Key takeaways:

• Real wages declined 0.3% for low-wage workers in 2025, a stark departure from the unusually strong wage gains they had experienced over the previous five years.
• This reversal was not inevitable—it was caused by policy decisions that weakened the labor market.
• Meanwhile, middle- and high-wage workers saw modest wage growth in 2025.
• Low- and middle-wage workers have suffered from decades of slow and suppressed wage growth. To improve affordability, policymakers can and must raise wages. 

Low-wage workers saw their real (inflation-adjusted) wages decline in 2025, a sharp reversal from the historically fast real wage growth they had experienced over the previous five years. Middle- and high-wage workers continued to experience modest gains in 2025, according to our new analysis (see Figure A).

We examine wage growth across deciles, using the Current Population Survey (CPS) Outgoing Rotation Group microdata. Note: Due to the federal government shutdown and a lack of funding at the Bureau of Labor Statistics (BLS), there are no CPS wage data for October 2025. That means that 2025 wages are calculated based on reported wages from the other 11 months of the year.

Figure A

A weakening labor market halted low-end wage growth in 2025

In 2025, the 10th-percentile wage—the hourly wage at which 10% of workers are paid less and 90% of workers are paid more—fell 0.3% to $14.56. The median wage—the wage at the middle of the wage distribution—grew 0.8% to $25.67 in 2025 while the 90th-percentile wage increased 0.4% to $64.52 (see Figure B).

Figure B

These outcomes are a stark departure from the post-2019 pattern in which low-wage workers consistently experienced faster real wage growth than those in the middle and upper parts of the wage distribution. During that period, policymakers engineered a fast and full recovery from the pandemic recession, which provided unusual leverage to low-wage workers as employers scrambled to hire or rehire the workers they lost in the pandemic. Low-wage workers were able to secure historically fast real wage growth, despite the pandemic- and war-driven inflationary spike in 2021–2022.

But in 2025, a softening labor market halted low-wage workers’ progress. The Trump administration chaotically imposed historically high tariffs, conducted cruel mass deportations, and implemented massive layoffs among federal agencies that provide key inputs to private-sector economic growth and security. All of this led to increased economic uncertainty, and promised growth failed to materialize—particularly in areas such as manufacturing employment.

Payroll employment growth slowed notably as average monthly job gains fell from 168,000 in 2024 to only 49,000 in 2025. The average unemployment rate ticked up from 4.0% to 4.3%. By December 2025, the unemployment rate stood at 4.4%, fully a percentage point higher than the 3.4% low point reached in April 2023. Groups that are often affected first by an economic downturn—such as young workers and Black workers—experienced a much faster uptick in unemployment. Most concerning is the depressed hires rate, which is currently at levels similar to 2013 when the economy was still recovering from the Great Recession. Should layoffs pick up even a bit, this low hires rate would see the unemployment rate rise quickly.

The 2026 outlook for wages remains uncertain. While higher unemployment and depressed hires point to a cooling labor market, it is unclear that conditions will continue to deteriorate: Unemployment insurance claims remain stable and consumer spending has held up. At the same time, economic risks loom large, including geopolitical instability, Trump’s immigration policy and cuts to social programs, and the possibility of a stock market and investment collapse if the AI-driven stock market bubble deflates rapidly.

Wage inequality has declined since 2019, but low-end wage levels are still insufficient to make ends meet

Even with last year’s decline, the 10th-percentile wage of $14.56 represents a significant improvement from 2019 in inflation-adjusted terms. And there was still substantial wage compression between 2019 and 2025, as wages at the 10th percentile grew twice as fast (15.0%) as wages at the 90th percentile (7.4%). These findings are consistent with economist Arindrajit Dube’s recent Substack article, which documents strong wage compression over this period. Importantly, Dube shows that these patterns persist even after controlling for compositional changes in the population, meaning that faster wage growth at the bottom is not explained by shifts in worker demographics such as age, education, or gender.

But this low wage is still far from sufficient to make ends meet. Even if that 10th-percentile worker worked full time throughout the year, their annual pay would only be $30,279—which is not enough to attain a modest yet adequate standard of living in any U.S. county or metro area, according to EPI’s Family Budget Calculator.

Low-wage workers have seen little wage growth for much of the last 50 years

While the gains over recent years are welcome, longer-term trends show lower-wage workers losing ground. Figure C shows wage growth at the 10th, 50th, and 90th percentiles from 1979 to 2025. Over this period, 90th-percentile wages grew by an average of 1.1% per year, compared with just 0.5% at the 10th percentile. In fact, it wasn’t until 2015 that lower-wage workers finally reliably surpassed their 1979 real wage. If wages at the 10th and 50th percentiles had grown at the same rate as the 90th percentile since 1979, they would have been $18.58 and $32.40, respectively, or about 27% higher. For all wage deciles over time, including data by demographic characteristics, visit the EPI data library.

Figure C

While workers at the 90th percentile have benefited more than lower-wage workers, their gains pale in comparison to those at the very top of the distribution. The gulf between the lowest- and highest-wage workers cannot be fully captured in the CPS data because it isn’t possible to accurately measure what’s happening with very high-end wages. Using Social Security Administration (SSA) data, we previously found that wages for the top 1% skyrocketed 182% from 1979 to 2023, roughly triple the growth rate of the 90th percentile and about seven times the growth of the 10th percentile.1 As a result, inequality between the top 1% and the rest of workers substantially worsened over this period.

It is also true that productivity growth—the change in the amount of goods produced or services provided in an hour of work—has outpaced the wages of the vast majority of workers since 1979.2 While higher earners have fared better, every worker should benefit when the economy expands and productivity increases. Had low-end wages grown in line with productivity since 1979 (as they almost surely did in previous decades), the 10th-percentile hourly wage would be $21.04—or 45% higher than it is now. Similarly, the median wage would be 43% higher—or $36.69. Our forthcoming wage calculator will allow users to input any wage level and find how much higher their wages would be if they had in fact grown as fast as productivity.

Policymakers can and must raise wages to address affordability concerns

Making life more affordable for working families is not just about slowing the rate of increase of prices, it’s about ensuring continued wage gains at the bottom and middle of the wage distribution after decades of slow and suppressed wage growth before the COVID pandemic. In 2025, wages for the lowest-wage workers lost the race against prices, making it more difficult for them to afford necessary goods and services.

Policymakers must identify raising wages as the key lever to making life more affordable for working families, and they can do that by raising the minimum wage, reforming labor law to ensure workers can freely exercise their right to unionize, and maintaining full employment. Ignoring these policy levers to raise wages makes the affordability proposition even more difficult to attain.

1. To be clear, this isn’t an apples-to-apples comparison because we are comparing two different data sets and one uses annual earnings (SSA) and the other uses hourly wages (CPS).

2. Productivity growth is the percent change between 1979 and the most recent four quarters. At the time of writing, 2025 Q4 was not available.

This paper focuses on the historical divergence between typical workers’ compensation and economy-wide productivity, and the implications of this divergence for workplace inequality. Commonly referred to as the “pay-productivity gap”, we show that while inflation-adjusted pay and productivity grew in lock step from the end of WWII until the 1970s, productivity has outpaced pay since 1979. We use publicly available data sources to document this gap and discuss key methodological considerations.

We argue that the equal growth in pay and productivity before 1979 and the subsequent fracturing were driven near-entirely by intentional policy decisions. Beginning in 1979, the efforts by employers and capital-owners to undercut the bargaining power of typical workers in labor markets through policies such as deregulation, corporate-led globalization, erosion of the minimum wage, deunionization, macroeconomic policies that tolerated excess unemployment, and tax cuts for high earners led to a divergence in pay and productivity that has yet to be remedied. Not only have these intentional policy decisions created a large pay-productivity gap since 1979, but they have also coincided with a productivity slow-down across the total economy during this time period, resulting in growth that was slower and less equal.

Productivity-Pay Gap

Rising inequality has been a major social and economic problem for the United States for decades. The policy agenda to stop or even reverse this inequality must start with a clear understanding of what drove it. In our view, inequality has come from the labor market, and more particularly from a range of intentional policy decisions that disempowered typical workers when they have sought wage increases from employers. In an effort to disseminate facts about labor market trends more broadly, the Economic Policy Institute has constructed the State of Working America Data Library, which allows users to easily browse and collate detailed data on wages, incomes, and employment. We hope this tool fosters a broadly shared understanding of the labor market changes we need to achieve a fairer society.

Despite recent reductions in inequality, decades of policy failures fueled public discontent and enabled billionaire public influence

In any single year since 1979 there has been a maddening gap between what the U.S. economy could be delivering for working families and what it actually delivered. The richest country in the history of the world has routinely failed to offer broad-based economic security and prosperity through a cascade of policy failures. This gap between the potential and the reality of what typical families eke out of the U.S. economy has shown up directly in skyrocketing incomes for the richest households.

There is a strong case to be made that the 2024 economy—despite just being a few years clear of two global economic catastrophes (the COVID-19 pandemic and the Russian invasion of Ukraine)—had re-attained its pre-crisis performance in record time and in a way that directed more benefits to the lowest-paid workers.

Yet the anxieties driven by the crisis, the unfamiliar burst of inflation, and a decades-long legacy of typical families rightly feeling they weren’t the main focus of policymakers made U.S. households extremely sour about the economy. A fairer distribution of economic rewards in the decades before the crises might have made the public more optimistic about the rocky economic road of 2020–2024 and less willing to take a chance on the extreme policy shift that will come as a result of their election of Donald Trump.

Read the full piece in The Briefing Book. 

But the evidence he cites is totally unconvincing on this.

First, he notes that wages for lower-wage workers have recently grown more rapidly than for middle-wage workers. But it’s been shown again and again that this is driven in large-part by those states that have raised their minimum wages. It’s also been shown that tighter labor markets disproportionately benefit the lowest-paid workers. The argument that changes in relative bargaining power and economic leverage have been the prime mover of wage trends in recent decades is not an argument that wages can never rise, period. When policies change—like minimum wages increase and the Fed allows labor markets to tighten without slamming on the interest rate brakes—good things happen. We just need to change a lot more policies.

Second, he cites a study that looks at wage and productivity growth in high-skill and low-skill industries between 1989 and 2017. The first odd bit of this evidence is that the wage growth he reports the study claims for high and low-skill industries is essentially identical: 26 percent versus 24 percent. The second odd bit is that this means even high-skill industries only gave average annual wage increases of 0.8 percent over that time, even as aggregate productivity grew by almost twice as fast over that time (about 1.4 percent annually). Finally, and most important, using industry-level productivity growth to infer anything about the productivity of individuals working in these industries cannot be done. To put it most simply, productivity growth within an industry can occur because each input used in production gets more productive, or, there is a shift in the mix of inputs. This might sound wonky but I’ll explain a bit more in the next paragraph:

This issue of changing the mix of a firm’s inputs also nullifies the conclusions he draws from his third and fourth bits of evidence—both of which highlight that there has been growth in the inequality of productivity between firms. Again, changes in a firm’s productivity often have nothing to do with individuals’ productivity or economy-wide productivity. Take the example of an auto company that once ran the factory cafeteria with its own employees. One day, it decides to fire the cafeteria employees and “rehire” them through a services staffing company. Measured productivity of the auto company will rise—the cafeteria employees didn’t make cars and so the firm’s output won’t change much. But, if the factory cafeteria was a necessary part of the business (say for retaining workers, or at least keeping them onsite during lunch to minimize time away from the assembly line) and must be kept running, then there has been no economy-wide productivity gain that’s occurred by spinning off the cafeteria workers into a separate firm. Instead, by throwing the cafeteria workers out of the lead firm, the auto company has reduced these workers’ bargaining power and wage declines are likely in their future even as their own productivity has not changed at all.

Finally, we should mention this bit of his column: “Today’s successful bosses are doing what they should be doing: increasing productivity, growing their businesses and offering great service.”

Does anyone really believe that pay for corporate managers is driven by their personal productivity? Are today’s CEOs really that much more productive relative to rank-and-file workers than CEOs were when, say, Microsoft and Apple were founded? CEOs of large public companies made salaries 45 times as large as the pay median workers in their industries in 1989. By 2018, they made 278 times as much. It seems awfully unlikely this was driven entirely by CEOs’ own productivity. We know, for example, that CEO pay is largely driven by luck. For example, oil company CEOs receive large pay gains when the global price of oil rises—which they obviously have little control over. And between 2006 and 2015, a study by Ric Marshall and Linda-Eling Lee found that the relationship between CEO pay and what a company’s shareholders earned on their investment was negative—meaning that CEOs’ pay was not determined by how much they generate in profits.

In its own way, the Brooks column is very useful, highlighting a key political cleavage over what people think drive disparate economic outcomes. Is it simply skills, talent, and hard work, or have institutional and policy changes that disempowered some while protecting others (rule-rigging, to be less polite about it) been the real story? Put me down for the latter view. And it is clear that more and more economists are adopting this view, because it conforms much more tightly to the real-world evidence.