Growth Lasts Only as Long as We Let Incumbents Lose

On October 13, the Royal Swedish Academy of Sciences^[1] awarded the 2025 economics Nobel to Joel Mokyr, Philippe Aghion and Peter Howitt “for having explained innovation-driven economic growth.” Growth lasts only while a society lets the losers from each new idea lose. A country that makes a habit of shielding incumbents is steering back toward the stagnation that ruled most of human history. Government’s job is to keep markets open to new competitors and to help the people displaced along the way get to the next job, not to preserve the old one.

The escape from stagnation is recent. In the Maddison Project Database,^[2] England’s output per person roughly doubled from 1300 to 1700, a pace of less than 0.2% a year. From 1800 to 2022, output per person, in 2011 dollars at purchasing-power parity, grew more than 11-fold in Britain and more than 30-fold in Sweden, compounding at about 1.1% and 1.6% a year. The early figures are reconstructions, so treat their levels as approximate.

GDP per person in the United Kingdom and Sweden, 1300 to 2022

Each half of the prize explains part of the break. Mokyr showed that before the Industrial Revolution, in the words of the committee’s popular science summary,^[4] “People knew that something worked, but not why,” so discoveries rarely built on one another. Once science could explain a technique, each improvement became the base for the next; in the committee’s example, steelmaking advanced after people understood how oxygen strips carbon from molten pig iron. Another condition, he argued, is a society willing to let the new displace the old. Many were not. In 1561, as the committee’s scientific background^[5] recounts, Nuremberg’s city council, pressed by the guild of red-metal turners, went after Hans Spaichl, a coppersmith whose improved lathe threatened the guild’s trade. Aghion and Howitt built that conflict into a 1992 model in which each successful innovation destroys the incumbent’s rents. The same document warns that “entrenched incumbents will have a tendency to block entry and hinder growth possibly with the help of government policies.” Keeping markets open therefore takes active policy: stopping monopolies from shutting out new competitors, and refusing incumbents the tariffs and rules that would do it for them.

The best case against me starts with the people who lose. In the U.S. local labor markets most exposed to Chinese imports, David Autor, David Dorn and Gordon Hanson^[6] found that wages and labor-force participation stayed depressed, and unemployment elevated, “for at least a full decade.” Churn looks like a feature from a desk and like a lost decade from a factory town. Security is the second objection. On October 9, China’s Commerce Ministry announced^[7] that from December 1, companies abroad would need its license to export rare-earth magnets and certain other products made outside China to destinations outside China if Chinese rare-earth materials on its list made up 0.1% or more of their value. China mined more than two-thirds of the world’s rare earths in 2024,^[8] by the U.S. Geological Survey’s estimate. Growth is not the only goal a country may rightly pursue. I accept both points, and neither justifies shielding incumbents in general. A security case covers a short list of inputs a rival power could cut off. Washing machines are not on it. Yet the 2018 safeguard tariffs on washers, which followed a petition from Whirlpool, cost consumers just over $1.5 billion a year in higher washer and dryer prices, Aaron Flaaen, Ali Hortaçsu and Felix Tintelnot^[9] estimate. Net of tariff revenue, that comes to more than $800,000 a year for each of the roughly 1,800 jobs manufacturers announced in response. At that price, a tariff is a poor way to give anyone a job.

Helping displaced workers directly is worth trying, though I will not pretend that retraining gives a factory town its lost decade back. Paying people to go back to school or learn a new trade costs real money, and I know of no perfect design. People want to be productive, and a society that leaves them idle wastes them. The committee’s popular science summary^[4] suggests “Protecting workers but not jobs,” and its scientific background^[5] describes the policy: insure workers whose jobs are destroyed, and retrain them for expanding firms. The same document expects AI-driven change to create many losers, at least in the short run, and calls helping them change jobs “while not hindering the transition” an important challenge for policymakers. Nuremberg’s council faced that choice in 1561 and sided with the guild. Two centuries of compounding make the case for the lathe.