The prologue to The Population Bomb (1968), a piece of neo-Malthusian apocalyptica written by Stanford biologist Paul Ehrlich and his (uncredited) wife, Anne, featured a prediction that “the battle to feed all of humanity is over. In the 1970s and 1980s hundreds of millions of people will starve to death.”

Extrapolation is useful as a predictive tool but useless on its own. Trapped in the dead end of finite-resource environmentalism, the Ehrlichs underestimated human ingenuity and human adaptability. Their message, which attracted a vast audience, turned out to be nonsense. It was not the first prophecy of doom to do so, nor will it be the last. And several panics later, depopulation is the new overpopulation. And just as the Ehrlichs had the numbers to back up their case, so do their mirror images.

Sinking birth rates have gone global. Almost all the high-fertility holdouts, mainly but not exclusively in sub-Saharan Africa, are now going down the same path. Birth rates are at present below replacement levels in nations that are home to over 70 percent of the world’s people. In some developed countries, the population is already declining. More are destined to follow suit. We are still some distance from peak human, but that milestone may be closer than it appears. The U.N. expects our ranks to top out at 10.3 billion in the mid-2080s and then fall after 2100, a forecast that could well understate the speed and scale of the transformation that is underway.

In the U.S., however, the decline may be more gradual than the headlines suggest, which could give it more time to adjust than its Organization for Economic Cooperation and Development (OECD) peers. Measuring birth rate trends is less straightforward than might be thought. Rather than rely on the “crude” figure (live births per thousand over a given period), most demographers look at movement in the total fertility rate. The TFR is an annual snapshot based on assumptions about the number of children that women of reproductive age will have borne by the end of their forties. To oversimplify, determining the U.S. TFR for 2024 involves calculating birth rates within different maternal-age groups in that year and then blending the results. In 2024 that led to a TFR of 1.63, materially below 2.1, which, assuming unchanged mortality rates and no net immigration, is generally regarded as the replacement rate: the rate needed to maintain a stable population in a developed country. The U.S. last reached a TFR of 2.1 in 2007. Hence those headlines.

But an increase in the age at which American women give birth undercuts the significance of that 1.63. There has been a dramatic fall in the rates of teen pregnancy and births (and abortions) in the U.S. The birth rate has also slumped for women in their twenties, but by more for those under 25 than over. The other side of the coin has been a leap in the birth rate for women in their thirties, especially among the over-35s. In part because of IVF, an example of the human ingenuity so underestimated by the Ehrlichs (if one that divides pro-lifers), women of that age group have in relative terms been outperformed by fortysomethings. There are now more births to women over 40 than to teens — not the worst thing.

These time-shifted births mean that the fall in the TFR is likely to exaggerate the true decline in the birth rate. In due course the later arrivals will make their presence felt in the completed cohort fertility rate (CCFR). This is primarily based on the number of births to women who were born in a given year before they turned 50. Data collected for the Human Fertility Database, the European project that is the leading source for such information, indicate that the CCFR for American women born in 1978 will end up at around 2.21, above the replacement rate and, in the OECD, second only to Iceland. As economist Jeremy Horpedahl notes, that exceeds the score of 2.0 for women born in the 1950s. If nothing else, that fact adds some more nuance, however dated, to narratives of imminent catastrophic decline. Horpedahl also describes an alternative method designed to adjust for the lag built into the TFR that yields a rate of 1.91 in 2022, below replacement but again comfortably above the “raw” TFR’s 1.63.

The CCFR is by definition backward-looking. We cannot be sure how many of the supposedly postponed births will actually occur, so add that to the long list of known unknowns about the birth rate in the years to come. Tellingly, the Congressional Budget Office has one set of population forecasts, the Census Bureau another, the Social Security Administration yet another, and they vary quite a bit.

We do know that Social Security and Medicare will run into serious financial difficulties early in the next decade. Immigration will not alter that fact, and neither would a surge in the birth rate, given that any above-trend children will be younger than ten years old when the ink turns deep red. We will have to navigate that horror with, more or less, the population we have now. How that could be accomplished is, mercifully, a discussion for a different occasion.

The good news is that, with one huge caveat, there is a decent chance — if Americans can adapt — that the U.S. will be able to work its way through the aftermath of that crunch and flourish despite an aging, and ultimately shrinking, population. The caveat is another known unknown. Will AI create mass unemployment, even if only “temporarily”? If it does, particularly among the educated class — typically the most dangerous class to antagonize — the problem will not be too few people but too many. The political turmoil that ensues will eclipse any that might be caused by demographic change.

If (a demanding if ) automation souped up by AI can, without triggering carnage in the labor market, increase productivity and generate more wealth, it will do a great deal to smooth the demographic transition. Thus China, facing an unusually treacherous decline in its numbers, has adopted a policy cheerily known as jiqi huanren (“replacing humans with machines”). That country leads the world in AI-assisted “dark factories,” facilities so automated that they barely need any lighting, and is, like many others, investing heavily in robotics. To take one application, the use of robots to help the elderly weather the trials of old age, pioneered by Japan for obvious reasons, is spreading beyond East Asia. Meanwhile, there may be fewer Japanese, but more of them are in the labor force than ever before.

Humanity adapted to a rising population, and it will adapt to a shrinking population. It will be a tragedy if depopulation panic induces Washington to succumb to the illusion that immigration (without which the CBO estimates the American population would start declining in less than five years) is the answer. Maintaining legal immigration at current levels would stave off population decline for some decades, but at what cost to social cohesion? The U.S. should always have room for the best and the brightest, but that is a choice, not a necessity. This country already has the technology, the talents, and the people it needs to continue to prosper.

But, until the materialization of Star Trek’s cornucopian economy, more of those people will have to be in the labor force and remain there for longer. The percentage of “prime working age” (25–54) Americans working or searching for work is over 83 percent, just short of 1999’s all-time high (84.6 percent). That is less reassuring than it seems, given how much the (welcome) surge in the female participation rate in the labor force (although to a level below that of the eurozone) has masked a troubling drop in the participation rates of prime-age men, from the high 90s in the postwar years to approximately 89 percent today.

Moreover, despite a recent uptick, only two-thirds of Americans ages 55–64 are in the workforce. In part, this can be blamed on how little progress has been made in how many years they can expect to live in good health — to use the jargon, their health-adjusted life expectancy (HALE). This failure goes some way to explaining why only about a fifth of over-65s are working (of whom just under 40 percent are part-time).

That said, Americans who make it to 60 can probably enjoy another 17 to 18 years in which they are in respectable shape. Increasing the “core” retirement age to 70 from 67 (which is only two years higher than its Depression-era starting point of 65) could, depending on how it is structured, make a considerable dent in the Social Security deficit, even more so if additional hikes are tied to a lengthening in HALE. The argument for pushing the age up at least some of the way and with tweaks to help out poorer retirees is theoretically irresistible, but support for that is not there. Judging by polling, propping up the system with higher taxes (ideally payable by someone else) would be a more popular option. Surprise! It would also be more damaging.

Older workers are already doing more to help out. Labor participation rates among U.S. over-65s are increasing and are above the OECD average, but it’s hard not to think that they could be higher — if not now, at least in time. About a third of Americans between the ages of 65 and 69 toil on, as do one in five of those between 70 and 74.

The HALE data are a reminder that Americans are not known for taking care of themselves, but once again, our species’ inventiveness is throwing us a life buoy. Remarkable pharmaceutical advances, including but not confined to GLP-1s, should boost HALE and the size of the potential workforce, and also, while they’re at it, reduce the pressure that extended longevity puts on Medicare. It won’t hurt that work is less physically challenging nowadays and that more of it can be done from home.

Contrary to myth, a stable or even declining population is compatible with strong economic growth, as many Eastern European states have demonstrated since the collapse of communism. Theirs may be a special case, but it shows that there is a lot more to a nation’s economy than the number of its inhabitants. Improvements in productivity, many of them attributable to technological innovation, will weaken the links between population size and prosperity, growth, and well-being. As automation accelerates further, those links will continue to fray. So long as that acceleration does not outpace our impressively adaptable species’ ability to keep up, that will be all to the good.

There is no “right” size for a country. The small can flourish as well as the large, but the U.S. will remain a giant, albeit one that towers a little less. Its economy will be reshaped as its people age — not least because consumption patterns will change — but there is no reason that it has to shrink. But America’s adaptation to demographic realities must incorporate acceptance, including fiscally, of the reality that, while the fall in the birth rate might be slowed down, there is little more that can be done. Regardless of occasional positive bumps such as in South Korea (so far) in 2024 and 2025, there is no politically feasible path to a material change of course. The decline may eventually level out, but no one knows when.

The causes of today’s birth rates are deep, varied, both novel and long-standing (they were going down in the U.S. and much of Western Europe well before easy access to contraception), fiercely debated, and far from fully understood. They are frequently, if inadequately, lumped together as “modernity,” a catch-all label so broad that its very imprecision helps explain why natalist policies have had so little effect on the birth rate. A change in societal expectations would do more, but such a vibe shift is not in sight, and there is not much that a government can (or should) do to influence that. Medals for the fecund should be reserved for countries where the state does not know its place.

So, what to do?

We should recognize how little we do know and can know about the implications of a declining birth rate, a topic where demographic, political, social, economic, and technological uncertainties collide. The best response will vary from country to country, but for now, the U.S. should rely on both automation and, as mentioned above, making better use of the people it has. The latter will include fostering higher and more prolonged participation in the workforce, a process that, if successful, would yield social as well as economic rewards. To the extent that automation cannot remedy any labor shortages, this ought to lead to higher pay and greater investment in employees’ skills, with those in the shortest supply (probably younger workers) gaining the most. There will be fewer children, but, in a wealthier society, more money to pay for their education and, if wanted, to encourage an increase in their numbers. It’s easy, for example, to justify a greater focus on tackling infertility, and many employers would benefit from doing more to enable their workers to combine motherhood and career, but there is no compelling need to spend taxpayer funds on natalist initiatives even if one could be conceived that generates more than a temporary return.

Each generation of Americans should aim to leave their successors a better, wealthier, and stronger country. Does that imply that it must be more populous? Mine is a minority view around here, but for many reasons, of which automation is merely one of the more persuasive, the answer in 2026 is no.