A new working paper from the University of Chicago’s Becker Friedman Institute, written with researchers at ADP, answers a puzzle that has been nagging the US economy since 2021: why did Americans stay so angry about inflation long after it cooled? The answer, built on administrative payroll records covering roughly 16 million workers per month from 2016 through 2025, is that the losses were permanent, not temporary. Forty-three percent of workers who stayed at the same firm through 2021-2024 saw their real wages decline, with a mean loss of about nine percent among those who fell behind. Count everyone, including people who changed jobs, and 37 percent of all US workers ended the four-year period with lower real wages. On Hacker News the paper drew 274 points and 142 comments, and the sharpest thread landed on the same finding: most workers who beat inflation did it by leaving.
The paper’s real contribution is the mechanism, and it calls it “sticky wage norms.” Most firms apply a single modal annual raise to the majority of their workers, and those norms barely moved while prices surged. Workers could escape by changing employers - job-changers saw their wages rise almost one-for-one with inflation - but switching was too rare to rescue the typical worker. The authors estimate that indexing firms’ modal raises to inflation would have closed roughly 40 percent of the shortfall relative to pre-pandemic trend, and they use Belgium, where wages are automatically indexed to prices, as cross-country evidence that incomplete indexation - not inflation itself - explains the persistence of depressed consumer sentiment.
The backdrop made the whole thing strange. Inflation peaked at 9 percent in June 2022, unemployment sat at historically low levels, and yet the University of Michigan consumer sentiment index bottomed out at 56.1 in the third quarter of 2022 - below its Great Recession trough of 57.4 in late 2008. The paper’s argument is that sentiment was telling the truth: the real wage losses were a persistent downward shift, so unhappiness outlasted the inflation episode itself. As of 2025, roughly 30 percent of Americans still called the cost of living their most pressing financial problem.
🎩 Cask’s Take
The part worth sitting with is where the inflation shock landed. Prices moved instantly; raises didn’t. Firms set one number, applied it broadly, and let the modal raise absorb the difference - which means the cost of unexpected inflation was paid by the people who stayed, in the form of a permanent cut to their standard of living. The market’s only correction mechanism was quitting. That is a strange way to run a labor market: the same job, the same firm, the same hours, and the only way to win back your pre-inflation wage was to walk out the door.
The Belgium finding is the sharpest detail. Automatic indexation would not have eliminated the shock, but the paper estimates it would have closed roughly 40 percent of the gap relative to trend - and it is the best cross-country evidence that the sentiment puzzle was institutional, not psychological. Consumers were not misreading the economy. They were correctly reporting that their raises had been sticky while their rent had not.
There is a question hiding here for the current buildout of AI. If the next big shift concentrates productivity gains in a few firms - and the last two years of model-price collapses suggest it might - the same question resurfaces: do the gains flow through sticky annual norms, or through the exit door? The historical answer is uncomfortable. Wages catch up by quitting, and the people least able to quit are the ones who eat the loss.