Rent and Its Discontents: How we turned our Economy rent-seeking and how AI might reverse it
When the pie grows for everyone, politics is about sharing it. When it grows only for the few, politics is about seizing it.
Something has gone wrong in the rich democracies, and the strange part is how widely it is felt and how poorly it is explained. Over the past decade, and far more steeply over the past five years, people across almost every Western country have grown more polarised, more separate, and, underneath it all, less happy. Ask why and you are handed a catalogue: inequality too wide, immigration too fast, corruption too deep, a governing class too far from the governed. Each answer has its partisans and its supporting chart, and each is held with the heat we reserve for explanations that feel almost complete.
They are almost complete in the way a fever is almost an account of an infection. They are real, and they are not the thing; they are what the thing looks like from the inside. A real explanation would have to say why all of it arrived together, on the same schedule, across countries that share little except a level of development. When the same symptoms appear at once in populations that agree on nothing else, the cause is rarely in the politics, where the argument is loudest. It is underneath, in the soil. And the thing that changed in the soil is what our economies do for a living.
For most of the postwar period the rich countries made things. They poured concrete and strung wire and built the grids and machines on top of which ordinary work grew more productive year after year. Capitalism, less an ideology than a ruthless method for raising output, was pointed at production, and the gains came out as more: more goods, more houses, cheaper nearly everything, wages that rose because an hour of labour was worth more each year than the last. Living standards climbed for almost everyone, because the surplus that production throws off is hard to keep in one place. It leaks, into wages and prices and the plain fact that a normal life buys more than it used to. Call this a productive economy: one whose central political fact is that almost everyone expects next year to be richer than this one.
Now look at what the economy has become. A rising share of it no longer makes anything; it collects.
Rent, in the old and exact sense, is income you draw not from making something but from standing between other people and something they need and cannot easily get around you. It is the toll at a chokepoint. Its size has nothing to do with how much you add and everything to do with how hard it is for anyone else to do what you do. A maker is paid for what he produces; a rentier is paid for what he controls. And by almost every measure, the controlling has been winning. The share of the economy paid out to people for their work has fallen by around five points since 1980, while the share taken as profit has risen by about as much; corporate profits that ran near six per cent of national income after the war now run above ten. Finance, insurance and property were a tenth of the economy in 1950 and are a fifth today. Health care has gone from a twentieth of output in 1960 to nearly a fifth. The share of jobs you may not legally do without a licence has climbed from one in twenty in the 1950s to roughly one in four. Half of all renters now hand over more than thirty per cent of their income for shelter, the highest on record. Economists fight over each figure, but the fight is always about the size of the move, never its direction. Every gauge tilts the same way.
Advertising is the most visible of these tolls and the least of them. The advertising bill has been a roughly constant share of the economy for a century; what changed was its concentration and its aim, not its size. The real mass of the skim was always in the older, duller chokepoints no one livestreams: the zoned lot, the licence, the financial toll, the patent, the cornered acre of land.
These tolls are not new, which is the first thing to explain, because the crisis feels new. The lines on most of those graphs bend in the 1980s, not the 2010s. What is new is the moment the rent caught up with the growth. A rising toll does no visible harm while the economy still grows faster than the toll climbs; people get their raise, only on a base quietly smaller than it should have been. The pain begins at the crossing point, the year the toll rises fast enough to swallow the whole of the growth and the raise stops coming. A society can absorb almost any extraction while it is still, on net, getting richer. What it cannot absorb is the year that stops. We crossed that line in the last decade, which is why a trend forty years old reached our politics as a sudden storm.
This is why the shift is not an economic curiosity but the engine of the unhappiness. A positive-sum economy breeds a particular politics. When the pie is growing, my gain does not require your loss; there is more next year by default, and the argument is only over how the more is shared. Cooperation pays, trust pays, fair dealing is an asset, because the stakes keep rising and you want to stay in the game. A quarrel over the division of a growing thing can afford to be generous, because no one is asked to accept less than he has.
A rentier economy is closer to zero-sum, and breeds the opposite. When the growth no longer reaches you, when the gains are skimmed in transit rather than added at the end, my gain begins to require your loss. No surplus arrives to paper over the conflict. Every transaction takes on an adversarial edge, because if the toll-taker is cutting from both of us, whatever I keep is something you did not get. Cooperation stops paying and positioning starts; trust becomes a liability. The society does not decide to turn suspicious and competitive; it is made so by the arithmetic it lives inside.
Look back now at the catalogue we began with, and watch the items resolve into expressions of this one shift.
Housing is the clearest, the one people feel in their bodies. A house was shelter; it has become a positional asset, the main store of wealth for those who got in and the main barrier for those who did not. The politics around it is naked rent-defence: incumbents block new building not out of malice but because supply would lower the value of the thing they sit on, and so the young are priced into a permanent antechamber while being told the door is open. Zero-sum made physical.
Immigration changes meaning depending on which economy you stand in. In a productive one, newcomers grow the thing, more workers and more demand, and the politics tends towards welcome. That was the logic on which the doors were opened, and it was sound for the economy we thought we still had. But in a rentier economy more hands do not enlarge the share that reaches people; they divide it. The newcomers arrived on the old promise and landed in the new arithmetic, where the aggregate rises while the slice per person stalls, so that people who would once have been adding now register as rivals for a fixed stock of housing and seats. The fight did not come from a change in human nature; it tracks the moment growth stopped reaching the people already here.
The rest run the same way. The credential race is what education becomes when good outcomes are fixed: a degree stops forming a capable person and becomes a ticket in a queue, valuable mainly because others lack it. The gulf between elite and everyone else widens because the very rich now get rich by owning the toll booths rather than building what the rest of us use, and people feel the difference even when they cannot name it. And the loss of trust beneath all of it is the residue: trust is a bet that cooperation will be repaid, rational in a positive-sum world and a sucker's move in a zero-sum one. The wariness everyone laments is not a moral failing of the age but an accurate reading of the incentives.
So the thesis, narrowly: the disease is not inequality, or immigration, or corruption, or the distance between rulers and ruled. It is the shift from an economy whose growth reached almost everyone to one whose growth is taken as rent before it arrives. Notice the precise form: the pie has not stopped growing, the headline numbers keep insisting it grows, but it grows into the hands that own the tolls, and what reaches everyone else stops moving. A society need not shrink to feel zero-sum; it need only stop the growth reaching it. The symptoms are real. They are just symptoms, the temperature and not the illness.
Which brings me to the technology everyone is arguing about, and to the only hopeful part of this essay.
Can a rentier turn even be reversed? Nothing so far suggests it. Rents do not loosen out of conscience; they rest on barriers, and a barrier does not fall because we disapprove of it. The only thing that has ever reliably dissolved a rent is a new way for more people to do the thing the rent was charging for, some capability that floods the narrow place the toll-taker was guarding until there is nothing scarce left to stand astride. Rents die when the scarcity behind them dies. So the question is whether anything dissolves scarcity, and here, for once, there is an answer, though a precise one, and the precision is the whole of the honesty in it.
There are two kinds of scarcity beneath the world's rents. The first is the scarcity of capability: the rent you charge because the thing is hard and few can do it, write the contract, read the scan, build the software, audit the books. It is real, but it is not fixed; it is only expensive to reproduce. The second is the scarcity of the physical and the legal: the rent you charge because the thing genuinely cannot be copied, or because the law forbids its copying, the acre of land, the zoned permission, the licence that caps how many may practise. The first kind of scarcity is a cost. The second is a wall.
Artificial intelligence is, before anything else, a machine for manufacturing capability. It takes what was hard and rare and makes it cheap and abundant. So it goes straight at the first kind of rent and does nothing whatever to the second, and the whole hope, properly chastened, lives in that asymmetry.
On the capability rents the news is good and large. The agent that writes the code is not an intermediary charging two parties for a meeting; it is a producer, and it produces where a barrier used to stand. A person who could not build a thing can now build it; the buyer can become a seller. And every profession whose fee was part skill and part cartel is about to learn how much of each it was charging. The lawyer's bill, the consultant's, the radiologist's, was always two things fused, genuine scarce expertise and a closed door, and because they came in one invoice no one could see the seam. Make the expertise abundant and the seam appears: the part of the fee that was skill falls towards the cost of the machine, and the part that remains is the door.
This is the strange gift, and not the one the optimists advertise. Artificial intelligence does not abolish rent; it unmasks it. It competes away precisely the rents that were doing real work, the ones resting on capability genuinely scarce, and leaves standing, naked and suddenly visible, only the rents that were pure barricade: the licence with nothing but law behind it, the lot with nothing but location behind it. For a century these hid inside the same fee as the scarce skill, and you could tell yourself the cost was the difficulty. Strip the difficulty out and what remains can no longer pretend. The machine performs, as a side effect, the most honest audit a rentier economy has ever sat through, telling us which tolls were paying for something and which for nothing but the toll.
That is more useful than it sounds, because it turns a fog of unfairness into a short and specific list. A society cannot fight a vague sense of being skimmed, but it can fight a zoning code, a licensing board, a patent term, a financial chokepoint held open by regulation. These are not laws of nature; they are votes, and votes can be taken again. It may even get worse first, since the newly rich will pour their winnings into what cannot be copied, land above all, so capability rents fall while land rents climb to meet the money. But even then the deeper change holds, because a rent you can see is one you can tax or break.
One last objection, the one I have made myself: why would AI not become the next rentier, moating the best models and taxing cognition as the feed taxed attention? Because this rent is of the breakable kind. No one pays ten times the price for a marginally better model when an open release months behind does the same work for a tenth of the cost; where the rent survives it slides to the chipmakers and the clouds, who will be rich but are weaker landlords than the last. The attention giants were unassailable because what they had cornered was a network of people, and there is only one set of those. A chip is not a network: its supply expands, a rival builds a better one in private, and buyers switch, because nothing holds them in place. And beneath it all, unlike beneath the feed, something is genuinely made. A heavily taxed productive economy is still productive: a toll on a road that is carrying traffic.
I owe one last honesty, having spent a whole essay on it. The same machine that floods the world with capability can, in the single mind, rot the faculty that made it worth having; an agent that writes the code spares you ever learning to, and a muscle spared long enough goes soft. But the timescales are, for once, on our side. The gains bank: the artifacts stay built, the firms do not un-start themselves because their founders grow lazy. Whether we compound them we answer ourselves, by keeping the hard part where it is what keeps us capable. Even the pessimistic case, ten or twenty years of real production before the bill comes due, beats the rentier decade we are leaving. The optimistic case is a different century.
So the choice is narrower and more hopeful than the ones we are usually handed. For forty years a rising share of everything we made was diverted to the people who owned the narrow places the rest of us had to pass through, and when the diversion finally outran the growth and the raises stopped, we turned on each other and called it a culture war, an immigration crisis, anything but the plain arithmetic of an economy that had stopped letting its gains reach the people inside it. The next machine will not, on its own, hand those gains back. But it is the first force in a long time pointed the right way: it makes the scarce thing abundant, competes the disguised rents to death, and drags the rest, the land and the licence and the law, into the light where we can decide whether we still consent to pay.
That is not salvation. It is the possible return of the conditions under which people can afford to be generous with one another, because each has reason to expect to be richer next year, together with a short and honest list of the barricades still in the way. We had those conditions once and gave them up without noticing. We might, with some care and some luck, be about to get them back. Only let it make things, and keep enough of our minds intact to know the difference, and to remember which of the walls we built ourselves.