How much for a marginal baby?

My back of the envelope estimate of the cost effectiveness of subsidizing egg freezing

In my last post I suggested that if we're serious about reversing population decline we should consider subsidizing fertility extension for women by giving them access to cheap or free egg freezing at a young age. Women are having their first baby later in life¹, and one issue with this trend is that it leaves them less time to have multiple kids. Encouraging women to start earlier is one possible solution to that problem and extending their fertility window through planned egg freezing is another.

Economically incentivizing women to start having kids while they're in their 20s may be a necessary part of a strategy aimed at bringing US fertility rates back up to replacement levels, but my intuition is that it would be a very expensive one. There's been a multi-decade trend² of increasing fertility rates among women in their 30s and decreasing fertility rates among women in their 20s. And when you look at abortion ratios (# of abortions per 1000 live births) by age group you can see that women in their 30s are significantly less likely to end a pregnancy than are women in their 20s.

Data as of 2020. From Table 4 of: Kortsmit, Katherine, et al. "Abortion Surveillance—United States, 2020." MMWR Surveillance Summaries 71.10 (2022): 1. Access excel file used for plot.

While financial limitations may partially explain the trend towards later pregnancies, there are generally other factors at play. As I mentioned in my earlier piece, when I think about what prevented women I know from having kids in their 20s it was generally some combination of the following:

1. They hadn’t found a stable relationship yet

2. Hadn’t achieved financial stability

3. Were attempting to advance in their career in a way that was incompatible with having young kids (junior roles in most industries in North America come with the expectation that you’re ready and able to work late etc.)

4. They just didn’t want kids yet and were enjoying their extended adolescence

Rarely was it just about their financial situation and I think it would take a very big subsidy to overcome the other factors, it certainly would've for me. Most women don't want to start a family without having found a worthy partner, and for college educated young women in urban areas (~28% of all young women given 46% of women aged 25-34 have a bachelors degree and 60% of college grads live in large metros) you probably won't have any peers that are having babies in their 20s.

Career Impacts for Young Mothers

Beyond the social and relationship factors, the concern that having kids young will be negative for your career seems to be a rational one. I found this paper on The Labor Market Returns of Delaying Pregnancy after

Ruxandra Teslo

mentioned it in a recent note. The paper estimates how disruptive births are to a woman's career based on age, whether they're enrolled in education and whether the birth was planned.

They find that even when births are planned they have a much more significant impact on the earnings and career trajectories of younger (under 28) vs. older women. Looking at Table 6 you can see that planned births for women 28 and younger resulted in an average estimated earnings loss of 26% per year over the 6 years following birth, while for women 28 and older the impact was only 12%. They also found that planned pregnancies under the age of 28 resulted in a 27% lower chance that the woman would be working in a medium or high skill occupation 5 years later, whereas that reduction was only 12% for older mothers.

This suggests that because the older mothers have had more time to build career capital they're less likely to be thrown off of their desired career trajectory following child birth. Even if we ignore the social factors preventing women from having kids young, an incentive that would compensate for this lost income and lowered career trajectory would likely be very expensive. The study used data on women in Sweden, but I'd assume a similar pattern would hold for the US where my perception is that the expectations of long hours and ample flexibility from junior staff (neither of which jive well with having a new baby) are even more pervasive.

Table 6 from: Gallen, Yana, et al. "The labor market returns to delaying pregnancy." Available at SSRN 4554407 (2023).

Cost per marginal baby of a state-funded egg freezing program

With all that in mind, I wanted to take a shot at estimating how much a program that paid for young women to freeze their eggs might cost per marginal baby produced (i.e. babies that wouldn’t have been born without the program).

Let’s say the program had the following parameters:

• Program covers 100% of the costs of an egg freezing cycle for women between the ages of 28 and 33. I estimated this cost to be $17,000 based on personal experience but this is likely the higher end of the range given I did it in New York.

  • I suggest this age range as I think it's the sweet spot between women being young enough that the procedure hasn’t lost much of its effectiveness but not being so young that we have to pay for storage for excessive periods of time. For the first 5-7 years of the program we could also pay for older women to do the procedure since it wasn’t available to them before.

  • In my earlier post I had suggested including a 50% claw back provision if eggs are unused by 50 years old which could also include an exception for women who are unable to pay (based on either an income or wealth cut off that say ~20% of women meet). But I’m going to leave this out for now - it’s unclear how this would impact program participation and

    Andras L
    brought up some reasonable points in his comment around why this would be unhelpful.

  • I also considered having some provision where unused eggs would be placed into a donor egg pool unless the woman opted to pay back some percentage of the costs involved in the retrieval. I like this idea because it would also make IVF easier and cheaper for couples who weren’t able to use their own eggs, but again given the potential optics and ethics of this I’ll leave it out for now. If you’re interested in freezing eggs for free though, this program appears to offer eligible women a free egg retrieval cycle if they’re willing to donate half of the eggs retrieved.

• Program covers 50% of the annual storage costs (about $1000/year) until the woman is 40 years old after which point she can take over storage costs herself.

  • I think it would make sense to have the woman pay something in order to maintain her investment in the eggs. Undergoing the procedure is not totally trivial, so I’m not overly worried about women doing it “just because it’s free” but I think having to pay the $500 a year would keep the eggs on her mind.

  • In the case that the woman has lapsed her payment for more than 3 years her eggs will be automatically transferred to a donor pool.

• Program covers 100% of future egg fertilization and IVF implantation if the woman decides to use her frozen eggs. I estimate this cost at $13,000 based on this source, but this could be higher or lower especially depending on whether the woman opts to do preimplantation genetic testing before implantation.

  • I think it’s preferable to pay for frozen eggs rather than embryos even if the woman has a partner. This allows unused eggs to be added to a donor pool and also means that the eggs are still useful to the woman even if she breaks up with whoever her partner is at the time of the retrieval.

  • If a couple prefers to freeze embryos there would need to be some additional test such as the couple having been married for more than 5 years.

I'm going to assume that this program has no effect on the overall number of IVF implantations done using donor eggs per year for simplicity. If women are able to freeze their eggs for free while young there would likely be a significant decrease in the need for donor eggs in the future but that would be balanced by more women doing IVF with their own frozen eggs.

Below are my estimates for the various factors we need to consider to get the estimated cost per baby born through IVF with eggs frozen through such a program.

1. Utilization rate for frozen eggs - in my last post I estimated that utilization rates for women who have frozen their eggs (mostly at their own cost) are likely above 50%. If IVF were paid for by the state I'd expect that number to be slightly higher, let's say 60% rather than 50%. For women who wouldn't have frozen their eggs on their own we might assume the retrieval rate would be lower. So let's say that women who use the program but wouldn't have frozen their eggs otherwise have only a 30% retrieval rate.

2. Percentage increase in women who elect to do the procedure - right now there's about 25,000 egg freezing cycles done for fertility extension per year in the US and about 10,000 more where embryos rather than eggs are frozen. For now I’ll assume that having egg freezing available for free triples the number of procedures done per year and that couples who otherwise would’ve frozen embryos will freeze eggs instead (unless they meet the threshold required to qualify to freeze embryos). But over a longer horizon I think it could increase much more. Assuming the ~35,000 egg and embryo freezing cycles done per year are primarily performed for women in the 35-39 age group, that implies that about 1.7% of women will have frozen their eggs. Given that the mean age of mothers at first birth has increased to 27.5 years and that I think egg freezing is relevant for any women that plans to start having kids after 30, I'd say this could be relevant for as much as 25% of women if it became normalized and free.

3. Expected number of years that the eggs are in storage and state is paying part of costs - 9 years given that the program covers women 28-33 and pays for part of storage until they’re 40.

4. Chance of a live birth per egg freezing cycle when fertilization and implantation is attempted (given age restriction) - 60%, based on estimate from my earlier post about why it makes sense to freeze eggs while young.

Plugging all those assumptions into excel (see my work here) and hoping I didn't make any mistakes (if you're interested in this please check my work and/or suggest edits to the model!) I got a cost of $191,000 per marginal baby. This is really high, but also probably worth it. Robin Hanson suggests, given that the average national debt per American is $300-$730k, “We should thus be willing to pay up to these huge amounts up front to induce the birth and raising-to-adulthood of a single child who would then pay average levels of future taxes to repay this debt.” And since the marginal babies from this program would be born to older parents who are more likely to be college educated and therefore likely to make more money than average we can probably expect their babies to be more productive than the average American.

However, even if this sort of program were cost effective per marginal baby, with the assumptions I’ve made it would barely made a dent in overall fertility. If this program led to a tripling in the current number of women doing egg or embryo freezing (as assumed in the table above) we’d get an expected additional ~15,000 babies per year. And even if it led to a 15x increase, which would translate to roughly 25% of women doing egg or embryo freezing at some point we’d still only get an extra ~100,000 babies per year vs. the more than 3.5 million born per year currently, a 2.6% increase. To get the total fertility rate from where it is now, 1.66, to replacement level, 2.1, we’d need to increase the number of babies born per year by more like 25%. So, while I think we should have a program like this given that it seems likely to be cost effective on a per baby basis, the assumptions I’ve made imply that even with mass participation it would only solve about 10% of our fertility deficit.

Footnotes below:

1. Data sourced from the CDC National Vital Statistics Reports, and extracted from pdf format with the help of Chat GPT: Osterman, Michelle JK, et al. "Births: final data for 2021." (2023)., Martin, Joyce A., M.P.H. et al. "Births: final data for 2010." (2012). and Mathews, T. J., and Brady E. Hamilton. "Mean age of mother, 1970–2000." National vital statistics reports 51.1 (2002): 1-14. Access excel file for plot.

2. Data sources: NCHS - Birth Rates for Females by Age Group: United States (1940-2018), Osterman, Michelle JK, et al. "Births: final data for 2021." (2023), and Hamilton BE, Martin JA, Osterman MJK. Births: Provisional data for 2022. Vital Statistics Rapid Release; no 28. Hyattsville, MD: National Center for Health Statistics. (2023). Access excel file for plot.