in the vagina. Kinda gross to spell it out that way, but it’s important to have a clear idea of what trends might explain the bizarre fact of “In 1870, married noncontracepting women had more than DOUBLE the fertility rates of married noncontracepting women today have even with the help of IVF.”

To the extent we are “near biological limits,” it’s because we’re talking about biological limits within the context of widespread intentional stopping behavior, or latent health/sexual/marital behaviors adjacent to the emergence of that behavior. Demographically, we don’t know exactly what’s going on here, but that’s the space of explanation.

So now we want to ask: given that we’re already smashing biological limits that have existed since at least 1955 and probably since the 1930s, how much more should we expect something like widespread egg freezing to accomplish?

The answer, after the paywall, will require some math.

---

The Math

SO LET’S DO SOME MATH!

Charting out the entire possibility space on IVF accounting for differential egg freezing date and differential attempt at conception date is a huge pain in the butt. But I did do all the math. We’ll take some simplified cases!

Imagine you’re a 24-year-old women considering egg freezing. Should you?

Provided you end up using the eggs, here’s a plausible success rate for you (I am indebted to Ruxandra Teslo’s excellent work for the empirical estimates underlying several key parts of this model):