Differences in fertility rates compound over time. A total fertility rate (TFR) of 1.8 might not look too different from 1.4, but it has drastically different implications. Let me illustrate the argument. I take two countries with 10 million inhabitants in 2026, each with the same age structure and life expectancy as the U.S. and no migration. The only difference is that country A has a TFR of 1.8 and country B has a TFR of 1.4. Then, I simulate both populations forward for 100 years. I run a full simulation, including age-specific fertility and mortality, the sex ratio at birth, etc., but I skip the details to save space. In 2126, country A, with a TFR of 1.8, has a population of 6.8 million, down 32%, and still falling by 0.5% per year. Country B, with a TFR of 1.4, has a population of 3.9 million, still falling by 1.3% per year. Country B’s decline is much sharper, 61%, and more will come in the following decades: country A has 71,000 births a year, while country B has only 27,000. But the real difference is the ratio of the population aged 65 and over to the population aged 20-64: 45 per 100 in country A vs. 60 per 100 in country B. Country A faces a challenging problem for social security and old-age health provision. Country B has a terrible problem. That is why I find the argument “pro-natalist policies cannot get us back to the replacement rate” so weak. Even pro-natalist policies with modest effects, which increase the TFR from 1.4 to 1.8, make a very large difference in just 100 years. This is, by the way, a common fallacy when we assess policies: the fact that policies do not succeed 100% does not mean we should not adopt them. We have flight regulations that have causally lowered death rates in commercial aviation, but we still have plane accidents. The point is to assess whether the policy passes a reasonable cost-benefit analysis. And I believe, after having read all the literature, that well-designed pro-natalist policies that lower the opportunity cost of having children (such as better access to large houses in medium-density locations and generous parental leave) do pass the test. I plan to write much more about this over the next few weeks, as I am preparing a literature review on the issue.