Payroll records covering roughly one in seven US workers show that during the recent inflationary period, most firms kept annual raises around 3%, causing real wages to fall behind rising prices and, for many workers, never fully recover.
The US economy has largely bounced back from the COVID-19 recession. Inflation has receded from its 9% peak to 3.5%, and unemployment now sits near a historic low. Consumer sentiment, however, has yet to rebound. Nearly 30% of Americans still report that the cost of living is their most pressing financial problem, and over 60% say inflation is a very big problem — even as price growth has returned to normal. What drives this disconnect — the “vibecession,” or “permacession,” as some are calling it? Why does sentiment remain poor when the economy, on paper, has recovered?
In this paper, the authors bring new data to bear on this puzzle. They use a sample of administrative payroll records from ADP, a global leader in HR and payroll solutions, which covers roughly 16 million US workers per month. The authors use these data to track the monthly contracted rate of pay for millions of workers from 2016 through 2025, spanning the pre-pandemic years, the inflationary surge, and its aftermath.
The authors open with Figure 1, above, which plots two indices of US real wages from 2017 through 2025, one computed using data from ADP and one using data from the Current Population Survey. The two lines track each other closely, demonstrating that the ADP data are representative of wages dynamics in the broader labor market.
As you can see, real wages fell during inflation as rising prices eroded consumers’ purchasing power. Although real wages resumed growing at their regular pace rate by mid-2023, they did so from a permanently lower base, leaving the typical worker’s real pay well below the path it had been on before 2021.
Building on this result, the authors proceed to use the ADP data to uncover what drives these wage dynamics — how firms set pay, how workers who stayed in their jobs fared against those who left, and, ultimately, why declining real wages, rather than inflation itself, appear to explain why public sentiment has remained so low. Their key findings, which are detailed in Sticky Wage Norms and the Real Wage Cost of Unexpected Inflation, follow.
FINDING 1: Job-stayers’ wages did not keep pace with inflation.
During 2017–2019, workers who remained at the same employer for at least 13 months (whom the authors term “job-stayers”) experienced median annual real wage growth of roughly 1%. This fell to about −4% in early 2022.
This marked decline in real wages resulted from the fact that nominal wages hardly budged as prices rose. As you can see in Figure 2, below, the distribution of annual nominal wage growth shifted only modestly during 2021–2023, before returning to its pre-pandemic shape by 2024–2025.
Some workers did edge into higher raise brackets: the share of job-stayers receiving a small increase (between 0 and 4%) fell from 40.1% in the pre-pandemic years to 28.4% during the high-inflation period. At the same time, however, nearly half of all job-stayers saw nominal wage changes of less than 4%, despite annual inflation greater than 7%.
FINDING 2: Forty-three percent of job-stayers had lower real wages after four years.
Because job-stayers’ wage growth fell behind the rate of inflation for multiple years, the effects compounded over time and added up to considerable losses. Figure 3, below, shows the distribution of cumulative real wage changes over two four-year windows, before and during the pandemic.
Forty-three percent of workers who remained at the same job from December 2020 to December 2024 had lower real wages at the end of the four years than they did at the outset, roughly double the 21% who fell behind over a comparable pre-pandemic window. Among those who lost ground, the typical decline was about 7% in real terms, versus about 4% before the pandemic. And even many workers whose real pay technically rose still ended up below where they would have been had their pay kept rising at the 1–2% a year that workers typically gain from added experience and seniority.
FINDING 3: Raises cluster at round numbers, especially 3%, and most workers at a firm get the same standard raise.
Why did nominal wages move so little? Part of the answer lies in how firms set pay. The data reveal that, instead of fine-tuning raises worker by worker, firms tend to apply annual, round-number increases, most often 3%. Fourteen percent of annual wage changes ranging from 1% to 6% granted between 2017 and 2019 were exactly 3.00%, while 21% came within a tenth of a point of 3%, and 42% landed on an exact whole or half number.
Those conventions are firm-specific and tightly applied, such that the size of a worker’s raise depends heavily on where they work. Among workers who received a single raise during the year, 60% came within half a percentage point of their firm’s norm (its most common raise), and 85% within 1.5 points.
FINDING 4: Firm wage norms changed little amid inflation.
When inflation surged, these norms barely moved. Figure 4, below, shows the distribution of firm norms across three periods. Even at the peak of inflation when prices were rising over 7% a year, 76% of workers were still at firms with a 2-4% norm, down only modestly from 89% before the pandemic.
FINDING 5: Off-cycle raises were more common during inflation.
While firms tended not to adjust their wage rules in response to inflation, they did offer workers more off-cycle raises. The share of workers whose pay increased during a month other than their firm’s main review month rose from about 17% before the pandemic to about 27% during the inflation, and these off-cycle raises were both larger and grew faster than standard on-cycle increases.
FINDING 5: More workers changed jobs, and those who did kept pace with inflation.
One dependable way to escape firms’ sticky norms and keep pace with inflation was to change jobs. As Figure 5, below, shows, job-changers’ pay stayed ahead of inflation throughout the inflationary period, while job-stayers’ fell behind, dropping to about −4% real at the 2022 peak. Because job-changers’ nominal pay rose in step with inflation, their real wages kept climbing at roughly their usual pace instead of eroding. And more workers chose to switch during the inflation, with monthly job-switching rising from about 2.1% before the pandemic to about 2.5% at inflation’s peak in 2022.
FINDING 6: Even counting job-changers, 37% of all workers fell behind.
While job-changers fared better, their gains were not enough to keep the workforce as a whole from falling behind. Figure 6, below, plots the distribution of total real wage change over a full four-year window for a sample of US workers representing both job-stayers and job-changers.
As you can see, 37% of all workers ended the four years with lower real wages than they started, compared to 24% before the pandemic. This share is only slightly lower than the 43% of job-stayers that experienced real wage declines during the 2021-2024 period, suggesting that incorporating job-changers does not meaningfully alter the findings.
These absolute losses understate the full cost of inflation on wages, as 58% of workers failed to achieve the real wage growth they would have realized had pre-pandemic annual gains continued.
FINDING 7: As real wages fell, corporate profits rose by roughly the same amount.
The real wages that workers did not receive resurfaced, in roughly the magnitude they were lost, as higher corporate profits. As Figure 7, below, shows, the corporate profit share of GDP jumped from a steady 11.4% before the pandemic to 13.1% during the inflation.
This rise of about 1.7 percentage points brought corporate profits to their highest sustained level in half a century, and it closely matches what the arithmetic would predict. Because wages are about 60% of output, a real wage shortfall of 2 to 4 percentage points relative to trend should lift the profit share by roughly 1.2 to 2.4 points, and the observed 1.7 sits squarely inside that range.
The authors are careful to call this a consistency check rather than a precise accounting, but the magnitude and timing align closely enough to suggest that sticky wage norms quietly transferred income from workers to firms during the inflation.
FINDING 8: Had firms indexed their standard raise to inflation, much of the damage would have vanished.
Findings 3 and 4 show that firms kept their standard raise near 3% even as inflation climbed. How much of workers’ lost ground is owed to that norm? The authors answer with a counterfactual, asking what real wages would have looked like if firms had tied their norm to inflation (and left everything else unchanged). They find that indexing firms’ raises one-for-one to inflation would have closed roughly 40% of the resulting shortfall relative to pre-pandemic trend.
Taken together, these findings suggest that low consumer sentiment is the result of lasting real wage losses generated during the recent inflationary episode. The authors confirm this by comparing consumer sentiment in Belgium, where wages are automatically indexed to inflation, to sentiment in Germany, the Netherlands, Denmark, and the broader Eurozone.
When Belgium and its neighbors faced the same inflation, labor market conditions, and shocks, Belgium’s real wages rebounded to pre-inflation levels by 2023 due to indexing, and consumer sentiment recovered with them.
By contrast, in the peer countries, where wages adjusted slowly and incompletely, neither real wages nor consumer confidence had recovered by the end of 2024. This story suggests that real wage losses rather than inflation itself, can explain the sharp and lasting fall in consumer sentiment during the 2021–2024 period.
The declining purchasing power of US workers helps explain why consumer confidence remained persistently low even as inflation returned to normal and unemployment has stayed historically low. Because firms’ wage norms are sticky, a temporary burst of inflation generated a persistent downward shift in the level of real wages, so the loss to workers was far longer-lived than the inflation itself.
This makes inflation more costly to workers than standard models imply, and the real wage losses understate the full cost. The workers who kept pace did so by searching for new jobs, bearing the disruption of switching, or negotiating against a reluctant employer, real resources spent simply to defend their wages, while the workers who did not take those actions simply fell behind. Even workers who “kept up” with inflation, in other words, were made worse off by it.
