Robert VerBruggen is off this week, so I’m filling in for today’s “Concrete Evidence” column. I’m Dario Marino, Manhattan Institute’s new Economic Policy Fellow. Let’s dive in.
It is widely known that fertility rates have fallen off a cliff throughout much of the rich world. Less well known is that they are also plunging in low- and middle-income countries like Thailand, Colombia, and Tunisia—places that were supposed to be decades away from this problem.
A new paper by Jesús Fernández-Villaverde of the University of Pennsylvania and Patrick Norrick of Northwestern University argues that, as of 2026, humanity has probably slipped below replacement fertility—the level of births needed to keep the population steady over time. If they’re right, the world is entering a demographic era with no real historical precedent, and one that will bring major economic consequences that we have to start preparing for immediately.
That would be a first for our species—something that has never happened even during wars, harsh dictatorships, or pandemics. The global population will still keep growing for another 30 years or so, as the large cohorts of women born decades ago are still having children and older generations are still alive. But the authors think the global peak arrives around 2056, at roughly 9 billion. That’s nearly three decades earlier than the United Nation’s central projection.
The old story was that rich countries have few kids and poor countries have many. As poor countries develop, they’ll gradually converge toward the rich-country norm of around two kids per woman. Fertility and income, in other words, were believed to be negatively correlated.
That relationship has flipped—not because rich countries are having more children than expected, but because middle-income countries are having fewer than expected. Mexico’s total fertility rate in 2025 was about 1.51, below that of non-Hispanic white Americans, despite its GDP per capita being a fraction of the United States’. Tunisia hit 1.45 with an income level roughly 18 percent of the U.S. In 2025, Thailand was at 0.87, Colombia at 1.01, Iran at 1.35, and Brazil at 1.52. These aren’t rich, aging societies like Japan, yet they’re approaching Korean-level fertility rates while still earning developing-country wages.
To draw a general lesson from the available date, Fernández-Villaverde and Norrick fit a statistical model to 236 countries going back to 1950. The model separates a common global force from each country’s own trend. The common force of worldwide fertility, it turns out, peaked in 1978 and has changed little over the past 15 years.
What is actually driving fertility down is not country specific. Of the 236 country trends, 219 point downward. The average country loses about 0.062 births per woman each year, and not one of the declining trends shows signs of leveling off. The fastest declines occur in countries near the bottom of the U-shaped relationship the authors identify. Thailand, China, Tunisia, Algeria, and Iran, which at some point had fertility rates between 3 and 4.5, are now losing roughly 0.13 to 0.15 births per woman per year.
This is what the authors call the “unpleasant integer arithmetic of replacement fertility.” You can’t have 1.8 children, so moving a society from a relatively healthy fertility rate of 1.8 to a very low 1.3 does not require a dramatic collapse in family life. A woman with two kids “may easily switch to one, or from one to zero, in response to relatively small changes in relative prices or opportunity costs,” the authors write. In other words, this means that a large decline in the fertility rate can result from a relatively modest share of women changing behavior.1
The authors’ main explanation for the trend is modernity, which they say “makes a third child expensive and childlessness cheap,” even in countries that aren’t rich yet. Children need years of costly investment in the modern world, and their economic usefulness within the household has diminished. Moreover, modern society “imposes next to no penalty on those who have no children,” while pensions and health care do much of what kids once did for aging parents. “In short,” they write, “modernity in its current form seems largely incompatible with replacement-level fertility.”
This is, however, a difficult theory to test because there are not many places where modernity has not happened. And in the few places where it arguably hasn’t, such as North Korea, we actually see a decline there too. North Korea’s fertility rate declined from 4 in 1967 to 1.77 today.
The rise of smartphones has become a fashionable explanation, and there’s some sophisticated causal work behind the claim. But Fernández-Villaverde and Norrick’s reading is that smartphones accelerated the decline rather than caused it. Italy’s fertility rate, after all, was already 1.19 in 1995 and stood at 1.14 in 2025. Where phones matter, they argue, is in spreading new norms faster.
The gloomy implication of the paper is that there is no obvious self-correcting market mechanism. The replacement rate has no special status as a natural societal resting point. A society can settle well below it and simply stay there. If Fernández-Villaverde and Norrick are right, sub-replacement fertility may simply be the new normal, and we’d better get good at managing a world with fewer people in it—for example, by reforming Social Security and Medicare before it’s too late.
Other Works of Note
How important are big firms to economic growth? Very.
Crime casts a long economic shadow: exposure to crime spikes can reshape political priorities and depress GDP decades later. Still waiting for the positive-effects paper on El Salvador.
“Fortunate Sons”: Members of Congress have long come disproportionately from economically elite families, and Progressive Era reforms didn’t change that much.
From a top econ journal: When Brazilian public hospitals outsourced management to private organizations, efficiency improved without sacrificing access or quality—and mortality fell.
On the Democratic Party’s shift away from “predistribution” policies like unions and minimum wages, and toward policies favored by educated donors. The rise of the New Democrats helped drive the partisan realignment by education.
On the importance of good managers. And many firms remain surprisingly slow to adopt the most advanced technologies.
Do children derail women from the CEO track?
An optimistic AI paper: The technology raises average wages 21 percent and reduces inequality by making difficult tasks easier for everyone.
What motivates people to become—and remain—police officers?
On Donald Trump tariffs: The AI investment boom may have offset what would have been a 10 percent drop in imports and a 0.7 percent economic contraction, while targeting consumption goods more heavily than capital goods reduced the output hit at the cost of higher inflation.
“What Women See in Men and Vice Versa”: Some apparent changes in marriage preferences over the twentieth century may actually reflect changing demographics.
Why did the long-running rise in U.S. test scores suddenly stop in the mid-2010s? Two suspects: weaker school accountability and the rise of smartphones, social media, and school-issued laptops.
What happens to marriage markets when dating moves online?
Real wages still haven’t caught up for everyone. A new Chicago Booth paper finds 37 percent of workers had lower inflation-adjusted pay in December 2024 than four years earlier.
A new urban economics paper finds that the concentration of high-skill workers in the big cities isn’t just rich versus poor: middle-skill workers are increasingly absent from high-skill metros too.
A paper in a top economics journal defends the Muslim practice of veiling.
What explains the recent stagnation in the college wage premium?
Consider 100 women who collectively have 180 children—a completed fertility rate of 1.8. Suppose five women are childless, 75 have at least two children, and 20 have three or four. Now shift the cohort modestly: 26 women are childless, 23 have one child, 46 have two, and five have three. Most women still behave much as before, and the median remains close to two children, yet fertility falls to 1.3. The drop from 1.8 to a very low fertility rate therefore does not require most families to change their behavior; it only requires a substantial minority to move from two children to one, or from one to zero. This demonstrates how relatively small changes in costs or incentives can produce large movements in the aggregate fertility rate.
Very interesting read. If the trend continues unabated, each new generation not replacing itself, the rate of population collapse should not be hard to plot, at least in the aggregate. I can’t help but wonder if societal issues like energy, environmental concerns and housing will ultimately solve themselves as the unaffordable today becomes abandoned in a world that is emptying itself out.
This is good for reducing strain on ecosystems and resources without too much malgenic selection. Low TFR in rich countries and higher in middle-income ones would have meant a large drop in global IQs for no population benefit. Now if we could just get Africa's fertility down...