The Subtractive Age
What happens to a civilization built on more when the world begins to deliver less
The Growth Assumption
There is a thought experiment I keep returning to, and it goes like this.
Imagine you could gather every major political thinker of the last two centuries in a single room. Marx and Mill, Keynes and Hayek, the developmentalists of the postwar South, the neoliberals of the Washington Consensus, the social democrats of Scandinavia, the Leninists, the libertarians, the technocrats of Beijing. They would agree on almost nothing. They spent their lives, and in some cases other people’s lives, fighting over property, class, markets, the state, the individual, the nation. And yet if you asked each of them one question, quietly, on the way out the door, you would get the same answer from all of them. The question is: next year, will there be more?
More people, more workers, more output, more of everything. Every one of them assumed yes. Not as a conclusion they had argued for, but as a premise so obvious it never occurred to them to state it. Liberalism, socialism, nationalism, developmentalism: these are rival theories of how to divide the fruits of growth. They are not, and never were, theories of what to do when the fruits stop coming. The fight was always over the surplus. Nobody asked whether there would be one, for the same reason nobody asks whether the sun will rise. It always had.
That premise is now failing, and it is failing in a way that cannot be argued with, lobbied against, or voted down, because it is failing arithmetically. The United Nations’ most recent projections have the human population peaking around the middle of the 2080s at roughly 10.3 billion, then turning down. More than half of all countries already have fertility below the replacement rate of 2.1 children per woman. South Korea’s rate touched 0.72 in 2023, the lowest ever recorded for a national population; in Seoul it was 0.58. China began shrinking in 2022, for the first time since the famine years of Mao. Italy recorded fewer births in 2024 than in any year since the country was unified in 1861. These are not forecasts in the sense that a weather forecast is a forecast. The workers of 2045 have already been born, or rather, have already not been born. What is coming is not a prediction. It is a schedule.
I will spend the later portions of this essay on that schedule and on what politics looks like when it arrives. But first I want to do something that sounds academic and is actually radical, which is to establish how strange, how recent, and how load-bearing the growth assumption really is. Because if growth is just one policy preference among many, then its ebbing is a problem to be managed, a matter for finance ministries and actuarial tables. If, on the other hand, growth is the hidden foundation on which the entire edifice of modern politics was built, its metaphysical basement, then we are not facing a downturn. We are facing a change in the physics of political life. I think the second description is the true one, and I want to earn that claim rather than assert it.
The trap, and the escape that rewired our minds
Start with the world before growth, because for nearly all of human history that was the only world there was. In 1798 an English clergyman named Thomas Malthus published An Essay on the Principle of Population, and its argument was brutal in its simplicity: population, left unchecked, multiplies geometrically, while the food supply expands only arithmetically. The result is that any improvement in living standards is temporary. More food means more surviving children, more surviving children means more mouths, more mouths means the gains are eaten, literally, until the average person is pressed back down to subsistence. What holds the line, in Malthus’s grim accounting, is “vice and misery,” to which he later, under some social pressure, added the gentler option of “moral restraint.”
It is fashionable to mock Malthus, because he published his theory at almost the precise moment it stopped being true. But this gets the history backwards. For the several thousand years before he wrote, Malthus was essentially correct. The peasant of 1700 was not meaningfully better fed than the peasant of the Roman Empire. Populations rose and fell with harvests and plagues, but living standards oscillated around a grim floor, because every productivity gain was converted into people rather than prosperity. Economists now call this the Malthusian trap, and it was not a theory. It was the human condition.
Then England escaped. Somewhere in the late eighteenth century, the pace of innovation crossed a threshold where output could grow faster than population could eat it. For the first time in the history of the species, the number of people and the income per person rose together, decade after decade. The economists Oded Galor and David Weil have modeled this as one continuous process, a phase transition in which societies tip from millennia of stagnation into self-sustaining growth, driven above all by the accumulation of human capital: educated people producing ideas that produce more than educated people consume. The trap did not merely loosen. It shattered, and the shattering spread, country by country, across the nineteenth and twentieth centuries.
Here is the part that matters for our story. The escape was so complete, so unprecedented, and so intoxicating that Western civilization did something very human with it: it forgot the escape was an event and began treating it as a law of nature. Two centuries is a long time in political memory and an eyeblink in historical time, but two centuries of compounding growth was enough to rebuild the entire architecture of collective life on the assumption that the curve points up forever. Constitutions do not say this. No treaty enshrines it. It is simply underneath everything, the way the ground is underneath a city.
The old man in Tunis
There is an older way of thinking about these matters, and it comes from a fourteenth-century North African who watched civilizations die for a living. Ibn Khaldun, writing the Muqaddimah in 1377, proposed that the rise and fall of dynasties follows a generational rhythm governed by something he called asabiyyah, usually translated as group solidarity or social cohesion. A people forged in hardship at the margins possesses fierce asabiyyah; that cohesion lets them conquer. The second generation consolidates what was won. By the third and fourth generations, raised inside the comfort their grandparents built, the solidarity dissolves. Luxury does what no enemy could. The dynasty softens, and some hungrier group from the edges takes its place. Three to four generations, he observed, is roughly what a dynasty gets.
I want to be careful here, because Ibn Khaldun was writing about political and military cohesion, not birth rates, and pressing his cycle into demographic service is an analogy, not an equivalence. But the analogy is worth making openly, because his deepest insight travels remarkably well. Ibn Khaldun’s point was that decline is structural rather than moral. Civilizations do not fall because people wake up one day and choose corruption. They fall because success dissolves the very conditions that produced it. The discipline born of scarcity cannot survive abundance; prosperity is its own solvent.
Now hold that thought against the central demographic finding of our era: everywhere on earth, without exception, the same package of achievements reliably drives fertility below replacement. Wealth. Education, above all the education of women. Urbanization. Female autonomy and opportunity. Falling child mortality, which removes the ancient logic of bearing many children so that some survive. These are the things every ideology promised, the very definition of development, the outcomes we measure progress by. And their reliable consequence, in every society that attains them, is that people choose to have fewer children than would sustain the population. Not because anything went wrong. Because everything went right. Success, again, is the solvent. The old man in Tunis would not have been surprised, though he would have been fascinated by the mechanism.
The stage the model never had
Demographers formalized the modern version of this story in what is called the demographic transition model, first sketched by Warren Thompson in 1929. A traditional society sits at stage one, with high birth rates and high death rates canceling each other out. Then medicine, sanitation, and food security arrive; death rates plunge while birth rates stay high, and population explodes. That is stage two, and it is the stage the twentieth century mistook for permanence. In stage three, birth rates fall as children shift from being farm labor to being expensive investments in education. Stage four is the promised land: low births, low deaths, a stable population. Thompson’s model ended there, at equilibrium, and for decades everyone assumed that was where the story ended too.
But by the 1980s, European societies began doing something the model had no stage for. Fertility did not glide down to replacement and level off. It fell through the floor and stayed there. Demographers now debate whether to speak of a stage five, a phase of sustained decline, deaths exceeding births, the population pyramid inverting into something top-heavy and unstable, a spinning top slowing down. Japan is there. Korea is there. Italy, Portugal, and most of Eastern Europe are there. The United Nations now classifies nearly a fifth of all countries, China and Spain among them, as being at “ultra-low” fertility, below 1.4.
The sharpest theoretical question of our time hides inside a piece of terminology. To call this a “stage” implies passage, something a society moves through on the way to somewhere else. But no rich society, anywhere, has ever returned durably to replacement fertility. Not one. The evidence, which I will examine in detail later, suggests that below-replacement fertility is not a dip but an attractor, a place societies fall into and do not climb out of. If that is right, stage five is not a stage. It is a destination, and the demographic transition model turns out to be a description not of a journey to equilibrium but of a slow, dignified exit.
The invention of “the economy”
Here is a fact that should be far more famous than it is: the idea that governments exist to grow the economy is younger than many people now alive. The economic historian Matthias Schmelzer has traced what he calls the growth paradigm, the belief that GDP growth is desirable, imperative, and essentially without limit, and found that it is an invention of roughly the last hundred years. Before that, there was no such thing as “the economy” in the sense we mean, a single national object with a size that could go up or down. There were harvests, trades, revenues, and prices, but the aggregate did not exist because nobody had built the instrument to see it.
The instrument was national income accounting. Simon Kuznets constructed the first modern version for a report to the United States Senate in 1934, in the pit of the Depression, so that policymakers could finally answer the question of how bad things actually were. And Kuznets, to his everlasting credit, attached a warning label at the moment of creation: this number measures productive capacity, he said, not welfare. The welfare of a nation can scarcely be inferred from it. Nobody listened, or rather, everybody found the number too useful to hear the caveat. By the Bretton Woods conference in 1944, GDP was the world’s standard measure, and within a decade its growth had become the standard by which governments themselves were measured.
What happened next is one of the great unremarked convergences of the twentieth century. The Cold War was advertised as a total conflict between incompatible systems, and in one respect the advertising concealed a deeper agreement. Washington and Moscow both staked their legitimacy on the same metric. Khrushchev did not promise to bury the West with superior justice; he promised to out-produce it. Five-year plans and quarterly earnings were dialects of a single language. Growth had become, on both sides of the Iron Curtain, the thing a state was for.
And once you see this, you begin to see growth’s fingerprints on every ideology in the modern inventory. Liberalism’s genius is positive-sum politics: rising prosperity lets everyone gain without anyone visibly losing, which is how a diverse society defuses its distributional conflicts short of violence. Marxism, whatever else it is, is a theory of the development of the productive forces, with the just distribution of an ever-larger product as its moral core; Marx’s quarrel with capitalism was never that it produced too much. Nationalism trades in vitality, in youth, in the promise of more of us. The developmental states of Asia offered growth in exchange for patience; China’s party-state, having quietly retired most of its founding doctrine, rests today on what scholars call performance legitimacy, the delivery of rising living standards in place of the vote. And the welfare state, the crowning achievement of twentieth-century social peace, is the purest artifact of all: pay-as-you-go social insurance is, in its very mechanics, a wager that tomorrow’s workforce will be larger than today’s. Each generation pays for its parents’ retirement on the understanding that a bigger generation will pay for its own. It is a chain letter that works beautifully so long as the chain keeps lengthening.
The economist Benjamin Friedman gave this arrangement its most unsettling formulation. In The Moral Consequences of Economic Growth, he assembled the historical evidence that growth does not merely make societies richer; it makes them better. Periods of broadly shared rising incomes, he found, are the periods when societies expand tolerance, open opportunity, extend rights, and deepen their commitment to democracy. Periods of stagnation are when they turn inward, hunt for scapegoats, and retreat from generosity. His metaphor was a bicycle: modern society stays upright because it is moving forward. Stop the motion and it does not stand still. It falls over, and Friedman was explicit that this holds even in countries that remain rich in absolute terms. It is the direction of travel, not the altitude, that keeps the machine balanced.
Take Friedman seriously and the stakes of this essay come into focus. If tolerance, pluralism, and democratic patience are not free-standing virtues but the moral surplus thrown off by expansion, then permanent demographic contraction is not an economic inconvenience. It is a threat to the liberal order at the level of its operating conditions.
The prophets who faced the wrong direction
It would be false to say no one questioned the growth paradigm. A dissenting tradition insisted, sometimes eloquently, that growth must end. The Club of Rome’s Limits to Growth landed in 1972 with Donella Meadows’s system-dynamics models projecting overshoot and collapse on a finite planet. Herman Daly, building on Georgescu-Roegen and Boulding, spent a career arguing that a bounded ecosystem cannot host unbounded quantitative growth, and sketched the steady-state economy that would permit qualitative development without aggregate expansion. Later, the degrowth movement went further still, arguing that rich countries should deliberately shrink their material throughput before the biosphere shrank it for them. And behind them all stood the ur-anxiety of the postwar decades, Paul Ehrlich’s The Population Bomb, which convinced a generation that humanity’s defining peril was too many people.
Notice what all of these visions share. They imagined growth ending because it hit a wall: resource exhaustion, ecological collapse, famine, the planet’s hard ceiling. Growth as a rocket running out of sky. What almost no one imagined was the engine simply, quietly, switching itself off from the inside; that the danger was not the wall but the fuel; that prosperity itself would persuade people, one household at a time, in Seoul and Milan and Chengdu and São Paulo, to stop replacing themselves. The twentieth century fought a war against overpopulation and won it so completely that it forgot decline was even on the menu. The bomb went off in reverse, and the people watching the sky never heard it.
There was one economist who glimpsed the true shape of the thing, and his reward was to be right too early. In 1938, Alvin Hansen coined the phrase “secular stagnation,” warning that slowing population growth would chronically depress investment: fewer new workers to equip, fewer new households to house, and therefore a shortfall of demand that no interest rate could cure. The baby boom promptly buried him, and for seventy years his thesis was a museum piece, the cautionary tale of a Depression-era pessimist. Then, in 2013, Larry Summers stood up at an IMF conference and exhumed it, arguing that the rich world had slid into precisely Hansen’s condition, and tying it directly to demography: a declining working-age population eliminates the demand for the new capital that a growing workforce once required. Robert Gordon added the supply-side half of the requiem, arguing that the century of transformative innovation from 1870 to 1970 was a one-time event, not a baseline. Summers has since wavered, telling interviewers that stagnation may have been a phase, and his uncertainty is itself instructive. Even the theory’s modern champion cannot tell whether we are in a pause or a permanence.
That uncertainty is where this essay lives, and I will not pretend to resolve it by assertion. What can be established, and what I have tried to establish here, is the shape of our exposure. The growth assumption is not one belief among many. It is roughly a century old as an explicit paradigm, two centuries old as a lived experience, and it is embedded in the deep structure of pensions, public debt, property, and democratic legitimacy itself, on the left and the right, in republics and autocracies alike. Every modern ideology is a theory of dividing more. We are about to find out what any of them has to say about dividing less, and the honest answer, so far, is nothing at all.
To see why the question is no longer hypothetical, we have to leave the history of ideas and go to the delivery wards of Seoul, the shuttered schools of Japan, and the emptying villages of the Spanish interior, where the arithmetic has already made its decision. That is where we turn next.
The Turn
Every so often, an institution whose job is to be boring says something extraordinary in the most boring possible way. In the summer of 2024, the United Nations Population Division released its biennial World Population Prospects, a document of tables and medium variants that most of the world did not read. Buried in its summary was a sentence that historians may one day treat the way we treat the minutes of some obscure committee meeting where an empire quietly decided to stop expanding. The world’s population, it said, would grow from 8.2 billion to a peak of around 10.3 billion in the mid-2080s, and then decline.
Sit with that for a moment. For the first time in the modern era, the official demographic authority of the human species has put a date on the end of human increase, and the date is within the lifetime of children now in school. And here is the detail that gives the projection its strange authority: each successive revision has moved the peak closer and lower. The 2013 edition assigned only a low probability to any peak this century at all. The 2022 edition said 10.4 billion in 2086. The 2024 edition says 10.3 billion in the mid-2080s. The models are not predicting the future so much as chasing a present that keeps falling faster than the modelers believed possible. When your forecasts err, the direction of the error is information. The direction of this error is always down.
I need to be honest about what kind of claim I am making, because the credibility of everything that follows depends on it. Figures for 2080 or 2100 are projections, model outputs built on assumptions, and models have been wrong before; the mid-century peak could slide by a decade in either direction. But the near-term arithmetic is a different kind of animal entirely, and the distinction is the load-bearing wall of this entire essay. The workers of 2045 have already been born. The mothers of 2050 are alive today, and we can count them. When I say the turn is not a forecast but a schedule, I am talking about the part of the future that is embedded in the age structure of the present, and that part cannot be revised, because the past cannot.
Consider the strangest number in the UN’s report. Global population is still growing and will grow for decades, which is the fact that lulls everyone. But the growth is running on fumes. Of the increase projected through 2054, seventy-nine percent comes not from today’s fertility but from demographic momentum, the built-in growth stored in the youthful age structure created by the high birth rates of the past. The engine has already cut out; the vehicle is coasting on the hill it climbed a generation ago. Fertility worldwide has fallen from 3.31 children per woman in 1990 to 2.25 today, brushing against the replacement threshold of 2.1, and more than half of all countries are already below it. Sixty-three countries containing over a quarter of humanity have already passed their peaks. One person in four now lives in a nation that is shrinking. The world is still filling, but it is filling the way a bathtub fills after the tap is closed, on the water already in the pipe.
The frontier
If you want to see where the rest of the developed world is heading, you go to South Korea, the way you would once have gone to Manchester to see the industrial future. In 2023, Korea’s total fertility rate reached 0.72. It is worth pausing on what that number means, because familiarity has dulled it. Replacement is 2.1. A rate of 0.72 means each generation is roughly one third the size of the one before it. Run that forward and a hundred Koreans of childbearing age become, in two generations, about twelve grandchildren. No war has ever done this to a population. No plague has. In Seoul, a city of ten million and one of the great success stories of the twentieth century, the rate was 0.58.
The rate ticked up to 0.75 in 2024, riding a rebound in marriages postponed during the pandemic, and this uptick was reported in some quarters with genuine relief, which tells you something about how far the goalposts have receded. Korea’s own statistics agency projects the population falling from 51.7 million to 36.2 million by 2072; some researchers put the century-end figure below 30 million. When the man nominated to run Korea’s finance ministry reached for a metaphor, he chose the iceberg that sank the Titanic, and the government has since declared a formal “Population National Crisis” and stood up a ministry devoted to it. This is not a country in denial. It is a country that has looked directly at the arithmetic and found that looking is not the same as changing it.
Japan is the veteran, shrinking since around 2010, projected to fall from 124 million toward 77 million by century’s end, and I will return to Japan later because it has had the longest time to build the institutions of decline. But the case that matters most for the history of this century is China. In 2022, deaths in China exceeded births, 10.41 million against 9.56 million, and the population of the world’s manufacturing heartland began to fall for the first time since the famine that followed the Great Leap Forward. Unlike the famine, this decline has no end date. China’s working-age population peaked all the way back in 2014, at 997 million, and had already fallen to 875 million by 2022. The UN’s medium projection has China’s total population at 633 million by 2100, a loss of more than half, and analyses of the workforce suggest the working-age cohort could fall from around 900 million toward something in the vicinity of 250 million. I want to flag, in keeping with my earlier caution, that these century-end figures are projections and should be held loosely. But the peak is not a projection. The peak is behind us. The country that spent thirty-five years enforcing a one-child policy out of terror of overpopulation now permits three and finds few takers, which may be the purest illustration in existence of the twentieth century winning the wrong war.
And around China stand its neighbors, Taiwan, Hong Kong, Singapore, all clustered at the bottom of the global fertility table, societies that achieved in two generations what Europe took two centuries to do, and discovered that the compression of the miracle compressed its consequence too.
The slow emptying
Europe’s version of the turn is quieter, older, and in some ways more instructive, because it shows what contraction looks like when it has had time to seep into the landscape. Italy recorded 369,944 births in 2024, the fewest in any year since national unification in 1861. Let the comparison register. Italy in 1861 had around 26 million people, no antibiotics, and infant mortality that would horrify us; Italy today has 59 million people and one of the longest life expectancies on earth, and it produces fewer babies. Its fertility rate has slipped to 1.18, beneath even the previous low-water mark of 1995, and nearly a quarter of Italians are now over sixty-five. The country’s response to its emptying villages has become a global curiosity: the one-euro house, offered in Sicilian and Sardinian towns to anyone willing to renovate, a price that is less a bargain than a confession. A house is worth one euro when the market expects no future buyer. That expectation, generalized, is the economics of the subtractive age, and we will meet it again.
Spain sits at 1.10, second-lowest in the European Union, and Spain has given the phenomenon its most politically potent name: la España vaciada, Empty Spain, the hollowed interior where more than half the country’s municipalities now hold fewer than a thousand residents. Emptiness there has become an electoral identity. A movement called Teruel Existe, whose entire platform is the assertion that its province still exists, won a seat in the national parliament in 2019, and by 2022 the Empty Spain coalition was the strongest party in Soria. This deserves more attention than it gets, because it is a preview of a new axis of politics, one that runs not between left and right but between the places people are leaving and the places they are leaving for.
Further east the process runs faster and crueler, because ultra-low fertility combines with emigration, the young departing for Berlin and London and taking their unborn children with them. Bulgaria has fallen from a peak of about nine million in 1988 toward six and a half, and is projected toward three and a half by 2100; the UN ranks it among the fastest-shrinking nations on earth, though the driver there is the compound of exit and mortality as much as birth rates. Ukraine, bleeding from war on top of everything else, fell from 43.7 million in 2020 to 37.9 million in 2024, with fertility around 1.0, the lowest in Europe. Lithuania, Moldova, Albania, Bosnia: the UN’s list of steepest projected declines reads like a map of the continent’s periphery, the places that spent the twentieth century supplying Europe with labor and are discovering that a nation can export its future one departure at a time.
Why the door is locked
Now to the heart of the matter, the reason I keep insisting on the word arithmetic. It would be reasonable to ask: if fertility fell, can it not rise? People change their minds. Cultures shift. Why treat the turn as locked in rather than as a pendulum at one end of its swing?
The answer is that demographic momentum, the same force currently masking the decline, runs in both directions, and in aging societies it has already changed sides. In the countries that have passed their peaks, the number of women of reproductive age is itself now shrinking, projected to fall by a third between 2024 and 2054. The births that did not happen in 2005 mean mothers who do not exist in 2030. Every year of low fertility narrows the base from which any future recovery would have to be built, so that each successive intervention pushes on a smaller lever. And there is a crueler wrinkle still. Suppose a miracle: fertility in Korea or Italy doubles tomorrow. For twenty years, that miracle produces no workers at all. It produces dependents, children to be housed, schooled, and fed by the very working-age cohort that is already stretched thin supporting the old. A fertility rebound, in its first two decades, makes the dependency arithmetic worse before it makes anything better. The corridor between here and any restored equilibrium passes through the hardest years, which is precisely why no democracy has ever voluntarily walked it.
This is what I mean by a schedule rather than a forecast. The contraction of the working-age population over the next twenty to thirty years is not a scenario. It is a census of people already alive, minus the people who were never born.
The two escape hatches
Against this arithmetic, modern states have tried exactly two remedies, and the record of both deserves to be stated plainly.
The first is paying people to have children, and its track record is one of the most consistent failures in the history of public policy. South Korea has spent, by a widely cited accounting, over 270 billion dollars across sixteen years on childbirth incentives, during which time its fertility rate fell from already-low levels to 0.72. Hungary has made pro-natalism the centerpiece of its national project, spending around five percent of GDP on baby loans, mortgage forgiveness, and a lifetime income-tax exemption for mothers of four; its fertility rose from 1.25 in 2010 to 1.59 in 2021, then slipped back, and in a recent year the country recorded roughly 127,500 deaths against 77,500 births. The scholarly literature on Russia, Poland, and Hungary converges on the same finding: cash produces bumps that are real but temporary, and mostly shifts the timing of births that would have happened anyway, pulling them forward without raising the lifetime total. What helps at the margin is the unglamorous whole-of-society package, affordable housing, childcare, labor markets that do not punish mothers. Nothing, anywhere, has restored a developed society to replacement. The demographer Nicholas Eberstadt, asked about Korea, gave the answer that policymakers cannot: he did not think there was any way population policy could raise Korean fertility appreciably. Two hundred seventy billion dollars is a serious experiment. The result is in.
The second remedy is immigration, and here I must tread carefully, because the key document has been so badly abused. In 2000, the UN published a study called Replacement Migration, an analytical exercise, not a policy proposal, though partisans of every stripe have misrepresented it as a blueprint. What it actually did was compute how many migrants would be required to offset aging under various targets, and its findings cut in a direction that surprises both sides of the immigration debate. Maintaining total population size requires inflows within historical experience for many countries; migration genuinely can do that. Maintaining the working-age population requires more, on the order of 6,000 to 6,500 migrants per million inhabitants per year for Germany and Italy, large but conceivable. But maintaining the ratio of workers to retirees, which is the number that actually holds up pensions, requires volumes the report itself called totally unrealistic: 524 million migrants to Japan by 2050, 674 million to the European Union, roughly 13 million people per year arriving in Europe indefinitely. Even the report’s critics agreed that migration at such levels is politically insupportable anywhere on earth. And there is a deeper problem the report could only gesture at: global fertility decline means the sending countries are aging too. Immigration redistributes the world’s young people; it does not create them. It is a mechanism for deciding where the shrinkage lands, not whether it happens. The report’s quietest recommendation, raising the effective retirement age toward 75, is the one governments are actually pursuing, generally without saying so.
The precedent
Has anything like this happened before? Once, and the exception illuminates everything. Between 1347 and 1351 the Black Death killed somewhere between thirty and fifty percent of Europe, the last sustained population collapse in Western history, and what followed is the closest thing we have to a controlled experiment in the politics of less. With workers suddenly scarce and land suddenly abundant, the bargaining power of ordinary labor inverted overnight. Wages surged. Serfs discovered they could walk off the manor and be welcomed elsewhere at double or triple pay; one indignant chronicler complained that the humble now turned up their noses at employment and could scarcely be persuaded to serve the eminent unless for triple wages.
Watch what the incumbents did next, because it is the template. England’s ruling class responded with the Ordinance of Labourers in 1349 and the Statute of Labourers in 1351, which capped wages at pre-plague levels, made it a crime to refuse work or to quit, and branded fugitive laborers. Confronted with a new scarcity, the possessors of the old abundance did not adapt to the new distribution of power; they attempted to freeze the old one in law. It failed, as such attempts tend to, but not before the resentment it generated helped fuel the Peasants’ Revolt of 1381, and within a few generations the labor obligations of feudalism had dissolved into money rents. The lesson is double-edged, and I want to keep both edges sharp for later. Depopulation genuinely raised the value of ordinary human labor, an egalitarian effect that some observers of our own era anticipate with something like hope. And the first instinct of every incumbent order facing that shift was coercion. Eberstadt draws the parallel across the centuries explicitly: for the first time since the Black Death, the planetary population will decline, except that the last collapse arrived on the backs of fleas, and this one arrives entirely through the accumulated choices of free people. Which makes it, depending on your vantage, either far less tragic or far harder to reverse. A plague ends. A preference may not.
The rupture that arrives in pieces
Which raises the final question of this section: if all this is true, why does it not feel true? Why is the largest structural transformation since industrialization living in the back pages, filed under business trends and lifestyle features?
Because it never arrives as itself. The turn has no invasion day, no falling wall, no single legible event for the cameras. It arrives disaggregated, as a school closure here and a maternity-ward consolidation there, as a pension tweak, a labor shortage, a feature about one-euro houses, a quirky statistic about diapers. Each fragment gets covered; the shape they compose does not. Even the UN, as the authors of Empty Planet argued, was slow, running models calibrated to a past that no longer applied, which is why the projections keep catching down. Our institutions are exquisitely tuned to detect the crises of a crowded world, famine, unemployment, housing shortage, and nearly blind to their inversions. The reviewer Zachary Karabell put the asymmetry in a sentence I have not been able to improve on: we are vaguely prepared for a world of more people and utterly unprepared for a world of fewer, and that is the world we are heading toward, fast.
So let us grant the premise its full weight. The turn is real, dated, and past the point of policy reversal on any timescale that matters to the people now alive. The question that remains is the one this essay exists to ask. What actually happens to the machinery, to pensions and public debt, to property and to democracy itself, when every gear in it was cut for a world that added and the world begins to subtract? The unsettling answer is that the machinery does not merely slow down. It runs backwards, and that is where we go now.
Contraction Politics
In March of 2024, a Japanese company called Oji Holdings made an announcement that belongs in the history books, though it ran as a business brief. The firm would stop manufacturing disposable baby diapers that September, after decades in the market, to concentrate on diapers for adults. Its baby diaper output had fallen from a peak of 700 million in 2001 to 400 million; meanwhile, adult diaper sales in Japan have outsold baby diaper sales every year since 2011. There is no better artifact of the age now beginning. A society’s consumer economy is a mirror held up to its age structure, and Japan’s mirror has, for over a decade, reflected a country with more people at the end of life than at its beginning.
I promised at the outset that the mechanisms of modern politics would not merely weaken under contraction but invert their sign, and this final section exists to make good on that claim. The method is simple. Take each of the great load-bearing institutions of the modern order, identify the demographic bet buried in its foundations, and then run the bet in reverse. Do this honestly and a pattern emerges that is more unsettling than any single fiscal projection: instruments designed to pump wealth and security forward through time begin, one by one, to pump it backward.
The chain letter
Begin with pensions, the purest artifact of the growth assumption and the largest financial commitment most states have ever made. On the 24th of May, 1889, the German Reichstag passed Bismarck’s old-age insurance law, the ancestor of every social security system on earth, and its founding arithmetic deserves to be famous. The law set the pension age at seventy, at a time when average life expectancy hovered around forty. Most contributors would never collect. The scheme was, actuarially speaking, a lottery in which nearly everyone paid in and the rare survivor drew out, sustained by an enormous base of young workers beneath a tiny apex of the old. That shape, many payers, few beneficiaries, is not a detail of the design. It is the design.
Now watch the shape invert. When American Social Security matured around 1950, there were roughly sixteen workers paying in for every beneficiary drawing out. By 1960 there were five. Today, per the program’s own trustees, there are about 2.7, heading for 2.1 by century’s end. The 2025 Trustees Report projects the retirement trust fund exhausted in 2033, the combined funds in 2034, at which point incoming revenue covers roughly 81 percent of scheduled benefits and the law, absent congressional action, imposes an automatic cut of about a fifth on every retiree in the country. Note the date. This is not a problem for our grandchildren. It is scheduled within the current decade, and the beneficiaries it strikes are the largest and most reliable voting bloc in American politics, which is precisely why no action has been taken.
And the United States is among the demographically luckiest of the rich countries. The OECD projects that by 2050 Japan will carry roughly eighty retirement-age people for every hundred of working age, and that Korea’s ratio will rise faster than any other, by almost fifty points. Italy, Korea, Poland, Latvia, and Lithuania are all projected to lose more than a third of their working-age populations by mid-century. The pay-as-you-go pension is a chain letter, and I use the term without contempt, because chain letters work magnificently as long as each generation of recipients is larger than the last. Every developed country is now living through the moment the letter reaches the end of the list, and discovering that the last generation holding it is also the one with the votes.
When g stops beating r
The second inversion is quieter and, I suspect, ultimately larger. Modern states run on debt, and the intellectual license for this was always a single inequality: as long as the economy’s growth rate, g, exceeds the interest rate on its bonds, r, a country can roll its debts forward forever, letting growth shrink the burden relative to income without a single dollar of principal ever being repaid. In 2019, at the peak of the low-rate era, Olivier Blanchard used his address to the American Economic Association to formalize the happy version of this logic, arguing that in such conditions public debt might carry no fiscal cost at all. The argument was sound. The premise was demographic.
Because what is g, decomposed? It is workers multiplied by output per worker. Contraction attacks the first term directly and, through the channels Hansen and Summers described, the second as well: a shrinking workforce needs fewer new factories, offices, and homes, which suppresses the investment demand that drives innovation forward. Let g sag below r and the entire logic of sovereign finance flips its sign. Debt no longer melts; it compounds. Sustainability then requires running primary surpluses, taxing more than you spend, year after year, which is to say it requires imposing austerity on an electorate whose median voter is elderly, dependent on the very spending that must be cut, and demographically guaranteed to grow more dominant with each election. The country that has traveled furthest down this road is, again, Japan, which sustains the largest debt burden on earth only by means of a captive financial system in which the central bank and domestic institutions hold roughly ninety percent of government bonds, a nation in effect owing the money to itself and agreeing, collectively and indefinitely, not to ask for it back. It is an elegant arrangement, and nobody knows whether it can be exited, or exported.
The house at the end of the market
The third inversion is the one ordinary families will feel first, because it lives where most household wealth lives: the home. For seventy years, property in the developed world has functioned as a one-way bet, and the bet, stated plainly, was always this: someone younger will pay more. Rising prices were treated as a law of nature, but they were a demographic fact wearing the costume of a financial one. A house appreciates because the queue of future buyers lengthens. Shorten the queue and the costume comes off.
Japan, once more, is the messenger from the future. Its government counts approximately nine million akiya, vacant homes, nearly fourteen percent of the national housing stock, and the Nomura Research Institute has projected that the share could exceed thirty percent by 2038, something on the order of twenty-three million empty dwellings, though that figure comes from a single think tank and should be held with appropriate looseness. In rural prefectures like Wakayama and Nagano, more than a fifth of homes already stand empty. The state now pays families around a million yen per child to leave Greater Tokyo for the countryside, a government bribing its own citizens to redistribute themselves across a landscape they are abandoning. Italy’s one-euro houses and Spain’s entire villages listed for the price of a suburban flat are the same signal at different volumes: a price is a prophecy about future demand, and these prices prophesy no one coming.
What makes this more than a rural curiosity is the wealth transfer now in motion. The largest generation in Western history holds the bulk of its savings in housing and is beginning to bequeath it to a smaller generation, which must sell into markets where, at the margin and outside the superstar cities, the buyer does not exist. Note the cruel geometry: the same young people priced out of thriving metros by scarcity will inherit abundance precisely where it is worthless. Housing wealth, the great middle-class escalator of the postwar era, does not simply stop in a shrinking society. In much of the territory, it reverses, and a family’s largest asset becomes its largest liability, a roof that must be taxed, maintained, and eventually demolished at the heir’s expense.
The electorate of the past
Beneath all three inversions lies a fourth, and it is the one that forecloses the exits. Every remedy for the problems above, later retirement, pension trims, fiscal consolidation, investment tilted toward the young, must pass through democratic politics, and democratic politics is aging faster than the population itself. The median voter is older than the median citizen, because the old vote and the young often do not, and in many systems rural malapportionment amplifies precisely the districts the young have left. The political scientist Samuel Moyn has argued that the deeper phenomenon is not elderly politicians, conspicuous as they are, but elder civic power: a structural condition in which the priorities of the old, by numbers, turnout, and map, simply are the priorities of the system. The empirical literature adds the uncomfortable finding that as cohorts age, their support declines for redistribution toward the young and for long-horizon investment of every kind, including climate.
Assemble the pieces and you have the political physics inversion at its starkest. The modern democratic bargain was struck by young, expanding electorates deciding how to distribute a growing surplus, a game in which nearly every player could win and the future was, demographically speaking, in the room and voting. The bargain now on the table is its mirror: old, shrinking electorates deciding how to allocate sacrifice, with the beneficiaries of the status quo holding a permanent and growing majority, and the future, the unborn and the young, holding almost none of the cards. Recall the Statute of Labourers. When scarcity shifts power toward those who lack it, the first instinct of incumbents is to legislate the old distribution back into place. The gerontocratic version of that statute will not look like branding fugitive serfs. It will look like what it already looks like: benefit formulas that cannot be touched, housing rules that protect incumbent values, debts issued in the names of people too young to object, and retirement ages raised for everyone except those currently near them.
The escape hatches, honestly weighed
An argument this dark owes its strongest objections a fair hearing, and there are three.
The first says technology will save us: fewer workers, each vastly more productive, machines filling the gap that missing people leave. I concede the logic without hesitation, because the logic is sound; g has two terms, and if productivity accelerates enough, the workforce can shrink while output grows. What I cannot concede is the evidence. The productivity acceleration required is enormous, permanent, and must arrive against the grain of the secular-stagnation record, which shows productivity growth slowing precisely as populations age. China, the world’s most aggressive adopter of industrial robotics, is the test case, and analysts of its trajectory doubt automation can fully offset a workforce contracting by hundreds of millions. Machines may yet soften the landing, and I hope they do. Softening a landing is not the same as repealing gravity, and a civilization betting its pension system on an unproven acceleration is not planning; it is praying.
The second objection says fertility may rebound, and here too something must be conceded. The number can move; Korea’s small uptick proves it, and certain religious subcultures sustain high fertility inside otherwise barren societies, a fact that may matter enormously over centuries. But no developed nation has ever returned durably to replacement, the pool of potential mothers is itself now shrinking, and, as established earlier, even a miracle rebound spends its first twenty years making the dependency arithmetic worse. A rebound is a hope for the twenty-second century. The politics of this one are already written.
The third objection is the most interesting, because it is not an objection to the facts but to the mourning. Dietrich Vollrath has argued that slow growth is a trophy, not a wound: roughly two-thirds of America’s slowdown, on his accounting, traces to rising living standards and women’s reproductive freedom, the twin successes of the modern project. A smaller, richer, older society, on this view, is what winning looks like. And the environmental ledger backs him further than my argument might suggest. Fewer people means less carbon, less extraction, more land returned to forest; even the UN permits itself to note that China’s coming shrinkage has obvious positive implications for sustainability. The degrowthers of the 1970s asked for a smaller human footprint, and they are getting one.
I do not think this objection is wrong. I think it is half of a truth whose two halves refuse to reconcile, and that refusal is the deepest thing this essay has to say. Population decline is simultaneously the answer to the twentieth century’s ecological prayer and the solvent of the twentieth century’s political order. Both are fully true at once. The biosphere and the pension fund are keeping different books, and the same entry that is credited in one is debited in the other. This is what tragedy means in its strict, older sense: not sadness, but a collision of goods. The environmentalists were right that the planet needed fewer of us. The economists were right that everything we built assumed more of us. The subtractive age is the place where both bills come due together, and anyone selling it to you as simply good news or simply catastrophe is reading only one ledger.
The early institutions of less
What, then, does the politics of the subtractive age actually look like? Not, if the early evidence holds, like collapse. Japan, two decades into national shrinkage, remains rich, orderly, safe, and pleasant, a standing rebuke to the catastrophists, and Nicholas Eberstadt, no optimist, concedes that depopulation is not a death sentence but a difficult new context in which countries can still thrive. What Japan shows instead is the texture of managed decline: roughly 450 schools closing every year, nearly nine thousand shuttered in less than two decades, their gymnasiums converted to community centers for the old; akiya banks matching abandoned houses to the occasional dreamer; villages consolidated, train lines pruned, an entire administrative craft emerging around the graceful shutdown of places. Korea has declared its situation a national crisis and built a ministry to face it. China has begun, for the first time in decades, raising its retirement age. These are the first institutions of a politics that no longer assumes more, and they share a family resemblance: they are all, in essence, techniques of subtraction, the statecraft of deciding what to close, and in what order, and who must consent.
That craft will need a politics worthy of it, and none of our inherited ideologies supplies one, for the reason established at the very beginning: all of them are theories of dividing an increase. The genuinely open question of the coming century is what fills that vacancy. The pessimistic answer runs through Benjamin Friedman’s bicycle: stalled societies fall toward intolerance, scapegoating, and the coercive instinct of the Statute of Labourers, and a politics of allocating sacrifice among the shrinking, refereed by the old, is a machine for manufacturing grievance. The hopeful answer runs through the other edge of the Black Death’s lesson: scarcity made ordinary labor precious, and a world short of young people is a world that must finally, structurally, value them, bidding up not just their wages but their bargaining power, their housing claims, perhaps their votes. Both futures are latent in the same arithmetic. Which one emerges will depend on whether any political movement, anywhere, learns to say honestly what none has yet dared: that the surplus is ending, that the promises calibrated to it cannot all be kept, and that the measure of a civilization was never how it divided its abundance, but how it divides its subtraction.
For two hundred years we have been the species of the ascending curve, and we built everything, our pensions and our parliaments, our mortgages and our meanings, on the slope of it. The curve is bending now, not because we failed but because we succeeded, and the bend will outlast every government currently in office and every reader of this sentence. The question it puts to us is not whether we can restart the ascent. We almost certainly cannot, and the decades of pretending otherwise will be remembered as the most expensive denial in fiscal history. The question is older and harder, the one Ibn Khaldun would recognize across seven centuries: whether a civilization that owes everything to growth can learn to be worthy of its own success, and to govern, with something like grace, the long exhale that follows it.
Magnificent piece of writing. Thank you!
This is the most thoughtful and thorough examination of economic and demographic and environmental trends I’ve ever read.