Synopsis

Findings: Between 2019 and 2024, there has been a notable reversal of long-term trends in wage growth. Low-wage workers experienced historically fast real wage growth (adjusted for inflation) and the strongest wage growth compared with workers at all other parts in the wage distribution. The hourly wage for the lowest-paid workers at the bottom 10% grew a tremendous 15.3% over this period. The wage growth at the lower end of the wage distribution was exceptional, significantly faster than their average growth in the prior 40 years and faster than higher-wage workers over the same five-year period. Wage growth for low-wage workers also far exceeded the 2.1% wage loss that characterized the five years following the start of the last pre-COVID business cycle (2007–2012).

Faster wage growth at lower-wage levels has resulted in a compression of wages (or a narrowing of the wage distribution among the bottom 90% of wage earners). In addition, Black and Hispanic workers, young workers, and workers with lower levels of educational attainment experienced relatively fast wage growth over the last five years. Nevertheless, because pay at the bottom of the distribution started at such a low point in 2019, low-wage workers today continue to suffer from wages that are grossly inadequate to sustain families, and significant wage gaps exist at all points in the distribution across demographic groups.

Implications: Policymakers responded to the pandemic recession with actions that made a real difference in people’s lives: Wages grew for those who needed it most. Thoughtful policymaking going forward can help ensure that lower-wage workers continue to see improvements in their standard of living.

Recommendations: Policymakers can choose to prioritize a strong labor market that continues to deliver these gains for lower-wage and historically marginalized demographic groups. Unfortunately, recent actions from the Trump administration will put downward pressure on wage growth and raise the risk of a recession. Policymakers should:

• reverse the general assault on the public sector, restoring federal employment levels and federal payments
• reject cuts to benefit programs like Medicaid and SNAP, critical parts of the safety net for low-wage workers and their families
• prevent an escalation of deportations, which will lower employment and wages across the economy
• oppose across-the-board tariffs that lower real wages without delivering key industrial benefits that a more strategic trade policy approach could realize

Introduction

The current business cycle is a notable reversal of historic trends that increased wage inequality in the United States labor market. Between 1979 and 2019, lower-wage workers experienced only a few short years of strong wage growth in real (inflation-adjusted) wages, and their wage growth over that period significantly lagged behind the wage growth of higher-wage workers. But, between 2019 and 2024, workers in the bottom of the wage distribution have seen fast wage growth compared with their historical norm and with higher-wage workers. This stronger relative growth for lower-wage workers has led to a compression, or a narrowing, of the gap between hourly wages near the bottom and the top of the wage distribution.

Policy choices in the wake of the pandemic and the strong labor market have made these unusually fast gains possible. Historically disadvantaged groups—such as Black and Hispanic workers, young workers, and workers with less than a college degree—have experienced particularly strong wage growth in recent years. Of course, this recent growth has only just begun to narrow these large wage gaps, and the nation’s lowest-paid workers still receive wages that are inadequate to meet most families’ basic needs. Policymakers need to strengthen labor standards so that workers can lock in the gains and continue to build on them, even in weaker labor markets.

Wage growth continued to be strongest for low-wage workers between 2019 and 2024

In this report, we analyze the wage distribution at deciles from the 10th to the 90th percentile of the wage distribution, using Current Population Survey (CPS) Outgoing Rotation Group microdata (EPI 2025a). In our analysis, we employ a new methodology to better smooth our wage deciles by using information from nearby percentiles, described in detail in the appendix. Notably, the labor market story the data tells is unaffected by these changes in methodology. Regardless of method, it is important to note that our estimates of the 90th percentile wage do not fully capture the earnings of those at the very top of the wage distribution, which is better captured with other data sets, discussed briefly later on.1

Our analysis focuses on changes in real wages between 2019 and 2024, as well as historical comparisons of real wage changes between 1979 and 2019. Figure A displays wage growth at each decile of the wage distribution. Between 2019 and 2024, hourly wage growth was strongest at the bottom of the wage distribution. The 10th-percentile real hourly wage grew 15.3% over this five-year period.

To be clear, these are real (inflation-adjusted) wage changes. Overall inflation grew 21.3%, or about 3.9% annually, between 2019 and 2024.2 Even with this historically fast inflation, particularly in the immediate aftermath of the pandemic recession, low-end wages grew substantially faster than price growth. Nominal wages (i.e., not inflation adjusted) for these lower-wage workers rose 39.8% cumulatively since 2019.

Lower-end wages grew faster than any other group within the bottom 90% of earnings. In fact, across the wage distribution, we see the pace of wage growth declining for each successive wage group until we reach the highest wage groups.3 Compared with the 15.3% wage growth at the 10th percentile, growth was less than half as fast at the median or 90th percentile.

The bounceback low- and middle-wage workers experienced was stronger than in any business cycle since 1979

Figure B shows just how exceptional this recovery has been in achieving strong wage growth for low-wage workers. The figure presents real changes in the 10th-percentile wage and the 50th-percentile wage five years from the prior peak in each business cycle since 1979. Wage growth at the 10th percentile was positive in only one other five-year recovery cycle (1989–1994), and even compared with then, the current 10th percentile wage growth is seven times as fast.

Middle-wage workers experienced slower gains in the recent business cycle compared with low-wage workers. However, the slower middle-wage growth over the last five years was still significantly faster than that found in the four prior business cycles, more than twice as fast as the next fastest growth rate across business cycles.

Wage compression in the most recent period contrasts sharply with the prior 40 years

This wage compression is in stark contrast with the experience of workers in the prior four decades.4 Figure C displays wage growth between 2019 and 2024 compared with wage growth between 1979 and 2019. This time we report annualized wage changes—which allow for comparison across periods that span different numbers of years (e.g., a five-year span versus a 40-year span).

The differences in wage growth between these periods are striking because the recent pattern is so contrary to historical trends. Whereas in the most recent period, wage growth was stronger among each successive lower-wage group starting with 70th percentile workers on down, the opposite pattern occurs in the earlier forty-year period. Each successive higher-wage group displays wage growth at least as fast as the previous one. In the most recent period, middle-wage workers’ growth was not far behind growth for the highest-wage workers—about four-fifths as fast—but in the 1979-2019 period, their wage growth was less than half as fast. The difference is even more extreme for the lowest-wage workers: only an average of 0.3% growth over the 40-year period versus nearly 2.9% annualized growth over the past five years. All wage groups experienced annualized wage growth faster in the most recent period as between 1979 and 2019 and much faster among roughly the bottom 60% of the wage distribution. Had the median wage grown over the last 44 years at the 1.1% rate it did from 2019–2024, it would be over $31 per hour today rather than its current value of just under $25 per hour.

While the strongest growth in the recent period swamps the slower growth in the prior period, the overall trend since 1979 is still one of rising inequality since the prior period is nearly 10 times as long as the most recent period being analyzed. (See EPI’s new State of Working America data library for the full complement of data years and trends (EPI 2025f).)

Wage compression over the last five years narrowed the 90–10 wage ratio

Another way to analyze wage inequality or quantify the extent of wage compression across periods is to analyze the wage ratios between different points in the wage distribution. The 90–50 and 90–10 wage ratios are measured by the 90th percentile wage divided by the 50th or 10th percentile wage, respectively. For instance, the 50th and 90th percentile wages were $24.87 and $62.75, respectively.5 Therefore, the 90–50 wage ratio is 2.5, which means that higher-wage workers are paid about two-and-a-half times as much as middle-wage workers on an hourly basis. The average annual change in the wage ratios is determined by comparing these ratios over time and dividing by the number of years in each period.

Figure D displays changes in the 90–50 and 90–10 wage ratios over the 1979–2019 and 2019–2024 periods. Over the 1979 to 2019 period, we see inequality ticking up, with a widening between high-wage workers and both middle- and low-wage workers through the growth in both the 90–10 and 90–50 wage ratios.

Between 2019 and 2024, there was a slight increase in the 90–50 wage ratio because wages at the middle grew just a bit slower than wages at the top. On the other hand, much faster low-end wage growth has shrunk the 90–10 wage ratio considerably over the last five years. While this wage compression is welcome news, it still does not reverse the decades-long growth in equality when measured over the entire 1979 to 2024 period.

The very top continues to amass larger shares of the overall pie

Changes at the very top of the wage distribution cannot be accurately measured using the Current Population Survey, given the censoring of high-end wages and the possibility that high earners do not accurately report their income, but Social Security Administration (SSA) data reveal what’s happening within the top 10%, 5%, 1%, and even 0.1% of the annual earnings distribution.6 Between 2019 and 2023 (the latest data available), average annual earnings also showed signs of wage compression: The bottom 90% grew by 5.0%, and the top 5% and top 1% grew by 2.3% and 0.6%, respectively. But, like the long-term pattern we observed in hourly wages, the five-year compression of wages pales in comparison with the four-plus-decade increase in inequality. Between 1979 and 2023, the average earnings of the bottom 90% grew 43.7%, while the top 5% grew 135.4% and the top 0.1% jumped 353.9% (Gould and Kandra 2024).

Faster wage growth for low-wage workers was driven by policy decisions and a tight labor market

The fast growth over the last five years, particularly for low-wage workers, didn’t happen by chance: It was largely the result of intentional policy decisions that addressed the pandemic and subsequent recession at the scale of the problem. Wage growth in the most recent five-year period far exceeded that of the last four business cycles, including notably, the downright losses following the start of the last pre-COVID business cycle (2007–2012). In the aftermath of the Great Recession, policymakers learned a lesson about the pitfalls of austerity: The pursuit of austerity led to a slow and prolonged economic recovery.

Several large spending bills were passed in the first year of the pandemic, which provided enhanced and expanded unemployment insurance, economic impact payments, aid to states and localities, child tax credits, and temporary protection from eviction among other measures (Gould and Shierholz 2022). These actions provided relief to workers and their families to help them weather the recession. These measures also fed the surge in employment, which gave low-wage workers better job opportunities and leverage to see strong wage growth.

Though ticking up slightly in 2024, the unemployment rate remained low, averaging 4.0% over the year, and the share of the population ages 25–54 with a job—the prime-working-age employment to population ratio (EPOP)—remained high in 2024 at 80.7%, surpassing even the pre-pandemic peak in 2019 of 80.0%. 

This tightening labor market further bolstered workers’ leverage. Low unemployment means that workers are relatively scarce, which requires employers to work harder to attract and retain workers and lessens their discretion to discriminate without facing a profitability penalty. In low-unemployment labor markets, lower-wage and historically marginalized workers experience better labor market outcomes and faster wage growth (Bivens and Zipperer 2018; Wilson and Darity 2022).

In addition, the sudden loss of millions of low-wage jobs at the start of the pandemic, followed by the extraordinarily fast employment recovery, meant that the frictions that tie workers to particular jobs—that is, the barriers that would normally keep workers from searching for better employment opportunities—were not constraining workers looking for work in this period. This “severed monopsony” in a time of furious rehiring reduced the normal drag on wage growth imposed by these frictions (Bivens 2023). High numbers of low-wage workers quit and found better jobs, increasing churn in the low-wage labor market (Autor, Dube, and McGrew 2023). This phenomenon increased low-wage workers’ leverage, which further contributed to faster wage growth. Employers, and particularly employers of low-wage workers, simply had to work harder to attract and retain the workers they wanted.

Higher minimum wages can lock in the gains made by low-wage workers

The minimum wage is an essential labor standard that establishes a baseline for earnings, strengthens the negotiating power of low-wage workers, and helps reduce gender, racial, and ethnic wage disparities. Robust labor standards—like the minimum wage—complement tight labor markets by accelerating wage growth for lower-wage workers. Higher minimum wages help sustain these gains, providing stability for low-income workers during both economic downturns and periods of growth.

Despite the federal minimum wage remaining stagnant at $7.25 per hour since 2009, since then more than half of U.S. states have implemented increases (EPI 2025d). By analyzing wage growth trends across states that have and have not raised their minimum wage, we can assess whether these increases have contributed to wage growth for lower-wage workers.

Before 2019, state minimum wage increases did more to bolster wages at the bottom

Between 2016 and 2017, wage growth for workers at the 10th percentile was twice as fast in states that raised their minimum wage compared with those that did not (Gould 2017). For women at the 10th percentile, wage growth was 2.5 times higher in states with minimum wage increases than in those without, contributing to a reduction in the gender wage gap among low-wage workers.

From 2013 to 2019, leading up to the pre-pandemic economic peak, low-end wages grew by 17.6% in states that increased their minimum wage at least once, almost twice as high as the 9.3% in states that kept it unchanged (Gould 2020). However, the gap in wage growth was smaller during the 2017–2019 period due to a tightening labor market. As previously mentioned, when the unemployment rate is low, the minimum wage has less of an impact, as employers must already offer higher wages to attract and retain workers, which reduces the number of employees directly affected by minimum wage increases. While these increases may not bite in good times, higher state minimum wages can help workers lock in those gains when the labor market softens and lower-wage workers lose the leverage they may have had in the tighter labor market.

A tight labor market and state minimum wage increases worked in tandem to generate immense low-end wage growth

In the most recent period between 2019 and 2024, 29 states and the District of Columbia raised their minimum wage through legislation, referendum, or indexing. To analyze the relationship between these state-level increase and low-end wage growth, we group all 50 states (plus D.C.) into three categories, as shown in Figure E: states with no minimum wage increase, states with a small increase, and states with a large increase.7

In the last five years, nearly every state with a higher minimum wage than the federal minimum of $7.25 experienced an increase in their minimum wage.8 The average nominal increase (not adjusted for inflation) in the minimum wage between 2019 and 2024 among states with any increase was 37.8%. Even with relatively fast inflation of about 21% over this period, average minimum wage increases outpaced inflation.

Figure F compares real wage increases at the 10th percentile, overall and for women, across these three groups of states.9 The key result is clear: Regardless of minimum wage changes, low-wage workers experienced extremely fast wage growth in all states. Even in states with no minimum wage increase, low-wage workers experienced an 11.7% wage increase between 2019 and 2024. Low-end wages grew at incrementally faster rates in states with small and large minimum wage changes compared with states without any change in their minimum wage, 13.7% and 14.8% compared with 11.7%, although these differences were not statistically significant at conventional levels of significance.10

For lower-wage women, specifically, the findings are stronger. The 10th percentile wage for women grew 10.8% in states with no minimum wage change, but low-end wage growth in states with large minimum wage increases was far faster at 15.6%, an extra 4.8 percentage points of wage growth due to the minimum wage. In this case, the wage growth in states with large minimum wage changes was statistically larger than wage increases in states without a minimum wage change.11

Research shows that a tightening labor market alone leads to stronger wage growth among lower-wage workers (Bivens and Zipperer 2018). The 10th-percentile wage grew across all states regardless of changes in state minimum wages because of enhanced relief measures and reduced frictions that boosted low-wage workers’ leverage. It is also important to note that low-wage workers in states with minimum wage increases saw even faster growth than low-wage workers in states without such increases.

Minimum wage increases are crucial to lock in low-wage workers’ gains and build on them

It is essential that we increase the federal minimum wage in order to secure the real wage gains for low-wage workers over the last five years. Unfortunately, Congress has failed to increase the federal minimum wage in the last 15 years, and it is now at its lowest value in real terms in 68 years (Cooper, Hickey, and Zipperer 2022; Zipperer 2024).

Many states and localities have continued to increase their minimum wages in response to federal inaction. On January 1, 2025, 21 states increased their minimum wage, benefiting more than 9.2 million workers (Hickey 2024). Among those affected, 20.4% are in families with incomes below the poverty line, while nearly half (48.5%) have incomes below twice the poverty line (Hickey 2024).

Low-wage workers experienced vital gains due to the tight labor market and legislative measures enacted early in the pandemic recovery. To secure these gains and have protection from weaker labor markets, these workers need strong labor standards such as a higher minimum wage.

Despite historic wage growth, low-wage workers continue to suffer from grossly inadequate wages

Although tight labor markets and, to some extent, minimum wages have bolstered wages at the low end of the wage distribution, millions still work for grossly inadequate wages. Federal policy action is needed to aid working families across the United States in making ends meet.

In 2024, the 10th-percentile hourly wage was $14.26. While this represents a significant improvement from 2019, it is still far from sufficient to make ends meet: Even if that 10th-percentile worker worked full time, their annual pay would be only $29,661. In states that saw increases in the minimum wage between 2019 and 2024, the average 10th-percentile hourly wage was $15.24 in 2024, more than 18% higher than in states that saw no minimum wage increase ($12.85).12

Even with 15.3% wage growth since 2019, it is still difficult—if not impossible—for a 10th-percentile worker to make ends meet. According to EPI’s Family Budget Calculator, whether a worker is making $12.85 an hour or $15.24 an hour, they are still not earning enough to attain a modest, yet adequate, standard of living (a basic family budget for a single individual with no children) in any county or metro area in the United States (EPI 2025b). In fact, there is nowhere in the country where a minimum-wage worker is paid enough to meet the requirements of their one-person local family budget on their wages alone (deCourcy and Gould 2025; Gould, Mokhiber, and deCourcy 2024).

Wage compression meant faster growth for historically marginalized workers

Long-standing discrimination and occupational segregation have led women and Black and Hispanic workers to be disproportionately overrepresented in the low-wage workforce (Bahn and Cumming 2020; Wilson and Darity 2022). Young workers and workers with lower levels of educational attainment also face higher unemployment and lower wages than their more experienced or more educated counterparts.

Table 1 provides wage levels at the middle of the wage distribution—the 50th percentile—for select demographic groups in 1979, 2019, and 2024. We use this to examine how wages within groups have changed in recent times (the last five years) compared with the prior 40 years, with wage changes calculated on an annualized basis for better comparability.

Wage growth has been fastest for Black workers, young workers, and less educated workers

Historically disadvantaged demographic groups experienced far faster wage growth over the last five years compared with the prior 40 years. Although women experienced significant wage growth between 1979 and 2019 due to the increase in labor market opportunities, their wage growth was even greater in the most recent period. Men and white workers’ wages grew in line with overall gains, while Black workers saw the greatest boost in wage growth.

Middle-wage Black men saw the biggest boost in wage growth compared with the earlier period. After increasing at an annualized rate of only 0.1% between 1979 and 2019, Black men’s wages increased at an annualized rate of 1.7% between 2019 and 2024. Black women saw the fastest wage growth over the past five years (1.8%), after experiencing only moderately paced growth in the earlier period (0.7%).

Middle-wage young workers — aged 16-24— and workers with lower levels of educational attainment—with less than a bachelor’s degree—also experienced dramatically faster wage growth between 2019 and 2024 than in the 40 years prior. Both groups saw extremely sluggish growth between 1979 and 2019 (0.2% and 0.1%, respectively), while young workers saw 20 times faster annualized growth (2.6%) and less educated workers saw 10 times faster annualized growth (1.1%) in the most recent period.

Key wage gaps narrowed but still remain large

The wage levels by race/ethnicity and by race/ethnicity and gender in Table 1 can be used to calculate the wage gaps between groups for 1979, 2019, and 2024. For example, the Black-white wage gap is calculated by subtracting the median Black wage ($21.40) from the median white wage ($27.28) and then dividing this number by the white wage. In 2024, this equates to Black workers being paid 21.6% less than white workers. We go one step further and calculate an annualized percentage point change in these racial and gender wage gaps over the most recent period and the prior 40 years (Figure G).

After widening between 1979 and 2019, the wage gaps between Black and white workers overall and between Black and white men, specifically, narrowed over the past five years. Notably, the gap between Black and white men closed at an average annual rate of 0.7 percentage points between 2019 and 2024, making a significant dent in the gap between these groups. The gap between Hispanic and white men also narrowed during this period, but at a slower rate (0.4 percentage points). Black and Hispanic women each experienced significant narrowing compared with white men in the last five years (0.6 and 0.7 percentage points, respectively), much faster than compared with the prior 40 years.

Although these data show promising signs for racial and gender wage equality, significant progress is still needed. In 2024, middle-wage Black workers are still being paid $5.88 less per hour than their white counterparts, while the gap for middle-wage Hispanic workers is even larger ($6.94). For full-time workers, this gap translates to more than $12,200 lower pay for Black workers and $14,400 lower pay for Hispanic workers than white workers. These gaps are even larger between White men and Black and Hispanic women. White men at the median are paid $9.09 and $10.36 more than 50th percentile Black and Hispanic women, which translates to annual earnings gaps of more than $18,900 and $21,500, respectively.

Policy matters

The source of much of the remarkable wage growth and compression over the last five years was a very tight labor market: high aggregate demand and “severed monopsony” (Bivens 2023), breaking the frictions that tie workers, especially low-wage workers, to certain jobs. An expansion of child tax credits, unemployment insurance benefits, food assistance, and direct payments all contributed to the economic recovery from the COVID-19 pandemic and made the labor market the strongest it has been in a generation (Gould and Shierholz 2022).

The Trump administration, however, is currently taking concrete and alarming steps to reverse these accomplishments. Its general assault on the public sector, by terminating the employment of tens of thousands of federal government workers, will directly reduce incomes and demand throughout the country (EPI 2025c). The administration’s elimination of grants and contracts will further weaken the labor market, shrinking universities, charities, and possibly even the entire domestic semiconductor industry (García 2025; Mickle and Swanson 2025). All these efforts will weaken the demand for employment and, therefore, reduce wage growth.

An escalation of deportations will reduce employment of foreign-born and U.S.-born workers alike, and the resulting decrease in demand due to lower incomes is likely to put downward pressure on average hourly wage rates (East et al. 2023). SNAP cuts and Medicaid reductions would eviscerate an already threadbare safety net for low-income families and reduce workers’ fallback position and bargaining power for higher wages (Bergh 2025; Bivens, Wething, and Morrissey 2025). In addition, the Trump administration’s broad-based tariffs will raise prices and reduce real wages in every state (Hersh and Bivens 2025) without providing any benefits that might accrue from a more strategic approach that targeted protection at the specific sectors that need it.

Even after two months there is not a single development or pronouncement from the Trump administration or Congress that is consistent with broad-based wage growth. Instead of coasting on a historically strong economy, the Trump administration seems willing to trash it, and the only question is this: How deep of a hole will they dig for the vast majority of workers who depend on a strong labor market to make ends meet?

Appendix

Calculating wage percentiles

In this report, we calculate real hourly wage percentiles (10th–90th) to analyze changes to wages across the wage distribution. One challenge that researchers face when calculating percentiles is that wages reported in survey data tend to be bunched at round values. For example, in 2019, about 80% of hourly workers paid between $20 and $20.99 reported an hourly wage of exactly $20.00. This bunching arises because people are more likely to be paid at certain round wages (Dube, Manning, and Naidu 2020); survey respondents are more likely to respond to questions about pay with round values; and beginning in 2023, the Census Bureau (Census) began rounding wage and earnings survey responses (Census 2022).

Bunching in survey data is problematic for estimating year-to-year wage growth at specific percentiles because it can increase variability or noise in wage growth estimates. Even in an environment where wages are growing, bunching can cause the median wage, for example, to stay constant for two years at $20 per hour and then in the third year suddenly jump to $21 per hour. In that hypothetical example, median wage growth would be 0% between year 1 and year 2, and 5% between year 2 and year 3, instead of the average growth of 2.5% each year.

To reduce bunching-related problems in measuring wage levels and wage growth, this report adopts the averaged percentile smoothing method recommended by Tedeschi (2024). Specifically, we calculate a given percentile using a weighted average of 9 neighboring percentiles in the distribution. For example, the median wage (50th percentile) is calculated using the weighted average of the 46th through 54th percentiles with weights 1/25, 2/25, 3/25, 4/25, 5/25, 4/25, 3/25, 2/25, 1/25, respectively. This smooths out wage clumps in the wage distribution with an unbiased estimate of the value of a percentile. The averaged percentile method is the method of choice in EPI’s new State of Working American data library (EPI 2025f).

Appendix Figure A shows real wage growth from 2019 to 2024 at each wage percentile using the smoothed averaged percentile method and an unsmoothed percentile. When no smoothing method is applied, wage growth is extremely volatile. On the other hand, averaging the percentiles before calculating wage growth greatly reduces the observed volatility.

Researchers have also used other methods to smooth percentiles estimated from bunched wage and income data. Autor, Dube, and McGrew (2023) use locally weighted regressions to smooth wage percentiles from the CPS. The Bureau of Labor Statistics smooths median usual weekly earnings from the CPS by linearly interpolating between $50 weekly earnings bins (BLS 2025). Similarly, in the past, researchers at EPI have estimated wage percentiles using linear interpolation between $0.25 hourly earnings bins and have estimated average wages within quintiles (Gould and deCourcy 2023; Gould and deCourcy 2024). This report and the new State of Working America data library uses the averaged percentile method described above (EPI 2025f).

An additional complication with CPS wage data is that high values of earnings are censored or “top coded” by the Census. For example, in 2022, about 6% of all wage earners and 13% of non-hourly wage earners had wage values that were top coded at the Census maximum of about $2,885. Beginning in April 2023, Census gradually transitioned to a dynamic top-code regime, adjusting the threshold every month to censor the top 3% of non-hourly and hourly wage earners.

Prior to the new dynamic top-code regime, we impute weekly earnings to those non-hourly workers by fitting year- and gender-specific Pareto distributions to weekly earnings above the 80th percentile and assigning those top-coded with the implied mean values; in 2022, those means were $4,803 and $5,903 for women and men.

Since April 2023, we impute wages for those top-coded by using the Census-provided mean above the top code; for those workers in this time period still subject to the static top code of $2,885, we impute their weekly wages by assigning the implied mean above $2,885 in a given month, using the sample of workers with dynamically assigned top codes and means. Given these new weekly earnings, we calculate hourly wages for non-hourly workers using usual hours worked at the main job.

In order to avoid how the imputation choices for high values of earnings may affect high values of hourly wages, when calculating the averaged 90th percentile in this report, we average only the 89th, 90th, and 91st percentiles, using weights 1/4, 1/2, 1/4, respectively, instead of averaging nine percentiles as we do for the other deciles.

Appendix Figure B shows the 90th percentile 2019–2024 wage growth using our preferred 89th–91st averaged percentile, as well as the 86th–94th averaged percentile, and the unsmoothed 90th percentile. Whether and how one smooths the 90th percentile has a noticeable effect on estimated wage growth because of the clumpiness of the wage distribution and perhaps because of the change in Census-provided top-code thresholds. Our preferred 3-bin 89th—91st percentile method suggests slightly faster growth at the 90th percentile than the 9-bin method and relatively similar growth to the unsmoothed method.

A focus of this report is describing changes to wages of people in the middle of the wage distribution. Appendix Figure C reports the median wage of the smoothed averaged percentile method, the unsmoothed classic method of calculating percentiles, the quintile averaging method (the average of all people between 40th–60th percentiles) used in Gould and deCourcy 2023, and the binned linear interpolation method used in earlier EPI reports. Just as with the 90th percentile, the exact choice of method yields slightly different results. Our preferred averaged method suggests the 50th percentile grew by 5.8%, slightly below the 6.9% growth rate of the 90th percentile, but the range of estimates across methods in Appendix Figures B and C suggests that the growth rates for the two wage percentiles were broadly similar.

Notes

1. Further, as we discuss in the appendix, wage changes at the top of the wage distribution are difficult to measure because of changes in how the Census Bureau censors high-end earnings.

2. Here and throughout this report, we measure inflation using an extended version of the Chained CPI-U, following the recently updated methodology the Census Bureau uses for its historical income series. Given this change, wage levels and trends are not directly comparable to earlier EPI reports, but differences between the points in the wage distribution—measured inequality—are unaffected. Specifically, in this report we use the annual values of the extended Chained CPI-U from version 0.19.0 of the R realtalk package (EPI 2025e).

3. In the appendix, we show that because of reported wage bunching and censoring, there is some uncertainty about the exact growth rate of the 90th percentile, with plausible estimates ranging from about 5.4% to 7.2%, all consistent with our conclusion in this report that lower-end wages grew much faster than those at the top between 2019 and 2024.

4. These findings are consistent with our State of Working America wage reports from prior years, as well as other research. See for instance, Gould and deCourcy 2024 and Autor, Dube, and McGrew 2023.

5. For all wage levels, please visit EPI’s State of Working America data library (EPI 2025f).

6. Wages in the CPS are censored—or hidden—for very high earners because of confidentiality concerns.

7. States with small increases saw nominal increases of less than or equal to 33.3% over the 2019–2024 period, while states with large increases saw nominal increases of more than 33.3% over the same period. We chose this nominal increase threshold because 33.3% is the unweighted median increase, which makes the number of states in each group relatively similar.

8. West Virginia is the one exception; the state’s minimum is higher than the federal, but its last increase was in 2015 (EPI 2025d).

9. In our state-level analysis, we exclude workers whose wages were allocated or imputed by the Census Bureau. The wage allocation model does not include the state (Census 2021), which can artificially mute or flatten differences in wages between states. 

10. In case there’s any confusion, the 10th percentile nationally is not just a weighted average of 10th percentiles in states with and without state minimum wage increases, which is why both growth rates can be lower than the overall 10th percentile reported earlier in this report. Here we calculate the 2024 employment-weighted average of the state-specific changes in the 10th percentile wage.

11. Among women, low-end wage growth in large change states was statistically significantly larger at the 5% level than in no change states, (using a one-tailed t-test, p = 0.032). For all workers, the difference was less precise (p = 0.075). For the 10th percentile for men (not shown), there was essentially no difference in wage growth between the state groups.

12. EPI analysis of Current Population Survey Outgoing Rotation Group microdata (EPI 2025a). The 10th-percentile wage in each state group is a weighted average of the states’ 10th-percentile wages. We exclude workers whose wages were allocated or imputed in these calculations.

References

Autor, David, Arindrajit Dube, and Annie McGrew. 2023. “The Unexpected Compression: Competition at Work in the Low Wage Labor Market.” National Bureau of Economic Research Working Paper no. 31010, March 2023. 

Bahn, Kate, and Carmen Sanchez Cumming. 2020. “Four Graphs on U.S. Occupational Segregation by Race, Ethnicity, Gender.” Washington Center for Equitable Growth, July 2020.

Bergh, Katie. 2025. Millions of Low-Income Households Would Lose Food Aid Under Proposed House Republican SNAP Cuts. Center on Budget and Policy Priorities, February 2025.

Bivens, Josh, Hilary Wething, and Monique Morrissey. 2025. Cutting Medicaid to Pay for Low Taxes on the Rich Is a Terrible Trade for American Families. Economic Policy Institute, February 2025.

Bivens, Josh. 2023. “Learning the Right Lessons from Recent Inflation: Why Macroeconomic Overheating Was the Problem and Austerity Is the Wrong Cure.” American Prospect, January 10, 2023. 

Bivens, Josh, and Ben Zipperer. 2018. The Importance of Locking in Full Employment for the Long Haul. Economic Policy Institute, August 2018. 

Bureau of Labor Statistics (BLS). 2025. “Handbook of Methods, Current Population Survey: Concepts.” Retrieved February 26, 2025.

Census Bureau. 2021. “Imputation of Unreported Data Items” (web page). Accessed February 2024. 

Census Bureau. 2022. “Improving Disclosure Avoidance Procedures for the Current Population Survey Public Use File.” Retrieved March 13, 2025.

Cooper, David, Sebastian Martinez Hickey, and Ben Zipperer. 2022. “The Value of the Federal Minimum Wage Is at Its Lowest Point in 66 Years.” Working Economics Blog (Economic Policy Institute), July 14, 2022.

DeCourcy, Katherine, and Elise Gould. 2025. “EPI’s Updated Family Budget Calculator Shows that States Like Virginia Need a Higher Minimum Wage.” Working Economics Blog (Economic Policy Institute), February 6, 2025.

Dube, Arindrajit, Alan Manning, and Suresh Naidu. 2020. “Monopsony and Employer Mis-Optimization Explain Why Wages Bunch at Round Numbers.” National Bureau of Economic Research Working Paper no. 24991, Revised January 2020. 

East, Chloe N., Annie L. Hines, Philip Luck, Hani Mansour, and Andrea Velásquez. 2023. “The Labor Market Effects of Immigration Enforcement.” Journal of Labor Economics 41, no. 4: 957–996.

Economic Policy Institute (EPI). 2025a. Current Population Survey Extracts, Version 1.0.61, https://microdata.epi.org. 

Economic Policy Institute (EPI). 2025b. Family Budget Calculator. Last updated January 2025.

Economic Policy Institute (EPI). 2025c. How Many Federal Employees Live in Your State? Last updated January 2025.

Economic Policy Institute (EPI). 2025d. Minimum Wage Tracker. Last updated January 1, 2025.

Economic Policy Institute (EPI). 2025e. realtalk R package, Version 0.19.0, https://economic.github.io/realtalk/.

Economic Policy Institute (EPI). 2025f. State of Working America Data Library. Last updated January 2025.

García, Uriel J. 2025. “Texas Refugee Aid Group Sues to Unfreeze $36 Million in Federal Funds.” Texas Tribune, March 3, 2025.

Gould, Elise. 2017. The State of American Wages 2016: Lower Unemployment Finally Helps Working People Make Up Some Lost Ground on Wages. Economic Policy Institute, March 2017.

Gould, Elise. 2020. State of Working America Wages 2019: A Story of Slow, Uneven, and Unequal Wage Growth over the Last 40 Years. Economic Policy Institute, February 2020.

Gould, Elise, and Katherine deCourcy. 2023. Low-Wage Workers Have Seen Historically Fast Real Wage Growth in the Pandemic Business Cycle: Policy Investments Translate into Better Opportunities for the Lowest-Paid Workers. Economic Policy Institute, March 2023.

Gould, Elise, and Katherine deCourcy. 2024. Fastest Wage Growth over the Last Four Years Among Historically Disadvantaged Groups: Low-Wage Workers’ Wages Surged After Decades of Slow Growth. Economic Policy Institute, March 2024.

Gould, Elise, and Jori Kandra. 2024. “Wage Inequality Fell in 2023 Amid a Strong Labor Market, Bucking Long-Term Trends: But Top 1% Wages Have Skyrocketed 182% Since 1979 While Bottom 90% Wages Have Seen Just 44% Growth.” Working Economics Blog (Economic Policy Institute), December 11, 2024.

Gould, Elise, Zane Mokhiber, and Katherine deCourcy. 2024. What Constitutes a Living Wage? A Guide to Using EPI’s Family Budget Calculator. Economic Policy Institute, January 2024.

Gould, Elise, and Heidi Shierholz. 2022. “The Economy Is Recovering Fast. But We Need to Ensure It Works for Everyone.” CNN Business Perspectives, March 3, 2022. 

Hersh, Adam S., and Josh Bivens. 2025. Tariffs—Everything You Need to Know but Were Afraid to Ask (fact sheet). Economic Policy Institute, February 10, 2025.

Hickey, Sebastian Martinez. 2024. “Over 9.2 Million Workers Will Get a Raise on January 1 from 21 States Raising Their Minimum Wages.” Working Economics Blog (Economic Policy Institute), December 17, 2024.

Mickle, Tripp, and Ana Swanson. 2025. “Trump’s Call to Scrap ‘Horrible’ Chip Program Spreads Panic.” New York Times, March 10, 2025.

Tedeschi, Ernie. 2024. “Introducing the Low-Wage Index: A Compositionally-Adjusted Look at Low-Wage Workers Since 1979.” Briefing Book, July 15, 2024.

Wilson, Valerie, and William Darity Jr. 2022. Understanding Black-White Disparities in Labor Market Outcomes Requires Models That Account for Persistent Discrimination and Unequal Bargaining Power. Economic Policy Institute, March 2022. 

Zipperer, Ben. 2024. “Real Value of the Minimum Wage (Adjusted for Inflation)” (web page). Accessed February 2025.