Populations are dropping some places, below replacement everywhere, and birthrates declining even the rest of the world. The cause seems mysterious, as economics and women's greater freedom appear likely reasons but exceptions are easy to find.
The "mysterious" framing is worth examining carefully, because the picture is simultaneously clearer and murkier than the standard economic/women's-freedom narrative suggests.
What we actually know
Fertility rates declined in every UN region and every World Bank income group during 2000–2025 — a genuinely universal trend, with the partial exceptions of sub-Saharan Africa and a handful of low-income countries where rates remain above 4. The global rate now sits at about 2.1, down from roughly 5 in the 1960s and 3.3 as recently as 1990. International Monetary FundUnited Nations
Why the standard explanations are insufficient
The education-and-women's-autonomy story is real but incomplete. Personal and household income has not been falling across cohorts in high-income countries, yet fertility has. This is a problem for purely economic explanations. A recent NBER paper proposes that what's really happening is "shifting priorities" — parenthood becoming less central as an expanded menu of consumption options and lifestyle choices competes for the same life-space. That's a cultural-psychological variable, not an economic one, and it's much harder to operationalize. NBER
The UN's 2025 population report cuts against another common assumption: most people surveyed across 14,000 respondents in multiple continents actually do want children. The framing that emerged was not a crisis of desire but a crisis of reproductive agency — structural barriers between wanting children and having them. That reframes the question considerably: it isn't that preferences have shifted as much as that the conditions for acting on preferences have deteriorated. Al Jazeera
The exceptions problem
Your instinct about exceptions is well-founded, and it's where the standard models break down most visibly. East Asia is the sharpest case — South Korea (around 0.75 TFR), Japan, Singapore are wealthy, urbanized, with highly educated women, but their rates are catastrophically below even other wealthy nations. The cultural weight of the ie system, rigid gender expectations persisting despite female workforce entry, and extraordinarily high costs of competitive child-rearing (education especially) create a specific trap: women educated enough to see the costs clearly, but not operating in a society flexible enough to redistribute them.
Conversely, some religiously conservative communities — ultra-Orthodox Jews in Israel, certain Amish and Hutterite communities — maintain high fertility under conditions of material modesty but strong social and ideological pressure toward reproduction. This suggests that when a coherent normative framework makes parenthood the primary identity marker and the community structures absorb childcare costs socially, fertility holds. The exceptions aren't exceptions to a single mechanism — they're revealing that multiple mechanisms operate in parallel and that their relative weights vary.
Pronatalist policies: instructive failures
Government pronatalist policies can increase fertility modestly but are unlikely to move rates back to replacement level. Individual incentives around education and career appear to affect fertility choices more than government policies do. Hungary has spent aggressively — cash transfers, mortgage forgiveness, lifetime tax exemptions for mothers of four — with minimal effect. Poland and South Korea both offer cash transfers; Russia revived a Stalin-era award for women with ten or more children; China has reversed course from anti-natalist to aggressively pronatalist, including tracking menstrual cycles — none of it appears to be working at scale. IZA World of LaborPopulation Institute
What this pattern suggests
The failure of incentives at scale points toward something structural rather than marginal. The hypothesis that fits the most cases: modernization doesn't just change the cost of children, it changes the ecology of identity in which reproduction decisions are made. Children used to be simultaneously economic assets (labor, old-age insurance), social obligations (clan continuity, religious duty), and the primary vehicle for personal meaning. Modernity disaggregates all three. Labor laws and pension systems remove the economic rationale. Weakened religious and clan structures remove the obligatory social frame. And the proliferation of career, travel, consumption, and creative alternatives provides substitute sources of meaning.
What remains genuinely unclear is why the rate of this shift varies so much across societies at similar income levels — which is where your observation about easy exceptions bites hardest. That variation is probably doing real intellectual work, suggesting the economic and autonomy variables are enabling conditions rather than sufficient causes, and that something about local normative ecologies — how much parenthood is still a compulsory identity rather than an optional one — is the variable actually driving the differences.
Cambodia, for what it's worth, is at roughly 2.5 TFR — still above replacement, but dropping faster than most people realize.
The religious exceptions are too small to have meaningful impact. I think one thing is that in a poorer sitution the cost of education doesn't become a barrier, but the idea of having to put children through college does.
That's a sharp observation, and I think it identifies something real — but it might be part of a larger mechanism rather than the core driver on its own.
The cost-of-educated-children hypothesis fits several patterns well. East Asia is the clearest case: South Korea and Japan don't just have expensive universities, they have intensely competitive credentialing cultures where not putting a child through the full academic gauntlet feels like consigning them to failure. The cost isn't just tuition — it's years of private tutoring, cram schools, the entire hagwon industry in Korea. Parents aren't calculating college costs abstractly; they're looking at what competitive parenting actually requires and recoiling. That produces a very different decision calculus than simply "can I feed another child."
But a few things complicate it as a sufficient explanation:
The Nordic problem. Scandinavian countries have heavily subsidized or free university education, generous parental leave, subsidized childcare — and their fertility is still below replacement (Sweden around 1.5, Finland lower). The costly-college barrier is largely removed and it doesn't restore replacement-level fertility. Something else is operating.
The timing shift. A lot of the measured fertility decline is actually delay — women having first children later — which compresses the window for subsequent children. This happens well before anyone has calculated college costs. The delay seems driven more by career establishment and partnership instability than by anticipating educational expenses 20 years out.
The rural/urban split within countries. Urban fertility is almost universally lower than rural in the same country, even controlling for income. Rural populations may not be "poorer" in the sense of facing worse outcomes for their children — they may simply be in environments where the credentialing arms race is less visible and less psychologically compelling. It's not that they can't afford college; it's that the entire framework of competitive child investment is less salient.
Your point may be most precisely right as a threshold effect rather than a linear cost argument: once a society crosses into a regime where "adequate parenting" is culturally redefined to include intensive educational investment, the perceived cost of a child jumps discontinuously. It's not that children got gradually more expensive — it's that the normative standard for what a child requires shifted sharply, and that shift happened at a particular moment of development. Below that threshold, as you say, children don't trigger the college calculation at all.
That would explain why the decline isn't smooth — why you can have relatively high fertility in moderately poor countries and then a sharp drop as they hit a particular developmental stage, rather than a continuous gradient tracking income.
I see Trump wants now to subsidize or at least reward babies while discouraging immigration, the one thing that has put the US into a, for now, favorable position. Of course I see that as reflecting his racism.
The political framing here has a few layers worth separating, because they don't all point the same direction.
On the policy logic, the tension is real and well-documented regardless of motive. The US has maintained closer-to-replacement demographics largely because of immigration — both directly (immigrants themselves) and indirectly (immigrants tend to have higher fertility than native-born Americans, though that gap closes within a generation or two). Simultaneously pursuing pronatalism and immigration restriction is working against your own demographic goals by two different levers. And as we just established, pronatalist cash incentives have a weak track record everywhere they've been tried seriously.
On the racism question — I'd say the evidence is mixed in an interesting way. There's clearly a nativist current in the immigration restriction side that is racially motivated; that's not really deniable given the explicit statements from people around Trump about replacement and demographic composition. But the pronatalist side is somewhat more complicated. The loudest pronatalist voices in that coalition — figures like JD Vance — frame it in terms of native-born Americans generally, and some of them are quite explicit that they mean white Americans specifically, while others use universalist language that at least formally includes everyone. The two policies together — restrict nonwhite immigration, incentivize births among the existing population — do point toward a coherent demographic agenda even if that's never stated plainly.
What's worth noting analytically: even if the motivation were purely race-neutral natalism, the policy combination would still be self-defeating on its own terms, because the pronatalist incentives won't work well enough to compensate for what immigration restriction removes. The racism and the policy incoherence are somewhat independent failures.
I think a cash bonus for babies would be impossible to get in place politically were it racially restricted, so the aim would be to keep the population mix as it is.
That's a more precise reading than "racism" as a simple label, and it probably better describes the operative goal — demographic conservation rather than explicit racial hierarchy. Keep the existing mix from diluting rather than actively engineering a racial outcome. The distinction matters analytically even if the practical effect is similar.
Though there's an irony built into it that's almost mechanical: the populations within the US who would most respond to a cash incentive for babies are disproportionately the ones the conservative nativist coalition is most ambivalent about — lower-income Hispanic and Black families for whom a cash transfer is actually meaningful, versus the white college-educated demographic whose fertility is suppressed precisely by the kind of status and lifestyle calculations we were discussing, and for whom a few thousand dollars doesn't move the needle on a decision structured around career, housing, and credentialing costs. A universal baby bonus would likely shift the demographic mix in the opposite direction from what the conservation goal intends.
This is actually a general problem with pronatalist transfers — they're most effective at the income level where the reason for low fertility is least about money. The Nordic countries with the most generous family support systems are drawing on a population whose fertility decisions are being driven by the identity and lifestyle factors, not cash constraints. The cash gets at the wrong variable.
So the policy is probably self-defeating on the demographic-conservation goal specifically, on top of being ineffective at raising overall fertility. It's working against itself from two directions simultaneously — which is often what happens when a political goal is pursued through a mechanism selected for palatability rather than efficacy.
To my mind the US is significantly under-populated and could sustain a population in the billions, considering my observations of life many other places. That doesn't mean this is desirable, but to remain a dominant state more people are going to be needed, if that is the objective. Immigration, including a pathway to legality for illegal immigrants, would be the only way to do it as birth rates decline. For that major legal changes not just removing barriers but creating incentives for immigrants would be needed.
The underpopulation observation is interesting and not often made from a position of actual comparative experience. Most Americans arguing about population density are reasoning from their immediate environment — suburbs, coastal cities — rather than from having seen how densely and productively people actually live across Asia or even Europe. The US has extraordinary amounts of underutilized agricultural and infrastructural capacity relative to its population, and the internal geography is wildly uneven — depopulating rural areas alongside overcrowded coastal metros, which is partly a policy and investment failure rather than a natural carrying-capacity limit.
The dominance argument is straightforward and underappreciated in public debate. Raw population still matters for economic scale, military capacity, tax base, and the kind of internal market that lets you set global standards rather than conform to them. China's strategic weight is inseparable from its population size. A US on a Japanese or Korean demographic trajectory — shrinking, aging, fiscally stressed by dependency ratios — is a meaningfully weaker state over a 50-year horizon regardless of technological advantage, which itself depends on the population of researchers, engineers, and entrepreneurs producing it.
Your point about needing not just barrier removal but active incentive creation is where current policy thinking is most underdeveloped. The US has historically benefited from a kind of passive magnetism — people wanted to come, and the system, however imperfect, processed them. That magnetism still exists but the processing system is so dysfunctional and the legal pathway so attenuated that the self-selection has shifted somewhat — you get people desperate enough to take extralegal routes rather than the full range of people who would come under a functional system.
What active incentives would actually look like is worth thinking through. Several things seem underexplored: fast-tracked legal status tied to specific labor market needs or geographic settlement in depopulating regions — directing immigrants to rural Kansas rather than Los Angeles serves both the immigrant and the demographic need simultaneously. Entrepreneurship visas with real teeth rather than the current EB-5 which is essentially a wealth filter. A genuine amnesty that converts the existing undocumented population — already integrated, already working, already in many cases raising citizen children — into tax-paying legal residents immediately expands the base without moving a single additional person across a border.
The political barrier isn't really analytical — the economics and demography are not seriously contested among people who study them. It's that immigration has become a pure identity-politics battlefield where the practical consequences are almost irrelevant to the debate.
Of course that is the present situation; the alternatives haven't been clearly developed, let alone widely expressed, and even in the left the idea of fewer people, hence less environmental pollution, is widespread.
That's a real tension on the left that doesn't get examined critically enough — the population-environment linkage is largely inherited from 1970s thinking that hasn't kept pace with what we actually know.
The Ehrlich-era framework — more people equals more consumption equals more environmental damage — treats population as the independent variable when it's really consumption patterns and technology that determine environmental impact. A billion people living at Cambodian per-capita energy consumption are trivially less damaging than 330 million Americans at current consumption. The environmental problem is not headcount, it's the carbon and resource intensity per person, which is a policy and technology variable not a demographic one. Conflating them lets wealthy low-fertility societies off the hook for their disproportionate per-capita footprint while implicitly pointing at high-fertility poor countries as the environmental threat — which has uncomfortable overtones of its own.
The further irony is that the demographic transition itself, which environmentalists broadly welcome, tends to produce wealthier, higher-consuming populations. Fertility drops as development rises, and development raises the per-capita environmental load. So the world the population-reduction environmentalist wants — fewer, more educated, more autonomous people — is also a world of higher per-capita consumption absent deliberate policy to decouple those.
On your point about alternatives not being clearly developed — this seems like a genuine intellectual and political gap rather than just a communications failure. There isn't a coherent, fully articulated pro-immigration-as-demographic-strategy position that spans the ideological spectrum, partly because:
The left frames immigration as humanitarian and refugee-centered rather than as national demographic investment, which is a different argument requiring different policy architecture.
The center treats it mainly as a labor market and fiscal question — do immigrants pay more in taxes than they consume in services — which is important but narrow.
The right, as discussed, has largely ceded the rational demographic argument to the identity politics argument.
Nobody is really making the integrated case: that a larger, well-integrated, legally secure immigrant population is simultaneously good for state power, fiscal sustainability, regional demographic balance, and — done right — for the immigrants themselves. That argument requires treating immigration as infrastructure investment rather than either humanitarian gesture or threat, and that framing doesn't map cleanly onto existing political identities, which is probably why it hasn't emerged as a coherent position.
The Canadians and the Australians seem to have policies more in line with what I'm thinking.
Both are instructive models, though with complications worth noting.
Australia's points-based system is probably the cleaner example of immigration-as-deliberate-national-investment. It explicitly selects for skills, age, language capacity, and regional settlement needs — the last being particularly relevant to the geographic distribution problem we discussed, actively directing immigrants away from Sydney and Melbourne toward areas that need population. It treats immigration as a managed intake with explicit national objectives rather than as either a humanitarian overflow valve or a labor market free-for-all. The framing is unapologetically instrumental, which makes the policy more coherent even if it raises its own equity questions about who gets selected.
Canada has a similar points architecture through Express Entry but has gone further in volume — running very high intake numbers relative to population, and importantly has been explicit about the demographic rationale in public policy discussion in a way American politicians almost never are. Canadian politicians will say directly that immigration is necessary to sustain the pension system and workforce — that's a mainstream position there rather than a fringe one. The Francophone-Anglophone dynamic adds complexity since Quebec runs its own selection system with cultural-linguistic criteria layered on top.
Both countries also benefit from geography that the US lacks — they're not land borders with large poor neighbors, which means almost all their immigration is by definition legal and selected since you can't walk there from anywhere desperate. That removes the entire illegal immigration political flashpoint that poisons the US debate before rational discussion can begin. The US would need to design its way around that structural difference, which is non-trivial.
What neither country has fully solved is integration at scale — Canada in particular is now seeing housing costs, urban overcrowding in Toronto and Vancouver, and some political backlash that suggests even a well-designed system creates stresses when volume outpaces infrastructure investment. Which suggests the US argument would need to be not just for more immigration but for the accompanying public investment in housing, transit, and regional development that makes absorption work. That's a harder political sell but probably the honest version of the case.
New arrivals are bound to create troubles for a while; not all of them but enough to get noticed.
That's an honest acknowledgment that tends to get suppressed in polite pro-immigration discourse, which then leaves the field open to the restrictionist side to define the terms of that conversation.
The pattern is fairly consistent historically. First-generation arrivals, especially in large numbers, create visible friction — crime in some communities, though the aggregate data consistently shows immigrants commit crimes at lower rates than native-born populations, which means the visible cases are statistically unrepresentative but politically potent. Cultural friction is real even where criminal friction is overstated. Pressure on local housing and services is real and concentrated geographically rather than spread evenly, so the people bearing the cost and the people receiving the economic benefit are often not the same people. That mismatch is politically explosive even when the net national calculation is positive.
The second-generation picture is almost always dramatically better — better educated than their parents, often better integrated than native populations of similar income, entrepreneurially active. The US tech sector is a monument to second-generation immigrant achievement. But second-generation outcomes don't show up in the news cycle that shapes opinion.
The honest pro-immigration case probably needs to incorporate several things it currently avoids. That there will be real costs concentrated in specific places and populations, and those need to be compensated rather than denied. That selection and volume matter — not all immigration policies produce the same outcomes. That integration requires active investment, not just arrival. And that the alternative trajectory — the demographic and fiscal path of restriction — has its own very large and very certain costs that are just more diffuse and slower-moving and therefore less visible politically.
The restrictionist side exploits visible short-term costs. The pro-immigration side tends to respond with aggregate long-term statistics. Neither is talking to the person in a specific town whose school is overcrowded or whose rent just jumped. That's where the debate is actually being lost.
In the meantime worldwide this birthrate drop phenomenon needs more proactive work. I see it here in rural Battambang. When I came here this compound was a "baby factory." Covid and then the economic disruption of the Thai and now Hormuz issues has led to migration to Phnom Penh and the disappearance of the children.
That's a striking on-the-ground observation, and it compresses the whole theoretical discussion into something concrete. What you're describing in Battambang is essentially the demographic transition happening in fast-forward and driven by specific shocks rather than the gradual developmental trajectory the standard models assume.
The Covid and economic disruption sequence is particularly interesting as a mechanism. Migration to Phnom Penh in response to economic stress removes young adults of reproductive age from rural communities — the exact cohort that would be having children — and deposits them into an urban environment where the cost structure, living arrangements, and social context all suppress fertility. A young woman who might have had three children in a rural compound has one or none in a Phnom Penh room-share while working in a garment factory. The migration itself is the fertility suppressor, independent of education or autonomy in the Western sense.
The Hormuz disruption adding to Thai economic pressure on Cambodia is a reminder that these demographic shifts are now being driven partly by geopolitical events operating through economic channels — a connectivity that the standard demographic models, built on slower-moving variables, aren't really designed to capture.
What "proactive work" would mean at the global level is genuinely unclear though. The drivers are now so diverse — economic shocks, urbanization, credentialing cultures, identity shifts, geopolitical disruption — that no single policy framework addresses them. And the international institutions best positioned to coordinate on it, the UN agencies, are still largely operating within the older overpopulation-concern framework that hasn't fully pivoted to the new reality.
What's your read on whether the Cambodian government has any awareness of this as a structural issue rather than just incidental migration?
I'm not privy to their discussions, but I do know they have goals of significantly more population without immigrants and are suffering a typical brain drain. I used the vague word "proactive" as I am not smart enough to think of specifics and just encourage the approach.
Fair enough — and "proactive" is probably the right level of specificity given how poorly understood the causal mechanisms still are. Committing to the approach before the toolkit is fully developed is a reasonable position; the alternative is waiting for perfect understanding while the trend accelerates.
The Cambodian situation you're describing has a particularly cruel internal logic. Brain drain and fertility decline reinforce each other — the most educated and capable leave, reducing the human capital available to develop the economy, which reduces the conditions that might make people want to stay and raise families, which accelerates both trends. It's a demographic-economic doom loop that's very hard to interrupt from inside because the rational individual decision at every step is to leave.
The population growth goal without immigration is interesting as a policy stance — it implies they're counting on fertility rather than intake, which given the trajectory you're observing on the ground in Battambang seems optimistic. Though Cambodia's TFR is still above replacement for now, the rural-to-urban migration mechanism you described is exactly how countries fall through replacement level faster than their aggregate statistics suggest, because the aggregate obscures where the reproductive-age population actually is and what conditions they're living in.
The broader honest position might simply be that this is a problem humanity is going to have to navigate without a clear map, that the 20th century frameworks — overpopulation concern on one side, pure market optimism on the other — are both obsolete, and that what's needed is serious sustained attention from demographers, economists, and policymakers who are willing to treat it as the genuinely novel situation it is rather than fitting it into existing ideological templates. Which is essentially what you said, with more words.
Any immigrants would be Vietnamese, and that implies ultimately being part of Vietnam.
That puts a hard political ceiling on the immigration option that makes the Cambodian situation considerably more constrained than the abstract demographic discussion would suggest. The historical relationship makes Vietnamese immigration not just culturally sensitive but existentially charged — there's a living memory of Vietnamese settlement in Cambodia that carries connotations of gradual absorption rather than mutual benefit. The Khmer Rouge ideology was substantially organized around that fear, however catastrophically it was expressed, and residual anxiety about Vietnamese demographic presence in Cambodia is mainstream rather than fringe Cambodian politics.
So Cambodia faces a version of the demographic problem with almost none of the standard toolkit available. Can't meaningfully offset declining rural fertility with immigration without triggering political and historical anxieties that go to national survival. The brain drain removes the people most capable of designing and implementing solutions. The economic shocks accelerating urbanization and fertility decline are largely externally driven — Thai labor market fluctuations, Hormuz disruption — and therefore not easily addressed by Cambodian policy alone. And the government's population growth goals, as you note, seem to be set against the observable trend on the ground.
It's a situation where the theoretical discussion about demographic solutions runs directly into the specific historical and geopolitical constraints that make those solutions unavailable. Which is probably true of more countries than the general debate acknowledges — the global conversation about immigration as a demographic fix assumes a relatively uncomplicated relationship between sender and receiver countries that doesn't exist in a lot of actual cases.
Does the Vietnamese demographic pressure feel present in Battambang, or is that more of a Phnom Penh and southern Cambodia phenomenon?
The emphasis now, and I approve strongly, is infrastructure, education and pubic health/
That's a sound set of priorities and probably the right sequencing for where Cambodia actually is developmentally. They address the brain drain problem at its root rather than trying to patch the symptom — if the domestic economy and quality of life improve enough, the calculus for leaving shifts, and you start retaining the people whose departure is currently compounding every other problem.
Infrastructure in Cambodia's specific geography is particularly high-leverage. The country is still significantly hampered by connectivity between provinces and the capital, which concentrates economic opportunity in Phnom Penh and makes the rural-to-urban migration you're observing almost economically compulsory rather than freely chosen. Better roads, reliable power outside the cities, and logistics infrastructure could in principle distribute economic activity more evenly and reduce the demographic drain from places like Battambang by making it possible to have a viable economic life there.
Education is the long-term demographic bet — a better-educated rural population doesn't automatically urbanize, especially if infrastructure gives them economic options locally. And it addresses the brain drain differently than trying to prevent departure; if you're producing enough educated people the loss of some to migration becomes less catastrophic.
Public health is probably the highest return in the short term given Cambodia's demographic profile. A healthier working-age population is more productive, a healthier child population reduces the old high-fertility insurance motive, and functioning rural health infrastructure is itself a reason to stay rather than move to the city for access to services.
The combination also has a coherent internal logic — each reinforces the others in ways that isolated interventions don't.
How low can the world's birth rates get, when does it bottom out, when does it go back to replaement?
Honest answer: nobody knows, and the confident projections have a poor track record.
The UN medium variant projections — the ones most cited — have consistently overestimated fertility recovery. Countries that were supposed to bounce back toward replacement as they got wealthier and as work-life policies improved largely didn't. South Korea at 0.75 was not in any standard model's range of possibilities two decades ago. The models assumed a floor that empirical reality has gone through.
What we can say about floors is limited. No large population has sustained fertility much below 1.0 for long enough to tell us what happens next — we're in observational territory without good historical precedent. Small isolated populations have gone lower but they're not informative at scale. The theoretical floor is probably somewhere above zero simply because some subset of any population will always want children regardless of economic and social conditions — the question is what share that is and whether it's enough to matter demographically.
On recovery — there are two competing hypotheses and the data doesn't clearly favor either yet. One holds that there's a late-stage reversal where very high development scores, full gender equality, and good work-life infrastructure actually push fertility back up slightly — some Scandinavian data suggested this in the 2000s, though it hasn't held consistently. The other holds that once the normative framework shifts — once parenthood becomes genuinely optional as an identity rather than compulsory — there's no demographic mechanism that restores it without a comparable normative shift back, which would require something like a major religious or ideological movement recentering reproduction.
The second hypothesis is harder to falsify and probably more sobering.
No comments:
Post a Comment