Key takeaways:

• The vast majority (85%) of young college graduates work in occupations that have seen strong employment growth in recent years.

• Young college graduates, like college graduates in general, are more likely to work in AI-exposed occupations than the overall workforce—and considerably more likely than young noncollege workers.
• But both young college graduates and young noncollege workers have experienced rising unemployment over the last three years, suggesting AI is not likely to be driving labor market weakness.

In the first blog post of our Class of 2026 series, we showed that the strong labor market for young college graduates of the early 2020s had begun softening in recent years. A growing share of young college graduates are seeking employment, but because their employment rates have not kept up with this job search, their unemployment rate has risen faster than the overall rate. The second blog post in the series discussed the industries where young college graduates worked. We found that recent graduates work in growing industries, but are forced to enter a weakened labor market with less job turnover, deteriorating their ability to break in. Young college graduates work in the tech sector at a similar rate to college graduates, and there is no clear evidence that tech sector employment is significantly decreased despite warnings about the advancement of AI.

In this blog post, we delve deeper into the occupations where young college graduates are likely to work.1 We examine whether it has been relatively more difficult to secure employment in these fields as the labor market has weakened. We also scour the data for signs that exposure to AI-related occupations may disproportionately affect the prospects for young college graduates as they enter the labor market.

Most young college graduates work in occupations with strong growth

Over 60% of young college graduates work in professional and related occupations or management, business, and financial occupations. Figure A displays the share of employment in each occupation or occupation grouping for young college graduates ages 22 to 27, all college graduates, and young workers without a four-year college degree. Occupations in the figure appear in order of the share of young college graduates employed in each, from largest to smallest. Over half (62.8%) of young college graduates work in professional, management, business, and financial occupations. Workers of any age with a college degree are slightly more likely to work in those two occupations (64.5%), though more likely in management occupations than professional occupations. On the other hand, nearly half (48.3%) of young noncollege workers are in service occupations or farming, construction, installation, and production occupations.

Figure A

Figure B shows the change in employment in each occupation between 2019 and 2026 and between 2023 and 2026, arranged in the same order as Figure A for comparison. Since 2019, management, business, and financial occupations and transportation and material moving occupations experienced the most growth, followed by professional and related occupations.

The top four occupations for job growth since 2023 account for 85% of young college graduate employment. The occupations with employment losses over the last three years were more likely to employ young noncollege workers than college graduates. It doesn’t appear that the occupations where young college graduates tend to work have been hit particularly hard in the last couple of years.

Figure B

While there has been job growth among occupations that tend to be filled by young college graduates, some worry about an increase in labor market underutilization, i.e., when workers with a college degree wind up working in jobs that typically don’t require one. Using O*NET data2, the New York Federal Reserve tracks this type of labor market underutilization. While the share of recent college graduates working at a job that doesn’t require a college degree has ticked up slightly over the last three years, it remains lower than it was for workers who graduated in the aftermath of the Great Recession. Even as late as 2017, young college graduates were working at these noncollege jobs at higher rates than they are today.

While college-educated workers are in more AI-exposed occupations, this does not appear to be driving labor market weakness

Much has been written in the last few years about AI exposure and its impact on the labor market. Using data from ADP, a large payroll processing company, Brynjolfsson, Chandar, and Chen find that entry-level workers in AI-exposed occupations—particularly AI uses that automate, not augment their work—have experienced an employment decline larger than that of older workers in the same occupations and all workers in less exposed occupations, explaining some of their stagnant overall employment growth. Atkinson and Yamco also find that declines in AI-exposed occupations are tied to lack of hiring rather than layoffs, hitting harder for young people attempting to enter the labor market. The second blog post in our series noted an across-the-board slowdown in hiring—which hurts the job prospects of all young workers, not only those in the industries most affected by AI.

On the other hand, researchers at the Yale Budget Lab argue that there has only been a slight increase in the shift in the occupation mix of employment, which would be evidence of AI automating jobs. They find that high AI-exposed occupations—determined by the top quintile of AI exposure—have yet to show declining employment, so no “dissimilarity” between young and older college graduates in terms of occupation mix has materialized. Raderman also finds that there isn’t strong evidence that AI is responsible for weaker labor market outcomes for recent college graduates, using evidence from Tillerman on college majors paired with change in unemployment.

Given the variation in assessments, we wanted to take a look at the data ourselves. Gimbel, Kendall, and Kulsakdinun have done an admirable job of summarizing the literature that attempts to classify AI exposure and propose a weighted aggregate measure of AI exposure.3 We employ this measure to investigate whether young college graduates may be more likely to be at risk in AI-exposed occupations than other workers.

In Figure C, we display the AI exposure of occupations weighted by the share of the entire workforce in each occupation. Moving from the left to the right on the figure increases AI intensity. For instance, professional and office & administrative support occupations are more AI exposed (to the right), while production, transportation, and service occupations are less AI exposed (to the left). Overall, the mean AI exposure score is 0.23.4

Figure C

In Figure D, we show the distribution of select demographic groups by occupation and AI exposure. As with earlier analysis, we compare young college graduates with all college graduates and young noncollege workers, in separate panels in the figure.

According to the aggregate measure, college graduates do have higher AI exposure in the labor market than the overall workforce. It is clear there is more mass in the direction of higher exposure (to the right) and their mean exposure is 1.07, higher than that of workers writ large. But the AI exposure of young college graduates isn’t any higher than that of college graduates in general. Mean AI exposure among young college graduates is 1.00.

Figure D

What is striking is that the AI exposure among young college graduates (1.00) is considerably higher than that of young noncollege workers (-0.61). If AI was driving labor market outcomes, we’d expect young college graduates to fare worse in today’s economy, e.g., see larger declines in employment or faster increases in unemployment. But, when we compare unemployment rates as we did in the first blog post of this series, both groups experienced similar increases in unemployment over the last two to three years. Trends in employment rates were also consistent across these groups.

Since the weakening labor market is hitting both young college and noncollege workers alike, it’s hard to argue that AI is uniquely causing job losses for new labor market entrants graduating from college now or in recent years. These findings are consistent with the literature, as there is currently no consensus about the effects of working in AI-exposed occupations on employment thus far.

1. Throughout this brief, we define young college graduates as people between the ages of 22 and 27 with only a four-year college degree. Unlike similar analyses of young workers, we do not exclude young college graduates who are currently enrolled in school, but the results here are robust either way. Unless otherwise noted, data for 2026 represent a 12-month average from April 2025 through March 2026 for the most up to date and reliable estimates, which removes seasonality and increases sample sizes.

2. O*NET or the Occupational Information Network provides the largest up-to-date database of information about workers sorted into detailed occupations. Information provided is about skills, abilities, education, training, and more.

3. We use an updated summary AI exposure PCA score (principal component analysis weighted standardized z-score) provided by the authors, May 13, 2026.

4. The PCA score scale is centered at 0, the unweighted mean across occupations.

Key takeaways

• The depressed overall hires rate is a key driver of new labor market weakness for young college graduates, as it makes it harder for them to break into the labor market. This is true across industries, not just in those that disproportionately employ young college graduates—suggesting the culprit is not a structural change in the economy like AI but a labor market in which employers are hiring less and workers are holding onto the jobs they have.
• The information sector—posited to be more AI-exposed—has experienced recent job losses but employs only 2.3% of young college graduates.
• High-tech industries, which employ about 1 in 10 college workers, expanded at a historically rapid pace in the early 2020s but have shown signs of softening over the last three years.

The early 2020s labor market for young college graduates was strong. But, as we showed in this series’ first blog post, the Class of 2026 is graduating college into a labor market that has notably weakened in the past two years. A growing share of young college graduates are looking for jobs, but their employment rates have not kept pace—meaning unemployment is rising faster for young graduates than for the overall workforce. While their outcomes remain better than those of their noncollege counterparts, the uptick in unemployment has been a rising concern.

In this blog post, we delve deeper into the industries where young college graduates are likely to work,1 examining whether it has been relatively more difficult to secure employment in these fields as the labor market has weakened. Our analysis first examines employment changes, then turns to labor market flows, including hires and separations rates. We also scour the data for signs of contraction in the tech sector that may disproportionately affect the prospects for young college graduates as they enter the labor market.

In the third blog post in the series, we will examine the occupations where young college graduates work with particular attention to occupations that may have grown or shrunk, as well as to those most exposed to AI.

Young college graduates work in industries with strong growth in this business cycle

Over half of young college graduates work in private education and health services, professional and business services, or public-sector jobs. Figure A displays the share of employment in each industrial sector or sector grouping for young college graduates ages 22 to 27, all college graduates, and young workers without a four-year college degree. Industries in the figure appear in order of the share of young college graduates they employ, from largest to smallest.

The types of jobs where young college graduates work look similar to those of college graduates generally. Young workers without a college degree (i.e., noncollege) are far more likely to work in trade, transportation, and utilities; mining, construction, and manufacturing; or leisure and hospitality. All groups of workers are least likely to work in the information sector, closely watched for signs of AI-induced displacement.

Figure A

Figure B shows the change in employment in each sector between 2019 and 2026 and between 2023 and 2026, arranged in the same order as Figure A for comparison. The two fastest growing sectors since the last business cycle peak occurred in the two largest sectors for young college graduates: private education and health services and professional and business services.

Since the rollout of ChatGPT, many have looked at industries and occupations likely to be exposed to AI to see whether this has led to weaker job growth. Among the most closely watched of these industries is the information sector, which has seen an 8.5% employment decline since 2023.2 While these losses are notable—and especially relevant to understanding AI’s fingerprints on the labor market for young college-educated workers—it cannot be overemphasized just how small this sector is in the overall economy. Less than 2% of overall employment is in the information sector, including only 2.3% of young college graduates. Further, the sector saw a rapid employment expansion between 2019 and 2023—the employment loss between 2019 and today is just 2.0%.

Figure B

High-tech sectors have relatively more college graduates but haven’t experienced large AI-induced employment losses

In recent years, the Census Bureau has created an experimental data series on high-tech industries to better understand business dynamics. These include both manufacturing and service sector components of high tech.3 We translate their NAICS classification for high-tech industries into Census Industry Classifications used by the Current Population Survey (CPS) to determine the likelihood of young college graduates working in these sectors.

In the 2026 economy, about 5.6% of workers were employed in what the Census considers high-tech industries. Just about 1 in 10 (9.9%) of the college-educated workforce works in the tech industry. Young college graduates are similarly represented: 10.3% work in tech.

Overall, the Current Establishment Survey tells us that tech industry employment tracked changes in overall private employment in the prior business cycle (between 2007 and 2019) but expanded sharply in the early 2020s and has softened a bit in the last three years. Since 2023, the tech sector has fallen by 0.7%. While overall employment using CPS does show modest growth, neither shows large swings that suggest a large impact for young college graduates.4

Weak hires may be the biggest culprit to labor market weakness for young college graduates

The Job Openings and Labor Turnover Survey (JOLTS) can shed light on the question of whether entry-level workers—with or without college degrees—are facing a harder labor market to break into. While JOLTS doesn’t include demographic characteristics, it presents jobs openings as well as rates of hiring, layoffs, quits, and other separations. Today’s economy has substantially less churn than during the recovery from the pandemic, when millions of workers reentered the labor market after mass layoffs—many quit soon after as they searched for, and generally found, better opportunities.

Figure C shows the hires and separation rates. The lighter colors represent the monthly seasonally adjusted data for each series while the darker colors represent a 12-month moving average that provides a better overall picture of recent trends, smoothing out some data volatility. Over the last five years, the hires rate has steadily fallen and now sits at levels last seen in 2013 and 2014, when the labor market was still struggling to recover from the Great Recession.

The total separations rate includes quits, layoffs and discharges, and other separations.5 As with the hires rate, the separations rate has been declining over the last few years and now sits about where it was in 2014. Much of this is due to reductions in quits. Quits are higher when workers feel confident that they will find better job opportunities. Right now, workers are sitting tight, more so than any point in the past 10 years. Taken together, there is simply less churn in the labor market. But reduced churn is not inherently bad. If the frantic labor market of the early 2020s led to many workers and employers finding satisfactory matches, it could make sense that the following years would see less churn than normal. But for young workers looking to enter the job market, a reduction in hiring can make it harder to find a job.

Figure C

Table 1 breaks down the change in the hires and separations rate over the last three years, again using 12-month moving averages to smooth some volatility in the data. The industries are listed in order of the share of young college graduates they employ, similarly to Figure A. Overall, the hires rate fell 0.8 percentage points and the separations rate fell 0.6 percentage points between 2023 and 2026. The industries where young college graduates are more likely to work saw smaller reductions in both hires and quits than the overall. Industries where young workers without a college degree are more often found—over a quarter are in trade, transportation, and utilities—saw greater losses. Finally, leisure and hospitality, where young noncollege are more than twice as likely to work as young college graduates, saw the largest declines in hiring.

It does not appear that the industries where young college graduates tend to work are experiencing more weakness than other industries. Job gains are just as strong, if not stronger, and hiring hasn’t fallen as far in other industries. In short, there does not seem to be any profound structural change in the economy affecting the industry composition of employment—AI or anything else—that would easily explain the softening of the labor market for young college graduates in recent years. What it does appear to be is a harder labor market for young workers to break into when employers are less likely to hire and workers are more likely to sit tight in the job they have.

1. Throughout this blog post, we define young college graduates as people between the ages of 22 and 27 with only a four-year college degree. Unlike similar analyses of young workers, we do not exclude young college graduates that are currently enrolled in school, but the results here are robust either way. Unless otherwise noted, data for 2026 represent a 12-month average from April 2025 through March 2026 for the most up to date and reliable estimates, which removes seasonality and increases sample sizes.

2. ChatGPT was first introduced in November 2022 but took several months for more widespread usage.

3. High tech industries: Computer and Peripheral Equipment Manufacturing, Communications Equipment Manufacturing, Semiconductor and Other Electronic Component Manufacturing, Navigational, Measuring, Electromedical, and Control Instruments Manufacturing, Aerospace Product and Parts Manufacturing, Software Publishers, Data Processing, Hosting, and Related Services, Other Information Services, Architectural, Engineering, and Related Services, Computer Systems Design and Related Services, and Scientific Research and Development Services.

4.  It’s not unusual for the CPS and CES to display small differences in employment levels or trends considering nontrivial differences in their methodologies. 

5. Other separations include separations due to retirement, death, disability, and transfers to other locations of the same firm.

But diverting public funds away from public K–12 schools and toward private schools does not guarantee educational opportunities will be expanded for all students—and this is especially true in rural communities. Most obviously, because students in rural communities often don’t have a private school option and therefore cannot use the vouchers, state voucher programs—which are financed by all the taxpayers in a state—amount to an education subsidy for wealthy urban families at the expense of strong public schools. Moreover, for rural areas that can support multiple school systems, voucher programs introduce a potentially large cost for the students that remain in public schools, as any sharp drop in public school enrollment will raise the fixed cost per pupil of running schools. For example, school facilities and staff that are efficient for 1,000 students in a school may no longer be efficient if enrollment were to drop to 800 or 900.

Voucher programs work like this: Parents who wish to send their kid to private school can receive public funding to cover part of the tuition or education-related expenses, rather than paying out of pocket. In states with vouchers programs, this added cost to government of paying for private educational expenses makes a big dent in state budgets—see examples here, here, and here. These programs also often entail fraud and abuse of funds and strip away funding for public schools. As a share of K–12 budgets, voucher spending accounted for as much as 26% in 2025, squeezing public schools of sorely needed funds. Moreover, recent reports have documented accounts of voucher funding getting used for high-end concert tickets and rideshare apps like Uber and Lyft. For wealthy parents in urban districts who were already planning to send their kids to private school, these slippery regulations and extra funding for education expenses are a feature, not a bug, of voucher programs. Vouchers are disproportionately taken up by students already attending private school, compared with those who consider a private school option when voucher laws get passed in their state.

For students in rural areas with no private school option, voucher programs simply mean there is less to spend on public schools, which leads to teacher shortages, fewer educational opportunities, and worse building maintenance. In rural communities with homeschooling or private school options, voucher programs impose an added cost to public education when students transition from public to private school.

We call this cost the fiscal externality of voucher programs, and it is borne by school districts, students, and their families when voucher-driven declines in student enrollment intersect with the fixed nature of many school costs. In rural districts, many key education costs—such as interest on bonds issued in the past, heating, electricity for school buildings, bus drivers, and even some staff—cannot easily adjust to student enrollment declines.

While public schools’ fixed costs do not decline when they lose students to voucher programs, their revenue does. Thus, when students in rural areas take up vouchers to leave public school for private school or homeschool, public schools have less revenue to cover the same level of fixed costs. The costs that can be adjusted—such as supplies or certain personnel—will get forced down due to shrinking school budgets. These variable costs are crucial for effectively educating children, meaning students who remain in public schools will pay the price of voucher program takeup.

This fiscal externality therefore leaves districts unable to deliver the same level of instruction to the remaining public school pupils. When students leave public schools in rural areas with voucher programs, there are fewer resources available on a daily basis to educate kids—fewer teachers and other staff members and fewer curriculum and education supplies. Education quality suffers.

How large is the fiscal externality that voucher programs impose on public schools in rural districts? Take the McComb Local School District in Ohio, which had 627 students in 2022 and is classified as a rural district according to the U.S. Census. Using EPI’s Fiscal Externality Calculator, we estimate that a 5% decline in enrollment would lead to an increased cost of $520 per pupil for the remaining students in the district, or a total of $309,530.

The key assumption is that there is some fraction of schools’ costs that is fixed and can’t be adjusted in the near term when enrollment falls. We assume that instruction and services costs (the cost of teachers and services like transportation, counseling, nurses, and school administrators) can only partially adjust to changes in enrollment. Specifically, we assume that when enrollment declines, instruction costs are only able to adjust by 50% of the enrollment decline, and service costs are only able to adjust by 20%. We assume that capital and building and maintenance costs can’t be adjusted at all. (Users can set their own adjustment rates for their school districts using the fiscal externality calculator here. The method behind this calculation is detailed in our report.)

Under these assumptions, aggregating all the rural Ohio districts using the rural categorization of school districts from the National Center for Education Statistics, a voucher-driven 5% enrollment decline would impose a fiscal externality of just over $206 million on Ohio public schools.

Rural districts have the most to lose when states enact voucher programs. For rural communities, vouchers are not a cost-free policy that simply expands education options for children—they are a subsidy for wealthy urban and suburban families at the expense of strong public schools. Voucher programs also introduce a large potential cost for the students that remain in rural public schools. The public spending declines associated with the introduction of vouchers will reliably cause significantly worse educational outcomes at a time when states should be spending more—not less—on public schools. States that promote voucher programs at the expense of funding for strong public education are signaling that rural students are not a priority. 

Key takeaways

• Proposed state legislation to extend full, equal collective bargaining rights to all state and local government workers can help Virginia:
  • reduce the state’s large public-sector pay gap
  • improve public services
  • reduce staff vacancies and turnover
  • decrease racial and gender wage disparities
• Strong collective bargaining rights and increased unionization rates are highly correlated with numerous, widely shared benefits including higher wages, more equitable state economies, and healthier democracies.
• Virginia currently has one of the largest public-sector pay gaps in the nation. State and local government employees in Virginia earn, on average, 26.7% less than private-sector peers with similar education and experience.
• The public-sector pay gap varies across states and is largest in states like Virginia where most public employees lack collective bargaining rights. In Virginia, collective bargaining is currently banned for state employees and only recently became permitted for some local government employees.

In 2026, Virginia lawmakers are poised to consider transformative legislation that would extend full collective bargaining rights to public employees at all levels of state and local government. The benefits of comprehensive collective bargaining rights would extend far beyond affected public employees who would enjoy better working conditions. Stronger state collective bargaining laws can help improve the quality of public services and economic outcomes for all Virginians.

Data show that strong collective bargaining laws help states address persistent public-sector pay gaps, reduce staff vacancies and turnover, and lead to higher unionization rates (Morrissey and Sherer 2024). Increased unionization rates are highly correlated with numerous, widely shared benefits including more equitable state economies and healthier democracies (McNicholas et al. 2025).

For Virginia, expanding public-sector bargaining is a critical next step in building an economy that works for all. The expansion will reverse a long history of anti-worker state policies that have suppressed wages and limited workers’ power in the labor market by blocking pathways to unionization. Such policies have resulted in greater income inequality and persistent racial and gender wage disparities (Bivens and Shierholz 2018; Mishel and Bivens 2021). Strong, comprehensive state legislation covering public employees’ labor rights is also especially important at a moment when the federal government has been attacking the jobs, working conditions, and union contracts of over 235,000 federal civil servants residing in Virginia, and threatening long-standing federal protections of all workers’ rights (EPI 2025b; Oakford and Poydock 2025). 

This report examines how public-sector workers with limited or no collective bargaining rights fare compared with public-sector workers with well-established collective bargaining rights. To do so, we estimate pay differences between state and local government workers and private-sector workers with similar education and experience. In this analysis, Virginia is among a minority of states in which state employees have no bargaining rights and only some local government employees have limited bargaining rights. By contrast, 27 states have well-established collective bargaining rights for state and local government workers (Table 1).

The right to bargain collectively over pay is associated with higher unionization rates (union membership as a share of the workforce) in a given state. This report shows that collective bargaining rights and union strength help state and local government workers narrow the pay gap with private-sector workers. We show that the pay gap for public employees is significantly larger when these workers have weak or no bargaining rights, like public employees in Virginia, which has one of the largest public-sector pay gaps in the nation (-26.7%).

Virginia public employees lack collective bargaining rights

Proposed legislation in Virginia would, for the first time in the state’s history, guarantee that all state and local government employees enjoy labor rights similar to those of their private-sector peers (whose rights to collectively bargain are covered under the federal National Labor Relations Act) (HB 1263 and SB 378). The Virginia Assembly passed similar legislation (HB 2764 and SB 917) in 2025, only to have it vetoed by then-Governor Glenn Youngkin (McGinley 2025). 

Virginia is one of a handful of Southern states that for decades explicitly banned public employees and employers from entering into collective bargaining agreements. In 2020, Virginia took an important step toward making collective bargaining newly optional for local governments, but the state’s policies remain out of step compared with most states. Virginia lacks a statewide collective bargaining statute covering all local government employees and still has a ban in place barring state employees from collective bargaining (Borja 2022).

As shown in Table 1, the majority of states and Washington, D.C., already ensure collective bargaining rights for most public employees, including state workers. Many states have had statewide public-sector bargaining statutes in place for decades (Rueben 1996; Sanes and Schmitt 2014). Such laws typically set clear, uniform guidelines for union elections and contract negotiation processes and establish state labor boards charged with fostering productive labor-management relations (including timely contract settlements), ensuring broad awareness of and compliance with statutory guidelines, and mediating or adjudicating disputes as needed.

Table 1

Who are public employees in Virginia?

In 2024, approximately 560,000 Virginians worked in state and local government occupations (BLS-CES 2020–2025). This includes teachers and school staff, firefighters, transit operators, law enforcement, administrative staff, and employees serving the state’s public safety, transportation, health care, judicial, corrections, and higher education systems. As shown in Table 2, 21.2% of Virginia state government employees and 20.4% of Virginia local government employees are Black, and 57.3% of state government employees and 64.5% of local government employees are women. Many public-sector workers in the state are highly educated. Virginia public-sector workers are more than twice as likely to have advanced degrees as private-sector workers and are much less likely to have a high-school level or less education.

Table 2

Anti-union state policies are rooted in racism and harm all Virginia workers

Virginia’s ban on union contracts for public employees is a Jim Crow-era policy with deep roots in the history of slavery and white supremacy. After the passage of federal labor laws accelerated worker organizing in the 1930s, Virginia joined several Southern states in adopting anti-union state laws designed to prevent multiracial union organizing and suppress Black workers’ wages and power (Childers 2023). The Virginia Assembly first took an explicit stance on public-sector bargaining in response to the unionization of Black hospital employees at the University of Virginia in 1946, via a joint resolution declaring it against the public policy of the state to negotiate with public employee unions (a stance later affirmed by state supreme court decisions and codified in statute in 1993). Virginia was also among the first states to adopt anti-union so-called right-to-work legislation in 1947, an anti-labor policy jointly promoted by white supremacist organizations and industry groups intent on slowing the growth of unions to maintain access to cheap labor—especially in Southern states (The Commonwealth Institute 2022; Watts 2021; Pierce 2017, 2018; Sherer and Gould 2024).

As a result of these long-standing anti-union state policies, unionization rates in Virginia for both public- and private-sector workers are well below national averages.

Figure A

Narrowing Virginia’s large public-sector pay gap

Across the country, public-sector employees earn less than their private-sector counterparts, and this pay gap has widened in recent years. In the latest available five-year period (September 2020–August 2025), state and local government employees earned, on average, 17.2% less than private-sector employees with similar education and experience. The size of the public-sector pay gap varies across states and is largest in states like Virginia where most public employees lack collective bargaining rights. Public-sector workers with strong bargaining rights experience a narrower pay gap (-14.3%) than those with weak (-19.6%) or no bargaining rights (-22.5%) (see Table 3).

Virginia currently has one of the largest public-sector pay gaps in the nation. Among all 50 states, Virginia’s -26.7% public-sector pay gap appears to be the second highest, though differences among states clustered at the bottom of the rankings are not statistically significant. (Pay gap statistics are for full-time wage and salary workers ages 18–64, based on the authors’ analysis of pooled September 2020–August 2025 Current Population Survey microdata downloaded from Flood et al. 2025 and EPI 2025a.)

Compensation packages of public employees, on average, include more robust benefits than those of private-sector workers, but Virginia’s public-sector compensation gap remains large, even when factoring in more robust benefits. Benefits are an estimated 32.0% of pay for private-sector workers in the South Atlantic region and an estimated 51.3% of pay for Virginia public-sector workers. Factoring in benefits, Virginia’s public-sector compensation gap shrinks to a still sizable 16.0% (authors’ estimate based on BLS-ECEC 2020–2025 and Public Plans Data 2020–2024; see Morrissey and Sherer 2024 for methodology).1

Table 3

Addressing high staff vacancy and turnover rates in Virginia’s public sector

The growing public-sector pay gap has become a particular concern at a moment when low pay has created serious challenges to recruiting and retaining teachers and other public employees across the country (Cooper and Martinez Hickey 2022; MissionSquare Research Institute 2023; Wething 2024a, 2024b; Martinez Hickey 2025). Virginia has faced particularly acute staffing shortages in public education and in units of state government in recent years (Manzanares 2025; Cantor 2025).

The latest (2025) biennial compensation report from the state’s Department of Human Resource Management notes that “Years without any salary adjustments in the past have made it difficult for state agencies to build a proactive and sustainable approach to addressing compensation, recruitment and retention concerns” (VDHRM 2025). Between 2001 and 2023, low average salary increases for state workers (just 2.9% per year, compared with 3.4% annually in the private sector) have led to high vacancy and turnover rates. As of 2024, roughly 1 in 5 (22.4%) state jobs were unfilled, and the median salary across the state workforce was just $61,305—a full $5,000 less than the median city employee salary in Richmond, which has adopted its own collective bargaining ordinance (McGinley 2025).

Chronic state employee staffing shortages are already having direct impacts on the health, safety, and quality of life of Virginians. To take one example, in 2025 the Virginia Department of Juvenile Justice (DJJ) drew headlines after reporting that inability to adequately staff large facilities for incarcerated youth had led to unsafe conditions, lockdowns, increased restrictions on out-of-cell time, and a lack of rehabilitative services. DJJ directors indicated ongoing difficulty recruiting for open positions, despite participating in job fairs, college and university visits, outreach to military and veteran communities, and offering signing bonuses and referral incentives (Manzanares 2025). The root causes of understaffing identified by Virginia state agencies like DJJ—from burnout and high workloads to low starting salaries, lackluster raises, and difficult or unsafe working conditions—are precisely the topics that a structured collective bargaining process would allow state agencies to address, with direct input from frontline employees.

Addressing racial and gender pay gaps to improve recruitment and retention of workers of color and women

The public-sector pay gap disproportionately affects Black workers and women, who are more likely to be employed in public-sector jobs and who are disadvantaged in the broader labor market (Childers 2025). Strengthening collective bargaining rights for government workers in Virginia therefore promises to narrow the pay gap and reduce racial and gender inequalities in public institutions and across the labor market.

For example, a 2024 RAND report showed that Black teachers nationally receive lower average salaries and pay raises than white teachers do, a difference linked directly to the fact that Black teachers were less likely to live in states where public educators had collective bargaining rights. The inadequacy of pay is one of the main reasons teachers report for leaving the profession, further contributing to the demographic mismatch between teachers and students (e.g., nationwide over half of all students are children of color, but the teaching workforce remains around 80% white) (Steiner et al. 2024; Gopalan 2025). Likewise, data show that in states like Wisconsin where legislators have weakened formerly strong collective bargaining rights in the past two decades, resulting decreases in unionization levels and worker wages have measurably widened public-sector pay gaps and gender pay gaps (Nack et al. 2019; García and Han 2021; Biasi and Sarsons 2022).

Building on success and remedying limitations of current state law that only permits local collective bargaining

In 2020, Virginia partially lifted its long-standing ban on public-sector collective bargaining with legislation creating an “opt-in” system that has allowed local governments to set their own policies on whether and how to bargain with their own employees. While this opening fell far short of creating a consistent statewide framework for collective bargaining, it has resulted in at least 17 of Virginia’s largest cities, counties, and school boards adopting collective bargaining ordinances and creating new pathways to union contracts for substantial numbers of public employees—including, for example, 27,000 teachers and school staff and 12,000 county employees in Fairfax County; nearly 4,000 City of Richmond employees; and others (Borja 2022; Sharma 2022; Khalil 2023; Lukert 2024; Pope 2025; Walter 2026).

The local “opt-in” system was a positive step forward that has already revealed high levels of interest in collective bargaining among Virginia workers. The system resulted in new union contracts covering tens of thousands of frontline educators and civil servants, providing an initial boost to Virginia’s historically low unionization rate.

Significant limitations of the new “opt-in” system have also quickly become clear. As noted above, the major shortcoming of the current law is that it does not ensure equal collective bargaining rights for all public employees. Virginia state employees remain barred from collective bargaining, and in local government where collective bargaining is permitted (but not required), workers continue to lack collective bargaining rights, unless they are able to persuade local officials to adopt and then implement a collective bargaining ordinance. When localities do adopt such ordinances, they may vary in strength and effectiveness (Overman 2023).

As a result, Virginia’s current “opt-in” system for local collective bargaining has generated an uneven patchwork of highly variable (and potentially unstable) collective bargaining policies across the state. Some local governments have continued to block workers’ path to a union contract by rejecting appeals from their own employees to adopt local collective bargaining ordinances (Murphy 2024; Cooper 2025; Lytle 2025; Wilkinson 2025). Because current state law leaves the burden of collective bargaining policy development up to each individual local government, some jurisdictions have expressed interest in or support for collective bargaining while remaining reluctant to invest the necessary time or scarce administrative or legal resources to developing and implementing a local ordinance. Even in larger local jurisdictions with strong collective bargaining ordinances now in place, the “opt-in” system remains fragile and highly vulnerable to instability whenever turnover occurs among elected leaders or administrators with experience necessary to maintain unique local labor-management systems.

Creating a state labor board to provide efficiency and stability for all Virginia public employers and employees

A proposed state labor board equipped to administer a statewide, uniform collective bargaining framework would serve all Virginia state and local government entities and provide consistency, efficiency, stability, and economies of scale. All Virginia public employers and employees would benefit from access to a central, independent state board with capacities to advise public employers and employees about collective bargaining procedures, administer union elections, and mediate contract negotiations. The creation of such state capacities is also especially important at moment when federal labor agencies with similar capacities have been eliminated or rendered non-functional. Indeed, across the country, many states are relying more than ever on their existing state labor boards, and in some cases, are exploring paths to expanding state labor board capacities in order to ensure consistent protections of workers’ rights to unionize and collectively bargain (H.B. 3005; Walter and Madland 2025).

Conclusion

In 2026, Virginia lawmakers should seize the opportunity to enact strong, comprehensive collective bargaining legislation that covers all state and local government workers and creates a state labor board to administer the new system. Under Virginia’s current state law (where collective bargaining is banned for state employees and allowed only for some local government workers), pay for Virginia public employees has lagged far behind that of private sector counterparts with similar education and experience.

Stronger collective bargaining rights can help shrink Virginia’s large public-sector pay gap, reduce racial and gender pay gaps, and improve recruitment and retention of qualified public employees. Removing barriers to unionization for public employees is also a critical step toward reversing the impacts of long-standing anti-worker state policies in Virginia that have for decades suppressed all workers’ wages and contributed to growing income inequality. Lastly, state action to shore up public employee rights is especially important at moment when the federal government is attacking civil servants, public education, health care, and all public services. By extending full collective bargaining rights to historically excluded state and local government workers, state lawmakers can help lead the way to a more vibrant, equitable economy rooted in multiracial democracy in Virginia, the South, and the nation.

Notes

1. If anything, our comparison likely minimizes the public-sector compensation gap in Virginia by comparing benefits for private-sector workers in the South Atlantic region to public-sector workers throughout the country because public-sector data for the South Atlantic region is not available from the Bureau of Labor Statistics. Our comparison does account for the fact that Virginia pension benefits are less generous than public pensions in many parts of the country, though possibly not in two Northern Virginia counties with their own retirement systems. It does not fully account for other differences in benefits between Virginia public-sector workers and their counterparts in other states, though it does account for the fact that public-sector workers in Virginia are covered by Social Security (not true in some states) and adds 1% for public-sector retiree health benefits that are not included in BLS compensation statistics. Finally, it compares all public-sector workers with all private-sector workers, when arguably the better comparison would be between public-sector workers and private-sector workers employed by large employers, who tend to provide more generous benefits than small employers.

References

Biasi, Barbara, and Heather Sarsons. 2022. “Flexible Wages, Bargaining, and the Gender Gap.” Quarterly Journal of Economics 137, no. 1 (February): 215–266. https://doi.org/10.1093/qje/qjab026.

Bivens, Josh, and Heidi Shierholz. 2018. What Labor Market Changes Have Generated Inequality and Wage Suppression? Employer Power Is Significant but Largely Constant, Whereas Workers’ Power Has Been Eroded by Policy Actions. Economic Policy Institute, December 2018.

Borja, Mel. 2022. “How Public-Sector Workers Are Building Power in Virginia.” Working Economics Blog (Economic Policy Institute), February 18, 2022.

Brannick, Priya. 2019. The Janus Impact and Grading of State Public Sector Labor Laws. Commonwealth Foundation, August 2019.

Brannick, Priya M., and Andrew Holman. 2022. The Battle for Worker Freedom in the States: Grading State Public Sector Labor Laws. Commonwealth Foundation, September 2022.

Bureau of Labor Statistics (BLS-CES). 2020–2025. Current Employment Survey, State and Area Employment Data. Accessed January 2026.

Bureau of Labor Statistics (BLS-ECEC). 2020–2024. Employer Costs for Employee Compensation data. Accessed January 2026.

Cantor, Dave. 2025. “Virginia Corrections Department has 2,400 Open Positions.” Virginia Public Radio, August 14, 2025.

Childers, Chandra. 2023. Rooted in Racism and Economic Exploitation: The Failed Southern Economic Development Model. Economic Policy Institute, October 2023.

Childers, Chandra. 2025. “Disinvestment in the Public Sector Undermines Opportunities for Black Women Across the South: Trump Cuts Further Threaten Key Services for Working People Across the Nation.” Working Economics Blog (Economic Policy Institute), July 7, 2025.

Commonwealth Foundation. 2021. “Public Sector Labor Law Database.” Accessed May 9, 2021.

Cooper, David, and Sebastian Martinez Hickey. 2022. Raising Pay in Public K–12 Schools Is Critical to Solving Staffing Shortages: Federal Relief Funds Can Provide a Down Payment on Long-Needed Investments in the Education Workforce. Economic Policy Institute, February 2022.

Cooper, Tracy. 2025. “Portsmouth Workers Frustrated by Private Collective Bargaining Talks.” 13 News Now, July 22, 2025.

Dippel, Christian, and Zachary Sauers. 2019. “Does Increased Union Power Cause Pension Under-Funding in the Public Sector?” (preliminary draft). September 23, 2019.

Economic Policy Institute. 2025a. Current Population Survey Extracts, Version 2025.12.18, https://microdata.epi.org.

Economic Policy Institute. 2025b. How Many Federal Employees Live in Your State? Accessed January 12, 2026.

Flood, Sarah, Miriam King, Renae Rodgers, Steven Ruggles, J. Robert Warren, Daniel Backman, Etienne Breton, Grace Cooper, Julia A. Rivera Drew, Stephanie Richards, David Van Riper, and Kari C.W. Williams. IPUMS CPS: Version 13.0 . Minneapolis, MN: IPUMS, 2025. https://doi.org/10.18128/D030.V13.0.

Frandsen, Brigham R., and Michael Webb. 2017. “Public Employee Pensions and Collective Bargaining Rights: Evidence from State and Local Government Finances.” Hutchins Center Working Paper no. 35, The Brookings Institution, October 2017.

García, Emma, and Eunice Han. 2021. The Impact of Changes in Public-Sector Bargaining Laws on Districts’ Spending on Teacher Compensation. Economic Policy Institute, April 2021.

Gopalan, Maithreyi. 2025. “Why Does Teacher-Student Demographic Representation Matter for Student Success?” HEDCO Institute for Evidence-Based Educational Practice, October 15, 2025.

Han, Eunice S. 2019. “The Impact of Teachers Unions on Teachers’ Well-Being Under Various Legal Institutions: Evidence from District–Teacher Matched Data.” American Educational Research Association 5, no. 3 (August). https://doi.org/10.1177/2332858419867291.

H.B. 1263, 2026 H.D., Reg. Sess. (Va. 2026).

H.B. 2764, 2025 H. D., Reg. Sess. (Va. 2025).

H.B. 3005, 2025 Gen. Assemb. (Ill. 2025).

International Association of Fire Fighters (IAFF). n.d. “Collective Bargaining” (web page). Accessed May 2024.

Khalil, Jahd. 2023. “Hundreds of Richmond City Employees Select Unions.” Virginia Public Media, June 23, 2023.

Lukert, Luke. 2024. “Fairfax Co. Teachers Vote Overwhelmingly for Collective Bargaining Rights.” WTOP, June 11, 2024.

Lytle, Derek. 2025. “Virginia Beach School Board Sits Divided on Collective Bargaining.” 13 News Now, November 11, 2025.

Manzanares, Keyris. 2025. “Virginia Juvenile Justice System Strained by Staffing Shortages.” Virginia Public Media, April 9, 2025.

Martinez Hickey, Sebastian. 2025. “The School Bus Driver Shortage Has Improved Slightly but Continues to Stress K–12 Public Education.” Working Economics Blog (Economic Policy Institute), November 3, 2025.

McGinley, Sophie. 2025. Building a More Equitable Commonwealth: The Case for Collective Bargaining Rights for Virginia State Employees. The Commonwealth Institute, August 2025.

McNicholas, Celine, Margaret Poydock, Heidi Shierholz, and Hilary Wething. 2025. Unions Aren’t Just Good for Workers—They Also Benefit Communities and Democracy. Economic Policy Institute, August 2025.

McNicholas, Celine, Lynn Rhinehart, Margaret Poydock, Heidi Shierholz, and Daniel Perez. 2020. Why Unions Are Good for Workers—Especially in a Crisis Like COVID-19: 12 Policies That Would Boost Worker Rights, Safety, and Wages. Economic Policy Institute, August 2020.

Mishel, Lawrence, and Josh Bivens. Identifying the Policy Levers Generating Wage Suppression and Wage Inequality. Economic Policy Institute, May 2021.

MissionSquare Research Institute. 2023. State and Local Government Employment Trends, 2023 (issue brief). May 17, 2023.

Morrissey, Monique, and Jennifer Sherer. 2024. The Public-Sector Pay Gap Is Widening. Unions Help Shrink It. Economic Policy Institute, August 2024.

Murphy, Ryan. 2024. “Virginia Beach Rejects Collective Bargaining for City Employees.” WHRO Public Media, April 30, 2024.

Nack, David, Michael Childers, Alexia Kulwiec, and Armando Ibarra. 2019. “The Recent Evolution of Wisconsin Public Worker Unionism Since Act 10.” Labor Studies Journal 45, no. 2: 147–165. https://doi.org/10.1177/0160449X19860585.

National Conference of State Legislators (NCSL). n.d. “Collective Bargaining database.” Accessed May 2024.

National Council on Teacher Quality (NCTQ). 2019. “Collective Bargaining Laws database.” Updated January 2019.

National Education Association (NEA). 2020. “Collective Bargaining Laws for Public Sector Education Employees,” November 2020 (unpublished document used by permission).

New Mexico Public Employee Labor Relations Board. 2023. Public Sector Collective Bargaining by State.

Oakford, Patrick, and Margaret Poydock. 2025. “Trump Is the Biggest Union-Buster in U.S. History: More Than 1 Million Federal Workers’ Collective Bargaining Rights Are at Risk.” Working Economics Blog (Economic Policy Institute), September 2, 2025.

Overman, Stephenie. 2023. “In Virginia, ‘Patchwork’ of Ordinances Makes Public-Sector Organizing a Maze.” Virginia Mercury, January 16, 2023

Pierce, Michael. 2017. “The Origins of Right-to-Work: Vance Muse, Anti-Semitism, and the Maintenance of Jim Crow Labor Relations.” The Labor and Working-Class History Association, January 12, 2017.

Pierce, Michael. 2018. “Vance Muse and the Racist Origins of Right-to-Work.” Expert Forum (American Constitution Society), February 22, 2018.

Pope, Michael. 2025. “Fairfax County May Soon Be the Home of the Largest Collective Bargaining Agreement in Virginia.” WVTF, September 10, 2025.

Public Plans Data. 2001–2024. Center for Retirement Research at Boston College, MissionSquare Research Institute, National Association of State Retirement Administrators, and the Government Finance Officers Association. Accessed January 5, 2026.

Rueben, Kim. 1996. “Extended NBER Public Sector Collective Bargaining Law Data Set” [Stata and Excel files]. Downloadable data available at www.nber.org/research/data/nber-public-sector-collective-bargaining-law-data-set (see last paragraph on this web page). Accessed January 12, 2026.

Ruggles, Steven, Sarah Flood, Matthew Sobek, Daniel Backman, Grace Cooper, Julia A. Rivera Drew, Stephanie Richards, Renae Rodgers, Jonathan Schroeder, and Kari C.W. Williams. IPUMS USA: Version 16.0 . Minneapolis, MN: IPUMS, 2025. https://doi.org/10.18128/D010.V16.0

Sanes, Milla, and John Schmitt. 2014. Regulation of Public Sector Collective Bargaining in the States. Center for Economic and Policy Research, March 2014.

S.B. 378, 2026 S., Reg. Sess. (Va. 2026).

S.B. 917, 2025 S., Reg. Sess. (Va. 2025).

Sharma, Rahul Chowdry. 2022. “Where Can Public Sector Employees Collectively Bargain in Virginia? Increasingly, the Commonwealth’s Most Populous Cities and Counties.” Virginia Mercury, July 28, 2022.

Sherer, Jennifer, and Elise Gould. 2024. “Data Show Anti-Union ‘Right-to-Work’ Laws Damage State Economies: As Michigan’s Repeal Takes Effect, New Hampshire Should Continue to Reject ‘Right-to-Work’ Legislation.” Working Economics Blog (Economic Policy Institute), February 13, 2024.

Shimabukuro, Jon O., and Julie M. Whittaker. 2014. Federal Labor Relations Statutes: An Overview. Congressional Research Service (CRS) R42526. Updated September 5, 2014.

Steiner, Elizabeth D., Ashley Woo, and Sy Doan. 2024. Larger Pay Increases and Adequate Benefits Could Improve Teacher Retention: Findings from the 2024 State of the American Teacher Survey. RAND Institute, November 20, 2024.

The Commonwealth Institute. 2022. History of Labor in Virginia: An Interactive Timeline and Map. Accessed January 12, 2026.

Valletta, Robert G., and Richard B. Freeman. 1988. “The NBER Public Sector Collective Bargaining Law Data Set.” Appendix B in When Public Employees Unionize, edited by Richard B. Freeman and Casey Ichniowski. NBER and Univ. of Chicago Press.

Virginia Department of Human Resource Management (VDHRM). 2025. Biennial Compensation Report. November 2025.

Walter, Karla. 2026. Virginia Workers’ Biggest Win in Decades Could Come in 2026. Center for American Progress, January 12, 2026.

Walter, Karla, and David Madland. 2025. Union Trigger Laws 101 How States Can Protect Workers if Federal Labor Law Falls. Center for American Progress, November 19, 2025.

Watts, Parker. 2021. “Labor Day Reflections on Race, Power, and Organized Labor in Virginia.” The Commonwealth Institute, September 1, 2021.

Wething, Hilary. 2024a. “Today’s Teacher Shortage Is Just the Tip of the Iceberg: Part I.”“Today’s Teacher Shortage Is Just the Tip of the Iceberg: Part I.” Working Economics Blog (Economic Policy Institute), October 9, 2024.

Wething, Hilary. 2024b. “Today’s Teacher Shortage Is Just the Tip of the Iceberg: Part II.” Working Economics Blog (Economic Policy Institute), October 16, 2024.

Wilkinson, Nolan. 2025. “Proposal to Allow Frederick City Employees to Unionize Tabled.” The Frederick News-Post, September 25, 2025.

While acknowledging that these legislative achievements led to “some very wonderful things,” President Trump recently mischaracterized this historic period as one in which white people “were very badly treated” amid “reverse discrimination.” The president’s unfounded remarks explain why this administration has directly attacked more than half a century of progress toward racial and economic justice. 

Here are six ways the Trump-Vance administration worked to undermine Dr. King’s legacy and curtail economic justice for people of color in 2025:

1. Making it easier for employers to discriminate by undermining the effectiveness of the Equal Employment Opportunity Commission (EEOC) to enforce Title VII of the Civil Rights Act of 1964 for historically marginalized workers, and by gutting the Office of Federal Contract Compliance Programs (OFCCP). 
2. Hindering equal access to education by dismantling the Department of Education and pushing policies that could limit diversity in higher education, a critical pathway to economic mobility.
3. Effectively eliminating the Minority Business Development Agency, the only economic development agency created to help minority-owned businesses overcome social, economic, and legal discrimination.
4. Cutting spending on the Supplemental Nutrition Assistance Program (SNAP) amid persistently high rates of poverty for children of color and rising food insecurity.
5. Slashing funding for Medicaid and the Children’s Health Insurance Program (CHIP), programs that disproportionately help families and children of color access health care.
6. Undermining health equity through massive cuts to the country’s public health infrastructure, setting the stage for the next health crisis.

The emboldened assertion of white supremacy in our political economy demands a renewed commitment to Dr. King’s legacy of racial and economic justice. In a 1966 essay, Dr. King described economic justice and security as rightful aims in the transition from equality to opportunity. Contrary to Trump’s unsubstantiated claims of pervasive discrimination against white people, both equality and opportunity continue to elude people of color at far greater rates as evidenced by disparate and suboptimal outcomes in employment, earnings, wealth, and even health. Moreover, none of those indicators suggest that white people have been disadvantaged by civil rights enforcement. The immortal words of Coretta Scott King capture the true spirit and impact of the civil rights era and expose Trump’s error and hypocrisy: “Freedom and justice cannot be parceled out in pieces to suit political convenience. I don’t believe you can stand for freedom for one group of people and deny it to others.”

Throughout 2025, the Trump administration has discriminated against Black and brown federal employees—and taken actions that make it easier for all employers to follow suit—by weaponizing the enforcement of antidiscrimination laws against the people they were justifiably created to protect. This includes redirecting EEOC priorities toward so-called “DEI-motivated race and sex discrimination and anti-American national origin bias,” restricting use of disparate impact liability, and effectively ending enforcement of equal employment laws for the civilian federal contracting workforce by gutting the Office of Federal Contract Compliance Programs. The administration’s actions clearly demonstrate how risky it is for workers not to have the protections of a legally binding union contract.

A key element of any union contract is protection from unfair and arbitrary dismissals. For school employees, as for so many, the greatest risk is an employer who plays favorites. Whenever an employer has the unfettered right to decide who stays and who goes, workers suffer. In K–12 education, the risk of layoff is a persistent issue because school districts face endemic funding challenges and are frequently forced to reduce staffing levels. Because educator unions don’t want to give principals and superintendents the right to pick and choose who gets laid off based on their own whims, they have traditionally fought for seniority protections, often known as “last in, first out,” or LIFO. Under LIFO contract provisions, seniority is the sole determining factor in layoff decisions, with newer teachers laid off before more senior ones.

However, LIFO has a tremendous drawback: It hinders efforts to recruit and retain Black and brown teachers. In Minneapolis, for example, only 20% of Minneapolis teachers are people of color, even though fully two-thirds of the student body is Black or brown. Similar patterns are observed nationally. Almost half (49.5%) of K–12 students in the U.S. are Black, Hispanic, or Asian American and Pacific Islander, compared with only 24.4% of teachers. As studies have long documented, LIFO contributes to this disparity because even if a school district is able to hire more Black and brown teachers, they will be the first let go as more senior white teachers are retained.

At the same time, however, teachers’ unions are right to fight for layoff provisions that take away the arbitrary power of school districts to pick and choose who they keep. A core function of unions has always been to protect workers across occupations from being subject to the whims of supervisors. Indeed, Black and Hispanic workers report higher levels of unfair dismissals, suggesting that racial inequities would persist or even get worse in the absence of union protections. Union protections are also critical to narrowing pay disparities. According to a 2024 Rand report, Black teachers received lower average salaries and pay raises than white teachers. This difference was further linked to the fact that Black teachers were less likely to live in states with collective bargaining. The inadequacy of pay is one of the main reasons teachers report for leaving the profession, further contributing to the demographic mismatch between teachers and students.

This is the conundrum that MFE sought to address when the union went on strike in 2022: preserving protections against unfair dismissal while mitigating the inequities of LIFO. The solution they reached in 2022, a solution approved by 76% of MFE’s majority-white membership, was elegant and fair. The contract does not guarantee that Black or brown teachers will be protected from layoffs, contrary to the claims of right-wing groups that the contract is “woke” and “racially discriminatory.” Rather, the contract states that, when the district is forced to lay off teachers, it will protect teachers from populations that are “underrepresented among licensed teachers in the district.”

This means the contract’s protections can and will shift over time, as the composition of the teaching workforce changes. If and when Black teachers are no longer underrepresented in the district, they will no longer be afforded special protections against layoffs. Indeed, if someday it is white teachers who are underrepresented, the same contract provisions would apply to them. Far from embedding racial discrimination into the contract, these provisions support the development of a diverse teaching workforce while protecting worker rights.

The goal of a diverse teaching workforce is not just a noble one but also supports the success and well-being of students of color. Research indicates that the presence of teachers who reflect the diversity of the student body is linked to lower rates of suspension, lower dropout rates, greater college aspirations, and improved test scores. The contract MFE fought for will support the careers of Black and brown teachers and will lead to a teaching staff that looks more like its students, while continuing to protect all educators from arbitrary dismissal. The Department of Justice’s claims are a complete distortion of reality. Sadly, that is what we have come to expect from this administration, which seems dead set on rolling back decades of civil rights protections and abdicating the 60-year position of the federal government in setting a higher standard for employing a workforce that represents the diversity of the U.S. population. Hopefully, the courts will recognize this and allow Minneapolis Public Schools to continue its innovative program to protect a diverse workforce.

Key findings:

• School bus driver employment has increased modestly in the last year but is still 9.5% lower than in 2019. 
• The recent increase appears to be driven by rising wages—school bus drivers have seen 4.2% real hourly wage growth in the past year, the quickest rate since the pandemic.
• However, the end of pandemic relief funds—in conjunction with the instability and attacks on public education by the Trump administration—threaten to reverse this recent progress.

The school bus driver shortage continues to play out across the country, making it more challenging for students to get to school and placing additional burdens on the K–12 public education system. As has been typical in recent years, the beginning of the school year brought forward a steady stream of reports documenting challenges schools are experiencing hiring bus drivers. Our latest analysis finds that school bus driver employment remains 9.5% below 2019 staffing levels.

But there are positive signs that school districts are taking steps to address the shortage. School bus driver employment overall has increased modestly in the last year, with growth in public K–12 schools likely being driven by increasing hourly wages for bus drivers.

It is important to note the available data likely do not fully capture the impact of the ending of pandemic relief funds or the instability for school districts created by the Trump administration. During the summer—a vital time for school district planning and hiring decisions—the Trump administration temporarily withheld $6.2 billion in funds from before- and after-school programs and teacher development. The Trump administration is also seeking to fully dismantle the Department of Education. Harsh anti-immigrant policies are also having harmful impacts on students and education staff. Under these circumstances, more time is needed to get a better sense of how policy changes during 2025 have impacted the K–12 education workforce.

Bus driver employment has grown modestly but remains far below pre-pandemic levels

The shortage of bus drivers is still acute and harmful to working families and their children. This fall, school districts in Missouri, Vermont, and Maine reduced bus routes and other bus services. These types of cuts can eliminate a student’s only way to attend school, including for students with disabilities who rely on buses to attend schools with enhanced special education services. Inconsistent bus schedules and routes can also contribute to absenteeism and missed school meals. Roughly half of all school children use a school bus to get to school, meaning a healthy public education system requires investment in these key support staff.

Figure A shows that there were 21,200 fewer (-9.5%) school bus drivers employed in August 2025 compared with August 2019. The private sector has experienced the largest decrease in employment, despite making up a small share of overall school bus driver employment. There are 12,800 fewer private school bus drivers than in 2019, a decrease of more than a quarter (-28.8%). State and local government school bus driver employment is down overall as well, but much less dramatically (-4.6%).

In the last year, school bus driver employment has grown modestly by around 2,300 jobs.1 This small increase (1.1%) is a step in the right direction, but the trend of the last few years remains mostly flat. Employment growth has been much stronger in the public sector than for privately employed bus drivers. State and local government school bus driver employment has increased by almost 9,900 since the fall of 2024, but private employment has fallen by 8,200 jobs over the same period.

To account for small sample sizes in the Current Population Survey (CPS), our employment analysis uses a 12-month rolling average of data, which means figures reported for August 2025 include data from September 2024. August 2025 data is currently the most recent CPS data available due to the ongoing government shutdown.

Figure A

Rising wages may be driving recent increase in school bus driver employment

The recent increase in school bus driver employment appears to be driven by rising wages for these workers. Recruitment for school bus drivers can be difficult because it often requires a “split-shift” schedule coinciding with the beginning and ending of the school day. It is also a low-wage job, which contributes to school bus drivers experiencing poverty at greater rates than other employed workers. However, Figure B shows that hourly wages have grown steadily over the last year. In August 2025, the median hourly wage for school bus drivers was $22.45, 4.2% greater than last year when accounting for inflation.2

This level of wage growth has not been the norm over the past 15 years. For much of the 2010s, wages for these workers mostly stagnated. Austerity and budget cuts in the 2010s not only contributed to a steady decrease in school bus driver employment but also meant there were few resources available for school districts to invest in school bus driver wages. The apparent wage growth in 2020 was likely influenced by the large compositional changes in the labor market during the pandemic, when large numbers of workers—including bus drivers—dropped out of the labor force.  These dramatic changes in the labor force, in conjunction with the challenges of administering the CPS during the pandemic, mean we shouldn’t draw meaningful conclusions about wages for these workers during that period. More recently, the wage growth for school bus drivers in the last year stands out as a much-needed investment in this critical segment of the education workforce.

Figure B

Other education support occupations still face shortages

Bus drivers were not the only education support occupation that experienced large declines in employment during the pandemic. In 2022, we documented significant employment losses in K–12 education overall, including teaching assistants, school custodians, and teachers. Figure C shows the change in employment between August 2019 and August 2025 for all K–12 education and key occupational categories. Overall education employment slightly exceeds its 2019 levels (1.4%), but the recovery has been uneven across occupation groups. The number of paraprofessionals (teaching assistants and early childhood educators) has grown 16.5% since 2019. However, administrative staff are slightly below their 2019 employment level (-3.0%), while teachers (-4.3%) and food service workers (-4.3%) have experienced more marked declines. Custodian employment is 12.4% below its 2019 levels, an even larger decrease than what school bus drivers have experienced.

Figure C

Federal policy changes threaten recent progress, showing need for state and local action

The recovery in overall education employment has been fueled by the use of pandemic relief funds provided by Congress in 2020 and 2021. The Elementary and Secondary School Emergency Relief Fund (ESSER) allocated $189.5 billion to public K–12 schools to address the impact of the pandemic and reopen safely and effectively. Even for occupations like school bus drivers which have not seen a full recovery, the progress made in the last year has been heavily supported by these federal dollars. These funds ran out at the end of the 2024–2025 school year, and it is too early to say whether the end of this support will reverse this progress.

The Trump administration’s actions are also creating instability for these workers. The Trump administration is gutting the Department of Education, and with it, the oversight of billions of dollars that go to low-income school districts, civil rights protections for students, and special education programs. More threats to public education are on the way, including the creation of a national school voucher program in the Republican-passed reconciliation bill. When fully implemented, this program is likely to expand the use of school vouchers, which will drain resources from public school systems.

A healthy K–12 public education system needs strong bus driver wage growth to continue to bring more workers into the occupation, but instability at the federal level could jeopardize those trends as school districts scramble to account for changes in funding. Bus drivers play a vital role in providing a safe, supportive, and effective K–12 education system. In the face of tremendous federal threats, state and local lawmakers must do everything they can to shore up resources for public schools.

Notes

1. Total bus driver employment includes federal, state and local government, and private-sector workers.

2. School bus drivers tend to work fewer hours than typical workers, but weekly wages are also growing steadily (4.4% growth year-over-year).

The bill reflects a broader national trend: In the lead-up to the fall 2025 school year, roughly a dozen states have passed anti-DEI legislation targeting higher education, including bans on DEI offices and trainings, censorship of equity topics in classrooms, and prohibitions on diversity statements in admissions and hiring. Ohio’s case, however, highlights how this agenda evolves differently across states.

Unlike some states, Ohio has historically maintained a strong state law protecting public employees’ rights to unionize and collectively bargain—including the legally protected right to strike. And unlike many states, Ohio has never adopted anti-union so-called “right-to-work” (RTW) legislation. Indeed, past efforts to strip public employees of their collective bargaining rights (or to introduce RTW legislation) were met with intense and successful public opposition from Ohio voters. That history helps explain why legislators this year combined anti-DEI measures with new restrictions targeting only faculty unions: It is the latest variation in a decades-long pattern of conservative efforts to chip away at worker power and limit equity gains—most recently by demonizing educators and public education systems.

These recent right-wing attacks on public universities, public schools, and unions are not random; they target institutions that have historically expanded racial, gender, and economic equity for working people—which is precisely why they are under assault today.

Cost to teachers and unions

While Ohio’s legislation targets higher education faculty, K–12 unions have also been on the front lines of these fights. Teachers’ unions have been at the forefront of defending against the Trump administration’s anti-union and anti-equity policies, serving as plaintiffs in nearly a quarter of the education-related lawsuits filed this past spring. These lawsuits seek to block the administration’s efforts to dismantle the U.S. Department of Education; eliminate diversity, equity, and inclusion initiatives; and bar collective bargaining for teachers at the more than 160 Department of Defense schools. On the state level, teacher unions have also been fighting to increase funding for early child care, K–12 education, and health care, and to block private school vouchers.

Given the explicit attacks on education at both the federal and state levels, educators and their unions across K–12 and higher education have been increasingly vulnerable to political and legal efforts to dictate what they can teach. In some cases, anti-equity laws are also being paired with anti-union provisions that weaken the right to collectively bargain. Earlier this year, for example, over 50 universities were under investigation from by the Department of Education for allegedly using “racial preferences and stereotypes in education programs and activities.” At the same time, K–12 teachers face the threat of losing their jobs or their schools’ state funding for using inclusive curricula that address race, gender, or sexual orientation. These efforts reflect a broader strategy by state and local governments to curb so-called “indoctrination” and censor education at every level.

In conjunction with state anti-DEI attacks, some states have taken a much bolder approach in disempowering public-sector employees and their right to unionize. This year, Utah Governor Spencer Cox signed a collective bargaining ban across all of Utah’s public sectors including education, transit, and law enforcement. Teachers—the largest group of public employees in the state—are disproportionately affected by the bill. Many educators see the measure as an effort to clear the way for an ultraconservative education agenda. The backlash has been immediate: Voters have already filed a ballot initiative to repeal the law—echoing Ohio’s 2011 SB5 fight, when an attempted rollback of strong public-sector bargaining rights was overturned at the ballot box. In recent years, Utah lawmakers have pushed to eliminate DEI programs, expand school choice vouchers, and restrict transgender bathroom access—the last of which passed earlier this year.

Across the country, state legislators have advanced broad anti-equity measures while simultaneously attempting to weaken education unions to push those agendas forward. This dual threat to public education and organized labor represents a direct attack on worker power and equity, placing both students and teachers at risk.

Undermining equity by undermining unions

Anti-equity policies will likely continue to affect postsecondary curricula, enrollment, and the strength of teachers’ unions. Of the 12 states that enacted anti-DEI legislation this year, many are also states where public-sector workers—including teachers and higher education faculty—have limited or no collective bargaining rights. Research shows that weaker bargaining protections are associated with lower union density, lower wages, and wider gender and racial pay gaps. By weakening organized labor, these policies diminish one of the strongest forces for challenging anti-equity agendas and create conditions for exclusionary laws to advance with less resistance. Across states, however, the more consistent factor is Republican-majority legislatures advancing a nationally coordinated set of anti-DEI and anti-union measures. This overlap underscores how linking attacks on equity and labor reinforces the harm inflicted in both areas.

Taken together, these efforts reveal a coordinated strategy: advancing anti-equity legislation while simultaneously undermining the unions best positioned to resist it. By weakening collective bargaining rights, lawmakers not only silence educators’ voices in shaping policy but also strip them of the power to fight for fair wages, inclusive classrooms, and supportive learning environments.

Ultimately, the erosion of both equity initiatives and labor protections jeopardizes the future of public education and public sector work. These attacks are designed not only to silence inclusive teaching and weaken unions, but also to strip away educational and economic opportunities for Black, brown, and low-income students and to erode good jobs for Black, brown, and working-class educators. Public universities, public schools, and unions have long been imperfect yet critical institutions for advancing racial and gender equity and expanding opportunity. It is precisely because of that role that they now face some of the harshest and most intertwined attacks. Protecting worker power and advancing equity are intersecting goals, and both are essential to building better schools, and a thriving economy, that serves all communities.

Providing teachers with compensation commensurate with that of similarly educated and experienced professionals is necessary to retain and attract qualified workers into the teaching profession. Worsening trends in teacher pay influence students’ career choice. While there are many important factors impacting teacher retention and the recruitment of highly qualified students into the profession, one that consistently lands near the top of any list is pay.3 And closing the growing pay gap between teachers and other college graduate professionals is critical to public education, as teacher quality is the most important school-related factor influencing student achievement.4

Data and relevant information

In analyzing differences in pay between public school teachers and other college graduates, I use two sources of data, both from the Bureau of Labor Statistics (BLS).5 First, I use Current Population Survey Outgoing Rotation Groups (CPS-ORG) data for the weekly wage analyses (BLS 2024a). I focus on weekly wages, as opposed to weekly hours worked or the length of the work year, to account for the “summers off” issue that affects teachers but not other college graduates.6 The sample is restricted to full-time workers (working at least 35 hours per week) aged between 18 and 64, with at least a bachelor’s degree, because teachers today need at least a bachelor’s degree to teach.

The sample is further limited to those who reported their wage information directly (those who didn’t respond and whose wages were estimated by BLS are excluded).7 To preserve data confidentiality, the BLS records weekly wages only up to a defined threshold, so the wage amounts above this threshold aren’t specifically identifiable in the data. This is called top-coding. Historically, the threshold was rarely updated. As a result, a growing share of workers are assigned top-coded wages that are below their actual wages, which has generated a growing understatement of college graduate wages relative to those of teachers. I replace original top-coded values with Pareto-distribution implied means above the original CPS top-code separately for men and women.8 My regression analyses also use CPS demographic variables (e.g., gender, race/ethnicity, state of residence, marital status, and age).

The BLS’s National Compensation Survey’s Employer Costs for Employee Compensation program (BLS 2024b) is the second data source. Specifically, I pull data on employer costs per hour worked for detailed categories of compensation for “primary, secondary, and special education school teachers” in the public sector, and the same data for “civilian professionals,” which is the broadest category available that largely corresponds to college graduates. “Benefits,” in this analysis, refer to employer costs for health and life insurance, retirement plans, and payroll taxes (covering Social Security, unemployment insurance, and workers’ compensation).

The remaining components of compensation are “W-2 wages,” a measure that corresponds to the wages captured in the CPS data used above. W-2 wages are the wages reported to employees and to the Internal Revenue Service. They include “direct wages,” defined by the BLS as “regular payments from the employer to the employee as compensation for straight-time hourly work, or for any salaried work performed” and other wage items, including “supplemental pay.” Supplemental pay includes premium pay for overtime, bonus pay, profit-sharing, and paid leave.

Findings

I present results of this research in four sections. I begin with trends in the simple (not regression-adjusted) average weekly wages for public school teachers and other college graduates from 1979 through 2024 (adjusted for inflation). Second, I report annual estimates of the national teacher weekly wage gap using standard regression techniques to control for systematic differences in age, education, state of residence, and other factors known to affect wage rates. Third, I present the regression-adjusted estimates of the teacher wage gap for each state and the District of Columbia in a figure and a map. Lastly, I factor in nonwage benefits are to estimate a total compensation penalty that accounts for the estimated teacher wage penalty, along with the teacher “benefits advantage,” to estimate a total compensation differential at the national level (which is not possible to calculate for each state).

Simple level differences: Weekly wage trends

The trends in the average weekly wages of public school teachers and other college graduates are shown in Figure A. These data are national annual averages adjusted only for inflation (i.e., not regression-adjusted). It is important to keep in mind that real improvements in living standards require wages to outpace inflation, which has been the case for other college graduates but not for teachers. 

Figure A

As shown in Figure A, the inflation-adjusted weekly wages for teachers were relatively flat from 1996 through 2021, indicating that teacher wages, on average, were just keeping up with the rate of inflation. By 2024, teacher wages were 5.3% less than they were on average in 1996. The average weekly wages of other college graduates also experienced a stretch of stagnation, but for a shorter time span (2002–2014), after which real increases ensued. Since 1996, the wages of other college graduates increased by just over 30%.

Addressing the long-term stagnation of teacher wages requires that future increases in pay exceed future rates of inflation to recover the loss in wages since 2021 and to drive an increasing trend in teacher wages.

Relative differences: Regression-adjusted trends

The average weekly wages discussed in Figure A are simple averages (i.e., they are not regression-adjusted) for teachers and other college graduates; they represent the underlying data used in the regression analyses. Regression estimation helps to account for ways the two groups may differ fundamentally which typically affect pay on margins such as age, educational attainment, race/ethnicity, and state of residence. For instance, all else being equal, one would expect experienced workers to earn more than younger workers who are just starting out in their careers. Controlling for age within a regression model therefore accounts for such differences across the two samples. Thus, standard regression techniques are used to estimate weekly wages of public school teachers relative to other similarly situated college graduates working in other professions, which can provide a more apples-to-apples comparison of earnings.9

Regression-based results are reported in Figure B. They show how much less (or more) teachers earn in weekly wages relative to other college graduates, estimated via regression analysis. A weekly wage “penalty” for teachers is reported when the regression estimates suggest that teachers, all else equal, are paid less than other college graduates. A penalty appears as a negative number in Figure B. When teachers are paid relatively more, the number is positive and is referred to as a “premium.” Estimates are reported for all teachers (which includes a gender control), as well as separately for women and men.

Figure B

The main takeaway from Figure B is the nearly 30-year trend of relative teacher weekly wages increasingly falling behind those of other similarly qualified professionals. Pre-1994, the teacher wage gap averaged 8.7%, but the shortfall worsened considerably starting in the mid-1990s. The teaching penalty hit a record of 26.9% in 2024, which was slightly worse than the penalty recorded in 2023 (26.6%). Otherwise, on average, teachers earned 73.1 cents on the dollar in 2024, compared with what similar college graduates earned working in other professions—much less than the relative 93.9 cents on the dollar that teachers earned in 1996.

Separating the analysis by gender shows that in the pre-1994 period, the relative female teacher weekly wage (i.e., comparing female teachers with other female college graduates) was at a premium that averaged 3.3%. But starting in 1996, the female gap quickly went from parity to a penalty, landing at a 21.5% penalty in 2024.

My previous research (using decennial Census data) confirmed that, over a longer timeframe, the relative wage estimates for female teachers moved from significant premiums to large penalties. For example, I documented that relative female teacher earnings were at a 14.7% premium in 1960, which lessened to 10.4% in 1970 and to near parity in 1980 (pre-1979 years not shown in Figure B). Using the estimates from 2024, the cumulative change has been a 36.2 percentage-point deterioration in the relative wage of female teachers since 1960.10

There is an important story behind the declining relative wages experienced by female teachers. Historically, the teaching profession relied on a somewhat captive labor pool of educated women who had few employment opportunities. This is thankfully no longer the case, but increased opportunity costs are a part of the story and reflected in these results. Expanding opportunities for women enabled them to earn more as they entered occupations and professions from which they were once barred.

In fact, the simple average weekly wages (inflation-adjusted) of female teachers compared with their nonteaching counterparts grew in lock step from 1979 until they started to diverge in the late-1990s. They were close to parity in 1996, when other female college graduates earned just 0.7% more than female teachers. But this divide grew nearly every year—reaching 40.9% in 2024.

Conversely, the trends in the weekly wages of male teachers compared with other male college graduates were never at parity. But like their female counterparts, men also experienced a considerable increase in the pay gap—from 24.1% in 1996 to 81.7% in 2024.11 Therefore, the regression-adjusted relative wages of male teachers have seen sizable penalties throughout the timeframe of this paper (1979–2024) and in my earlier analyses using 1960, 1970, and 1980 decennial Census data. Over the long run, the male relative penalty worsened from 20.5% in 1960 to 36.3% in 2024.12

The growing male teacher penalty partly explains why approximately three in four teachers today are women—a ratio that has not changed much since 1960. The pay penalty experienced by male teachers is unfortunate given the recent statistics and reporting of boys struggling in school. Performing poorly in school is associated with problems encountered later in life—including addiction, mental and physical health issues, and involvement with the criminal justice system.13 Further, Thomas Dee (2010) found that a teacher’s gender has large effects on student test performance, teacher perceptions of students, and students’ engagement with academic material.

So, it is not surprising that today a much smaller share of educated women choose the teaching profession over expanding opportunities with better pay—even as three of four teachers are women. Moreover, the very large male teaching penalty that persists today goes a long way in explaining why men who may want to teach are compelled to choose other career paths, which are on average much more lucrative.

Relative teacher weekly wage penalties by state

Thus far I have reported that the relative teacher weekly wage penalty in the United States was 26.9% in 2024. But there is much variation across the country. To produce regression estimates by state, I pool six years (2019–2024) of CPS data to assure ample sample sizes for each state. Again, I compare public school teachers with nonteacher college graduates within each state and estimate regression-adjusted weekly wage gaps for each state and the District of Columbia.

Figure C

As in previous reports, Figure C shows that in no state does the relative (i.e., regression-adjusted) weekly wage for teachers equal or surpass that of their nonteaching college graduate counterparts. The results are sorted from the largest (38.5%) to the smallest (10.0%) penalties across the United States.

The teaching penalty was at least 25% in 20 states and at least 30% in nine states. In those nine states, teachers on average earn less than 70 cents on the dollar compared with similar college graduates in their respective states—ranging from 69.2 cents on the dollar in Kentucky to 61.5 cents in Colorado.

Figure D depicts a map of the state penalties reported in Figure C.

Figure D

Adding benefits to the analysis

In this section, I examine the teachers’ “benefits advantage” and how it impacts total compensation. The benefits advantage refers to the view that, on average in the U.S., teachers generally receive a larger share of their total compensation as benefits—such as health or other insurance and retirement plans—compared with other professionals. Keep in mind that a larger share of total compensation via benefits means a smaller wage share, given that total compensation is made up of these two components. Here, I calculate how the relatively more generous benefits package for teachers may partially offset the large teacher wage penalty.

The BLS Employer Costs for Employee Compensation (ECEC) series measures the average employer cost per employee hour worked for total compensation, wages and salaries, benefits, and costs as a share of total compensation. I compare benefits packages of primary, secondary, and special education public school teachers with those of comparable workers (specifically, workers in professional occupations).14 Table 1 shows a summary of my calculations.

Table 1

The first two columns in Table 1 under “W-2 wage share of compensation” report the share of W-2 wages that make up total compensation for professionals in all occupations and for state and local K–12 public school teachers. The shares of compensation for W-2 wages and benefits add up to 100. The W-2 shares allow for an examination of how important wages are relative to benefits in the total compensation package.

In 2024, W-2 wages made up 69.6% of teachers’ total compensation, while the share was 78.9% for nonteaching professionals. That means that for every dollar of teachers’ total compensation, 69.6 cents went to wages and 30.4 cents went to benefits. For nonteaching professionals, 78.9 cents went to wages and 21.0 cents went to benefits. Therefore, for every dollar of total compensation, public school teachers receive more in benefits than other professionals, but less in wages. I refer to this as the “benefits advantage.”15

The columns under “public school teachers” in Table 1 provide the information needed to assess total compensation on average for the United States. The “wage penalty” column reports the teacher wage penalty estimates from Figure B, followed by the benefits advantage calculation for teachers. Combining the two gives us a measure of how teachers compare with other professionals on total compensation, which is reported in the last column. Per usual, the benefits advantage for teachers partially offset their estimated relative wage disadvantage, but still left teachers with a significant total compensation gap of -17.1% in 2024—up slightly from -16.7% in 2023. This slight change was due to a 0.2 percentage point decrease in the teacher benefits advantage, and a 0.3 percentage point increase in the teacher wage penalty.

Over the last five years (2020–2024), the benefits advantage that favors teachers varied from 8.8% to 9.9%, but over the same timeframe the teacher wage penalty grew substantially. Thus, in 2024, the teacher total compensation gap widened to -17.1%—the largest on record. Of course, even if the teacher benefits advantage could exceed the large teacher wage penalty, the standard of living for teachers would likely fall, as they would have little in the way of earnings to make ends meet.

Final thoughts

The success of teachers and public education is critically important to students, their families, and communities. It is hard to think of a profession that is more consequential than teaching. After all, one of our highest ideals as a country is to educate each and every child regardless of means, and the future of the U.S. economy depends on this. The highest standard is still worth fighting for, even as we have repeatedly fallen short of the ideal.16

To that end, are teachers sufficiently supported and compensated in the U.S to retain current staff and recruit a pool of highly skilled college students into the profession? The trends documented in this series over the last three decades have no doubt already had profound consequences on teacher retention and recruitment as evidenced in research on teacher staffing challenges (Fortin and Fawcett 2023; NCES 2023), college students forgoing teaching careers citing pay as a main barrier (Croft, Guffy, and Vitale 2018), parents actively steering their children into professions that pay better than teaching (PDK 2019), fast-tracking credentials in response to shortages of permanent teachers (Povich 2023), the heavy use of unqualified teachers (Tamez-Robledo 2023; Lopez and Van Overschelde 2024), and the reliance of unqualified substitute teachers (Franco and Kemper Patrick 2023).

The quality of a public education greatly hinges on our efforts to sufficiently invest in our schools and teachers. This includes the public school workforce and its infrastructure along with all the essential wrap-around services. I have long asserted that providing teachers a standard of living commensurate with similar nonteacher professionals is not simply a matter of fairness. Teacher pay is a central issue in public education; it affects our ability to retain currently credentialed teachers, address teacher shortages, and ensure teaching remains an attractive career option for a large pool of highly qualified students.

Targeted and sustained investments in public education are needed to mitigate (let alone reverse) the growing teacher pay penalty. Funding efforts at the local and state levels, along with support from the federal government, are needed to improve teacher pay and compensation. Additionally, public-sector collective bargaining should be upheld and expanded, given the role of unions in advocating for improved job quality and better pay.

Regrettably, sustained and effective policy interventions capable of mitigating, much less substantially improving, the trends outlined in this long-running series have been lacking. This is a troublesome reality, especially in the United States—a country that has more than enough resources and wealth to be the envy of public education around the world.

Notes

Notes

1. See Allegretto, Corcoran, and Mishel 2004, 2008; Allegretto and Tojerow 2014; Allegretto and Mishel 2016, 2018, 2019; and Allegretto 2023 and 2024.

2. The teacher sample does not include kindergarten or pre-kindergarten; if included, the teacher pay penalties would even larger. 

3. See Blad 2024; Merod 2023; and Steiner, Woo, and Doan 2023.

4. For example, high quality teachers can increase test scores (see Rockoff 2004); students taught by highly effective teachers are more likely to attend college, earn higher salaries, and are less likely to have children as teenagers (see Chetty, Friedman, and Rockoff 2014); international evidence points to a positive association of teacher cognitive skills and student performance (see Hanushek, Piopiunik, and Wiederhold 2019).

5. Allegretto and Mishel 2019, Appendix A provides a comprehensive discussion of the data and methodologies that were used to produce our teacher weekly wage and total compensation estimates. 

6. In Allegretto and Mishel 2019, we provide evidence that teachers work weekly hours similar to those of other professionals.

7. Our earlier work documents that BLS’s imputation method overstates teacher earnings, which is not the case for the other college graduate sample (Allegretto, Corcoran, and Mishel 2008, 9).

8. For more about top-code adjustments, see Economic Policy Institute 2024b.

9. The wage model includes controls for both public and private school teachers. The weekly wage penalty estimates are based on the coefficient on the public school teacher indicator. Regression for all teachers includes a gender control. The percentage gap is calculated as (eb -1) x 100. See Allegretto and Mishel 2019, Appendix A, for specification details.

10. See Allegretto, Corcoran, and Mishel 2008 for 1960, 1970, and 1980 estimates using decennial censuses.

11. Not shown but available upon request from the author.

12. The 1960 results are not shown in Figure B. They can be found in Allegretto, Corcoran and Mishel 2008, 7.

13. See Abrams 2023.

14. The ECEC provides compensation data for a narrower category of “primary, secondary, and special education school teachers” and for a broader category of “teachers.” I analyze the narrower category, which closely matches the definition of teachers in the CPS-ORG data, using data limited to state and local public-sector workers. The inclusion of kindergarten and special education teachers in the benefits analysis does not produce any more substantial differences than if they were excluded (as they are in the CPS sample used to estimate the wage penalty). Greater methodological detail is provided in Appendix A of Allegretto and Mishel 2019.

15. My analysis accounts for differences in annual weeks worked, as it is based on the usual weekly wages of teachers and other college graduates, not hourly wages or annual earnings. One reason health and pension costs are higher for teachers is that teacher health benefits are provided for a full year, while teacher salaries are for less than a full year.

16. See Allegretto, Garcia, and Weiss 2022. This paper describes inequities in public education funding. We also argue that the federal government should play a larger role in funding public education.

References

Abrams, Zara. 2023. Boys Are Facing Key Challenges in School. Inside the Effort to Support Their Success. American Psychological Association, April 2023.

Allegretto, Sylvia A. 2023. The Teacher Pay Penalty Still Looms Large. Economic Policy Institute and the Center for Economic and Policy Research, September 2023.

Allegretto, Sylvia A. 2024. Teacher Pay Rises In 2023—But Not Enough to Shrink Pay Gap with Other College Graduates. Economic Policy Institute and the Center for Economic and Policy Research, September 2024.

Allegretto, Sylvia A., Sean P. Corcoran, and Lawrence Mishel. 2004. How Does Teacher Pay Compare? Methodological Challenges and Answers. Washington, D.C.: Economic Policy Institute.

Allegretto, Sylvia A., Sean P. Corcoran, and Lawrence Mishel. 2008. The Teaching Penalty: Teacher Pay Losing Ground. Washington, D.C.: Economic Policy Institute.

Allegretto, Sylvia A., Emma García, and Elaine Weiss. 2022. Public Education Funding in the U.S. Needs an Overhaul: How a Larger Federal Role Would Boost Equity and Shield Children from Disinvestment During Downturns. Economic Policy Institute, July 2022.

Allegretto, Sylvia A., and Lawrence Mishel. 2016. The Teacher Pay Gap Is Wider Than Ever: Teachers’ Pay Continues to Fall Further Behind Pay of Comparable Workers. Economic Policy Institute, August 2016.

Allegretto, Sylvia A., and Lawrence Mishel. 2018. The Teacher Pay Penalty Has Hit a New High: Trends in the Teacher Wage and Compensation Gaps Through 2017. Economic Policy Institute, September 2018.

Allegretto, Sylvia A., and Lawrence Mishel. 2019. The Teacher Weekly Wage Penalty Hit 21.4 Percent in 2018, a Record High. Economic Policy Institute, April 2019.

Allegretto, Sylvia A., and Ilan Tojerow. 2014. Teacher Staffing and Pay Differences: Public and Private Schools. Monthly Labor Review. U.S. Department of Labor, Bureau of Labor Statistics, September 2014.

Blad, Evie. 2024. Teachers Report Lower Pay, More Stress Than Workers in Other Fields. Education Week, June 19, 2024, sec. Teaching & Learning, Teaching Profession.

Bureau of Labor Statistics (BLS). 2024a. Current Population Survey.

Bureau of Labor Statistics (BLS). 2024b. Employer Costs for Employee Compensation Historical Listing: National Compensation Survey, data tables accessed July 11, 2024.

Chetty, Raj, John N. Friedman, and Jonah E. Rockoff. 2014. Measuring the Impacts of Teachers II: Teacher Value-Added and Student Outcomes in Adulthood. American Economic Review 104, no. 9 (September 2014): 2633–2679.

Croft, Michelle, Gretchen Guffy, and Dan Vitale. 2018. Encouraging More High School Students to Consider Teaching. ACT Research & Policy, June 2018.

Dee, Thomas S. 2010 The Why Chromosome. Education Next, January 26, 2010.

Economic Policy Institute (EPI). 2024a. Current Population Survey Extracts, Version 2025.7.10, https://microdata.epi.org. Accessed August 9, 2024.

Economic Policy Institute (EPI). 2024b. “Methodology: Wage Variables.” EPI Microdata Extracts documentation.

Fortin, Jacey, and Eliza Fawcett. 2023. How Bad Is the Teacher Shortage? Depends Where You Live. New York Times, August 29, 2023.

Franco, Marguerite, and Susan Kemper Patrick. 2023. State Teacher Shortages: Teaching Positions Left Vacant or Filled by Teachers Without Full Certification. Learning Policy Institute, July 2023.

Hanushek, Eric A., Marc Piopiunik, and Simon Wiederhold. 2019. Do Smarter Teachers Make Smarter Students? Education Next, February 20, 2019.

Lopez, Minda, and James P. Van Overschelde. 2024. Unlicensed Teachers Now Dominate New Teacher Hires in Rural Texas Schools. The Dallas Observer, June 20, 2024.

Merod, Anna. 2023. Low Pay, Long Hours Top Reasons Teachers Consider Leaving. K-12 Dive, September 12, 2023.

National Center for Education Statistics (NCES). 2023. Most Public Schools Face Challenges in Hiring Teachers and Other Personnel Entering the 2023-24 Academic Year. October 2023.

Phi Delta Kappan (PDK). 2019. Teaching: Respect but Dwindling Appeal. The 50th Annual PDK Poll of the Public’s Attitudes Toward the Public Schools. Supplement to Kappan magazine.

Povich, Elaine S. 2023. Plagued By Teacher Shortages, Some States Turn to Fast-Track Credentialing. Stateline, July 24, 2023.

Rockoff, Jonah E. 2004. The Impact of Individual Teachers on Student Achievement: Evidence from Panel Data. American Economic Review 94, no. 2; 247–52. May 2004.

Steiner, Elizabeth D., Ashley Woo, and Sy Doan. 2023. All Work and No Pay — Teachers’ Perceptions of Their Pay and Hours Worked: Findings from the 2023 State of the American Teacher Survey. RAND Corporation, September 12, 2023.

Tamez-Robledo, Nadia. 2023. These States Have the Most ‘Underqualified’ Teachers Stepping in to Fill Open Positions. EdSurge, April 4, 2023.

The Trump administration is seeking to hobble or even abolish the Department of Education and undermining federal efforts to support public schools. Earlier this summer they impounded $6.2 billion previously allocated by Congress and put their full support behind efforts to privatize public education in their recently passed 2025 reconciliation package. This provision would provide dollar-for-dollar federal tax credit designed to support private and religious K–12 schools, costing the federal government tens of billions of dollars per year (Davis 2025).

At the state level, the increasing use of universal voucher policies threatens to divert even more funds away from public-school systems and to worsen the quality of education for students remaining in public schools. And in recent years, debates about banning books and diversity, equity, and inclusion curricula decisions have eclipsed discussions on the damage caused by pulling back resources for public schools. Wherever one comes down on debates about the precise curriculum taught in public schools, arguing that it should be taught less with less funding and under-resourced teachers and classrooms, is misguided at best. Yet that is where too many school systems in the U.S. are heading, making it urgent to pay more attention to current trends and debates about the level of resources available to our public schools.

The terrible irony is that these efforts to reduce investments in public schools come on the heels of a recent wave of academic research highlighting just how valuable new investments in public education can be. For decades before the 2010s, it became mainstream to claim that the level of spending on K–12 education was not a primary determinant of student success and that incremental additions to per-pupil spending would disappoint in terms of the extra academic and later life success they would generate. In this view, it was the composition and the use of spending, not the levels, that should be a first-order concern of policymakers. The new wave of research has revealed that this is incorrect: The level of spending per pupil has large effects on student success, and additions to this level of spending are the most reliable educational policy tool we have to help children succeed academically and as adults. This report examines data on educational spending and reviews research linking this spending to various economic outcomes. The sweep of the evidence is clear: The U.S. should devote more resources to ensure that all children have access to an excellent public education and veer off the current destructive path of starving public schools of needed resources.

Key findings:

• The best measure of quality-adjusted spending on public education is per-pupil public spending on K–12 education as a share of per capita potential gross domestic product (GDP). This measure accounts for the fact that as the economy grows richer, key educational inputs (like teachers’ salaries) must rise in line with the rest of the labor market to keep quality constant.
  • This quality-adjusted measure rose throughout the 1970s and 1980s, reflecting a continuous increase in public education investments. However, this spending growth slowed in the 1990s and even fell after the Great Recession in 2008, reflecting society’s lagging investment in public education.
• Increases in educational spending between the 1960s and 1990s were due in large part to an expansion in the scope of services public schools provided their communities—e.g., spending on special education programs or food—and not an increase in instructional resources. Given this, it shouldn’t be surprising that these increases did not directly translate into higher test scores for the general population of public schools.
• Currently, the U.S. ranks extremely low among its peer countries in educational spending, with 22 countries providing more investment effort for public education, as measured by per-pupil spending divided by per capita GDP.
• Disparities in school spending between districts are stark. High-income districts spend more than is necessarily required to meet educational standards, while low-income districts are underspending relative to what is needed to meet basic adequacy thresholds. In 2019, spending in higher-poverty districts was $4,000 below the per student threshold, while spending in lower-poverty districts was $5,700 above the per student threshold. 
• While researchers have debated for decades about the importance of the level of per-pupil spending on student outcomes and the social rate of return of incremental spending level increases, a new wave of the robust research has shown that additional money in public schools directly increases the test scores of public school students, with greater effects for students in low-income districts.
• In high-poverty and medium-poverty districts, there often exist large gaps between current spending levels and the levels needed to lift all students in the district to “adequate” performance on achievement measures.
  • Between 1995 and 2008, the average “adequacy gap” in per-pupil spending expanded as many states significantly reduced education budgets.
  • This increase in the average adequacy gap was much more pronounced in states with a Republican trifecta governing.
• Privatization and universal voucher programs threaten the funding of public schools. There is no evidence that students that take up vouchers do better academically and voucher programs burden state budgets by subsidizing private education.

How much do we spend on public education? Where does that money go?

Figure A shows average per-pupil spending in the United States as well as its composition in 2021 (expressed in $2024). The average per-pupil spending for public schools was $18,777—an unusually high spending level due to the federal Elementary and Secondary School Emergency Relief Funds that went to state and local governments to help them make investments to allow children to regain educational ground potentially lost because of school closures and disruptions during the COVID-19 pandemic. On average, pandemic-related fiscal aid boosted per-pupil spending by $597 in 2021. 

Of the $18,777 spent on average per pupil, nearly $16,422 went directly toward supporting children’s elementary and secondary education, including expenses for instruction and instructional equipment (like technology and school supplies), student services (like transportation, food, and medical services), general administration costs, and operation and maintenance expenses. Less than one-third of per-pupil spending went to salaries, which average roughly $6,089 per student. An additional $1,848 per student went toward capital outlays, which include the construction of buildings. Of the $2,217 “Other” expenditures, $1,131 went to operation and maintenance of buildings, including electricity, heating, and plumbing. Transportation and food, while crucial to student success, were a very small share of overall funding at $576 and $500 per pupil, respectively.  

Figure A

Trends in school funding over time and how to assess its effectiveness

To get a sense of how funding has changed over time, Figure B shows total spending on public education, adjusted for inflation, divided by the number of school-age children in the U.S. since 1960. Based on this measure alone, one may conclude that the nation has displayed an admirable willingness to boost society’s resources for providing public education. For example, between 1960 and 2021, spending per child for public education rose from $3,801 to $18,777. 

Figure B

This upward trend in real education spending is sometimes compared unfavorably with measures of student achievement that didn’t rise as quickly over this period to foster the impression that this additional spending has had disappointingly small payoffs (Elbaum 2023; Mueller 2024). But this conclusion is too simplistic and fails to account for key functions performed by public schools and the fact that they must compete with other sectors for key inputs (like labor).

First, public schools in the United States serve many functions; fostering success in academic achievement evaluations is only one such function. Public schools also are a primary provider of nutrition and health needs for millions of students: Roughly 30 million students received school lunch every day in 2024 (USDA 2025), with over 20 million of these for free or reduced-price. More than 15 million students received breakfast each day in 2024, with over 12 million for free or reduced-price (USA 2025). Schools also are the locus for integrating new residents of the United States into society. And they provide credentials and skills to help students achieve success in the labor market as adults.

Since 1960, the number and scope of functions provided by public schools and the populations they serve have increased dramatically. Rothstein and Hawley (1995) and Rothstein (1997) highlighted the public-school spending functions that grew the fastest in nine representative school districts between 1967 and the mid-1990s. Over time, the functions and populations that have seen the fastest growth have not necessarily been those that might be expected to boost the achievement scores of the general student population. For example, a disproportionately large share of spending growth from 1967 to 1996 could be accounted for by the rise spending on special education—the integration of children with disabilities or particular needs who were historically kept out of public schooling. This particular rise in spending stemmed from society’s increasing recognition that these children and their families deserve the right to public education and its benefits. But this type of spending, while important, will likely not translate directly into improved test scores for the general population of public schools. Of the seven districts surveyed in Rothstein’s report, special education accounted for 3.6% of total expenditures in 1967, but by 1996, this share increased to 19%. Food similarly increased as a share of the budget in these seven districts, from 1.9% in 1967 to 4.8% in 1996.

While these exact spending breakdowns do not exist in the NCES data we examine in this report, the data points we do have confirm that special education remains a large spending priority in K–12 public schools. In 2000, the Special Education Expenditure Project conducted a nationwide survey and confirmed the findings of Rothstein and Hawley (1995) and Rothstein (1997), finding that special education accounted for 21% of overall education spending (Lieberman 2024). Using district-level data on special education spending in 2021 for the 70% of districts that reported special education spending, we found that special education spending accounted for 14.7% of all elementary and secondary education expenditures, still a fivefold increase over what prevailed in the late 1960s. Food spending has also remained high, representing 3.6% of budgets in 2021.

The second reason why comparing school spending with educational outcomes over time can be misleading for drawing conclusions about the efficacy of spending is that schools are in competition with other economic sectors to purchase inputs, like labor. This means that school spending needs to rise in line with overall economic growth to attract the same level and quality of inputs over time. In other words, it is not enough for spending to just rise with inflation—it must also rise with gains in economy-wide productivity. This is especially relevant when we are talking about the teaching workforce. If teachers’ pay only rose with the rate of inflation, their pay relative to other college-educated professionals would rapidly fall and schools would find themselves with a smaller and lower-quality pool of potential teachers to hire. This dynamic whereby teachers’ wages and other school inputs must grow in line with overall economic growth just to maintain a constant level of quality is called the Baumol effect.

Given that public-school spending per pupil must keep pace with economic growth to avoid eroding in its real effectiveness, assessing how well spending trends have maintained educational effectiveness over time requires one to scale this spending against some measure of overall economic growth. Figure C shows potential GDP per capita and education funding per pupil since 2010, in aggregate and in parts. We use 2010 as our starting year because substantial federal fiscal aid to state and local governments was provided in 2009 and 2010 as part of the American Recovery and Reinvestment Act. While this aid was not earmarked for education, state and local fiscal resources are fungible, and any aid meant that it was easier to keep education spending stable in the face of the recessionary shocks. Further, the 2010 elections saw a wave of Republican state governments take power, and this partisan shift is strongly predictive of sharp cutbacks in education spending (as we show later in this paper). We use a measure of potential GDP rather than actual GDP because the steep recession of 2008–2009 left actual GDP significantly depressed, but decisions about long-run investment decisions—like public school investments—should not be conditioned on how high the current unemployment rate is. Instead, they should be conditioned on what is the long-run sustainable base of GDP to finance these investments from.

Figure C

Following the Great Recession in 2008 and the slow ensuing recovery, growth in per-pupil expenditures stagnated and fell further behind GDP growth.

One key change in the overall economy that has made it more difficult for schools to attract the teaching workforce they need is the increase in opportunities for women in the wider job market. While women’s labor force participation rose sharply from the early 1960s until the late 1990s, occupational segregation often confined women to a narrow set of occupations, which included teaching. As women have been able to diversify their career paths, they have also had greater opportunities to compete for higher-paying jobs than teaching. Today, the gap between women teachers’ pay compared with other similarly qualified college educated workers is at its all-time high. Women teaching public K–12 education earn 21.4% less than women with comparable education and demographic characteristics (Allegretto 2024). This “pay penalty”—which has accelerated sharply over the last 25 years—makes teaching a less desirable job, as evidenced by surveys of teacher satisfaction. According to the Survey of the American Teacher, the share of teachers that were very satisfied with their careers was at an all-time high 2008 at 62%, but has subsequently declined by at least 20 percentage points throughout the 2010s and is currently at some of the lowest levels recorded over the last half-century (Kraft and Lyon 2024).

This historical analysis provides insight into how policymakers should determine levels of public support for education. The growth in school-age population has slowed considerably in recent years, a trend likely to continue going forward. Fewer students will somewhat reduce the need for educational investments, but this won’t offset the need to attract quality teachers. Meeting these two needs—to keep up with productivity growth generally and to keep up with a labor market producing more equitable outcomes by gender—is clearly positive for society at large and should lead to more resources being available for public education. But this also means that higher spending levels will be needed just to maintain the current education quality; if policymakers fail to recognize and act on this, the quality of public education will inevitably decline.

How to know if we’re spending enough on public education

The previous section noted that trends in public school spending need to be scaled against economically relevant benchmarks to assess whether the quality of educational services being provided to students over time is rising or falling. We find that when they are properly scaled, quality-adjusted educational investments have been falling in recent decades.

But examining trends over time does not tell us what level of spending is enough for U.S. public K–12 education. Because public education is not provided in markets and because there are so many positive spillover effects associated with education spending, it is inaccurate to assume that the current supply of public educational services in the United States reflects the broad social need for these services. Public investments in education are the outcome of politics and policymaking, and steep shortfalls in these investments from socially optimal levels can persist indefinitely.

Useful benchmarks for assessing whether the U.S. is spending enough on public education include:

• Historical trends: Has the public educational investment “effort”—i.e., the share of the economy’s resources that society is willing to dedicate to public education—kept pace with economic growth in a way that would at least hold educational input quality constant?
• International comparisons: Does the United States expend as much public investment effort as our advanced country peers, and what do recent trends in this effort look like?
• Adequacy measures: Does the United States spend enough to ensure that every school district in the United States can achieve average scores on achievement tests that meet common targets for educational adequacy?
• The social return to additional spending: Do additional increments of educational investment lead to measurably higher social benefits?

Most of these benchmarks indicate that more spending would deliver large social benefits. Before delving deeper into each of these questions, it is important to highlight a clear sign that public spending on education may be insufficient: the well-established fact that teachers routinely dip into their own pockets to supplement school supplies for their students. Figure D shows the average amount that teachers spend out of their pay to supplement school supplies by state for the 2020–2021 school year. Teachers in South Dakota spend the least of their own money on school supplies that school year ($315), while teachers in Nevada and Alaska spend the most ($587 and $691, respectively).

Figure D

Historical trends in public investment effort in education

As noted above, the stability of public education spending as a share of GDP does not guarantee a constant quality of education over time (unless enrollment is shrinking). As productivity rises in sectors of the economy that are more amenable to automation and other labor-saving processes but largely remains constant in sectors that require intensive face-to-face contact (like education), maintaining the same amount and quality of inputs into public education will require spending to be a rising share of GDP.

How rapidly public education should be rising as a share of GDP depends on the growth of the share of school-age children (ages 5–18), but also depends on productivity growth in the rest of the economy. If this growth is fast in historic terms, then public education’s share of overall GDP should rise rapidly just to keep pace in terms of the amount and quality of inputs needed to retain or attract away from other sectors. If this productivity growth is slow in historic terms, then public education’s share of overall GDP can rise more slowly and yet remain constant in effective terms. One rough proxy that accounts for both enrollment and productivity growth is per-pupil public spending on K–12 education divided by per capita GDP. To keep recessions (which reduce GDP) from pulling this measure around in uninformative ways we use a measure of potential GDP from the Congressional Budget Office (CBO), which estimates how much GDP could be produced in the country if unemployment was maintained at low levels. If this measure of per-pupil spending as a share of per capita GDP is falling, then society’s educational investment effort can be said to be lagging. Figure E highlights the recent history of this measure of spending.

Figure E

This measure of educational investment effort rose throughout the 1970s and 1980s, reflecting a continuous investment in public education. However, spending on K–12 education as a share of per capita GDP slowed in the 1990s and outright stalled after the Great Recession in 2008, reflecting a lagging public education investment effort.

International comparisons

Another benchmark for assessing whether the U.S. invests enough in public education is how we compare with our international peers. A well-educated workforce is a key driver of economic growth and competitiveness. If the United States is failing to match a high standard compared with international peers, this could make U.S. relative economic performance suffer.

Figure F provides a simple ranking of countries based on average public per-pupil spending on K–12 education, expressed as a share of per capita GDP in each country in 2019. This is in line with our measure above on educational investment effort. The United States does not rank highly, with 22 countries surpassing its investment effort for public education.

Figure F

Figure G shows the change in this per-pupil spending since 2010 for the 29 countries with data in each year. Again, the United States ranks in the bottom third of countries on this metric, with an outright decline in educational investment effort. The U.S. is therefore not only spending less on this measure than most other advanced economies, but it is also reducing this spending while dozens of our advanced country peers are increasing it.

Figure G

Estimated returns to new increments of public education spending

In the 1990s and early 2000s, the conventional wisdom in many education policy discussions was that the U.S. government was spending “enough” on K–12 education, and that marginal additions to spending per pupil would likely not generate improved student outcomes (Hanushek 2006). This conclusion led policies and initiatives to focus on changing how educational resources were deployed rather than increasing the average level of public resources.

Some of these initiatives statistically monitored teacher performance to remove teachers deemed ineffective. Others focused on increasing the share of students educated in charter schools or private schools—including through the expansion of voucher programs which diverted public funds to pay for private-school tuition. But the outcomes of such initiatives have been largely ineffective or show little sign that they could be widely scaled as improvements for public education.

More recent research overturns this older conventional wisdom and highlights just how damaging a detour these initiatives may have been, demonstrating that in fact incremental increases in education spending translate directly and reliably into significant gains in student outcomes and provide a much more reliable lever with which to improve educational quality in the United States.

Crucially, this wave of new research has overcome the statistical challenge of trying to isolate the causal effect of higher spending levels of educational and social outcomes. Public education spending is not random and depends on a number of factors that are correlated with student success. For example, spending in a given district might rise as higher-income families move into the area and property values (and taxes) increase. These higher-income families might also be able to provide greater in-home resources that will aid their children’s academic performance. Simple correlations between the level of district spending and student success might hence show a positive relationship, but the causality would not necessarily be running from district spending decisions to student success; both might instead be driven by a third variable, which is simply the level of family resources on average across the district.

Running the opposite direction, much of school funding is explicitly compensatory, targeting students facing greater socioeconomic disadvantage to attempt to even out total resources (both in home and public) available to students for academic success. But if this greater spending targets students with fewer in-home resources, it could show a negative relationship between levels of spending and student performance, and again would not be reflecting the causal effect of this spending.

Newer research has found natural experiments that allow truly exogenous changes in school spending to be identified, and hence the effects on student achievement to reflect the causal effect of this spending.

This research can be divided into three strands:

• Pre-COVID-19 research focused on court-ordered school finance reforms (SFRs) as the natural experiment;
• Pre-COVID-19 research on other sources of exogenous variation in school spending as natural experiments; and
• Post-COVID-19 research on the large increase in federal fiscal aid to state and local governments earmarked for education as the policy experiment. 

Pre-COVID-19 research on court-order school finance reforms

Jackson, Johnson, and Persico (2015) used longitudinal data to trace the educational trajectories and adult earnings outcomes for a nationally representative set of students who were exposed to court orders mandating changes to a state’s school finance system. These court-ordered reforms imposed discrete and significant changes in a state’s school finance system that differentially affected students in different districts within a state and/or who were born in different years—allowing researchers to isolate any variation coming only from changes in spending. The authors found that a 10% increase in school spending for 12 years led to 7.7% higher wages and a 9.8% increase in family income in adulthood. Large positive effects were also found for high school graduation, completed years of education, and a reduced incidence of poverty in adulthood.

Other studies have used the school finance reforms as part of their quasi-experimental design to understand the relationship between educational funding and student outcomes. Candelaria and Shores (2019) studied more recent reforms and found positive effects on high school graduation rates. Rothstein and Schanzenbach (2022) studied the recent wave of adequacy reforms and found a 2.0 percentage point increase in high school graduation and a 1.4 percentage point increase in college enrollment. Similarly, Hyman (2017) concluded that students exposed to 10% more funding were 3.0 percentage points more likely to enroll in college and 2.3 percentage points more likely to graduate from college. Lafortune, Rothstein, and Schanzenbach (2018) studied the recent wave of adequacy reforms and found an increase in test scores for students in low-income districts which closed the gap between low- and high-income districts by one-fifth over the 10 years following the reforms. Biasi (2019) studied the effects of these reforms on intergenerational mobility and found that reducing the gap in funding between low- and high-income districts disproportionately increased the income of children with parental income in the 10th and 25th percentile later in life as adults. The new evidence therefore shows how much public schooling can deliver for students when schools are adequately resourced.

Other pre-COVID-19 natural experiments

Court-ordered school finance reforms are not the only source of exogenous variation in spending that can be used to assess the causal effect of more resources for public education. Jackson and Mackevicius’ (2024) literature review of 32 papers highlighted research that uses discontinuities in state funding formulas or changes in local elections to assess the effect of higher spending levels. Their overall finding is that higher levels of school spending cause higher levels of student achievement, measured either by test results or the probability of attending college. Their baseline estimates are that each $1,000 in additional per-pupil spending raises achievement scores by roughly 3.2% of a standard deviation, and raises the probability of college attendance by nearly 3 percentage points—implying that the benefits outweighed the costs of this higher level of spending.

Other papers also use the introduction of new capital investments to assess the value of higher educational spending. Cellini, Ferreira, and Rothstein (2010) look at referenda on bond initiatives to finance new education investment that either passed or failed by narrow margins. Districts where bond referenda narrowly passed and new investments came online should differ from districts where such measures narrowly failed only by the impact of this new investment. They find that treated districts—those that saw a narrow passage of a bond referendum to finance new education investments—saw identifiable increases in student test scores after just a few years. Most impressively, treated districts saw immediate and large increases in home prices. This strongly indicates that the public highly values well-funded schools, and that no market exists to let them easily “buy” these well-funded schools directly—after all, it took a political intervention (the passage of a bond issue through public referenda) in order for these homeowners to secure the gain in value spurred by higher school investments. Specifically, Cellini, Ferreira, and Rothstein (2010) found that every $1 in increased educational investment made possible by narrowly passed bond referenda results in an increase of home prices in the district of more than $1.50. If educational investments were set in an efficient market, one would expect these investments to rise until the return to them (in the form of higher home prices) was just $1. This sizable gap in the cost versus returns to educational investment signals clearly that the public would value higher educational investments.

Lafortune and Schönholzer (2022) took a similar approach to Cellini, Ferreira, and Rothstein (2010) in examining a large increase in educational facility investments in the Los Angeles Unified School District (LAUSD) between 1997 and 2008. Over this period, voters in LA approved nearly $20 billion in state and local bond issues to finance increased educational expenditures. This large investment effort proceeded in numerous phases. Lafortune and Schönholzer (2022) used the staggered rollout of new facility investments to essentially see which students in the LAUSD were treated with new investments. They then measured the causal effect of these new investments on student achievement and surrounding home prices. They found significantly positive effects on test scores, attendance, and home prices. Like Cellini, Ferreira, and Rothstein (2010) they found that each $1 in new spending resulted in $1.62 in increased household value, mostly through upward pressure on home prices as demand for homes in districts receiving greater investment increased. Again, these findings strongly suggest that relative to efficient market outcomes, there is an underinvestment in public education relative to the public’s underlying demand.

Biasi, Lafortune, and Schönholzer (2025) also used the bond referenda approach described above. They find that bond authorization raises test scores and is most beneficial in districts with more disadvantaged student populations. This disproportionate effect arises for two reasons. First, these districts prioritize bonds to finance spending that directly improve student learning instead of simply improving school amenities (like an athletic field or school theatre). Second, underinvestment in these districts in the past have led to investment deficits, leading to disproportionate gains when these deficits are addressed.

Assessing the effectiveness of COVID-19 fiscal relief for education

During 2020 and 2021, three separate legislative aid packages provided fiscal aid to state and local governments with funds earmarked for education. This dedicated education aid was spurred by policymakers seeking to blunt the effect of the COVID-19 pandemic on students’ education. Most of the aid was given directly to local school districts without state or federal oversight and was not tied directly to district-level education disruptions caused by the pandemic. The fact that aid was not targeted by pre-existing situations in local school districts means that it can be treated as a purely exogenous change in spending and hence its effect on subsequent educational performance can be interpreted as causal.

The Education Recovery Scorecard (ERS) project (a joint research effort of Harvard and Stanford) assessed the contribution of federal educational aid to subsequent outcomes (see Dewey et al. 2025). Essentially, the ERS approach measures the relationship between federal pandemic relief spending per student and the change in average student achievement between 2022 and 2023, while controlling for several district-level characteristics that might affect achievement (like measures of community poverty).

The overall ERS finding is almost exactly in line with the effect of extra money estimated before the COVID-19 pandemic: Each $1,000 increase in per-pupil spending increased student achievement by just under 1% of a standard deviation. The effect was more pronounced for math than for reading scores. When they restricted their assessment to high-poverty districts, the per-dollar effect was nearly double. They concluded that their results “imply that the federal pandemic relief contributed to academic recovery during the 2022–2023 school year, and that the impacts were in line with what would have been expected from prior research.” What makes this especially informative, however, is that the federal aid pushed per-pupil expenditures significantly above pre-pandemic levels, and even this significant increase saw no decline in the marginal boost it provided to student achievement. This discredits any argument that we have saturated public education spending and are nearing the point where more money would stop yielding social returns.

Where net new spending on public education could do the most good

The burgeoning literature we highlighted shows that higher levels of public spending on K–12 education could generate large social benefits—benefits that are reflected directly in higher measurable student outcomes. Moreover, when schools get more investment, home prices increase, reflecting the families’ underlying demand for investment in public education. These higher home prices reflect buyers’ valuation of being in districts with well-funded public schools and show just how much they’ll pay in order to be near one. This new literature argues for paying more policy attention to the level of public resources for education, not just how these resources are deployed. But this research can also inform where new additions to public resources for K–12 education could do the most good.

The literature cited in the previous section generally found that the overall improvements in student outcomes spurred by higher levels of spending are driven disproportionately by students raised in low-income households (Jackson and Mackevicius 2024). Lafortune, Rothstein, and Schanzenbach (2018) studied the recent wave of adequacy reforms and found that the test score gap between low- and high-income school districts within a state fell by roughly 10% of a standard deviation 10 years after the reforms. Biasi (2019) studied the effects of these reforms on intergenerational mobility and found that a $4,500 reduction in the funding gap between the richest and poorest districts in a state—slightly larger than the average funding gap in 1990 ($4,012)—improved upward mobility: Students exposed to these changes born in the bottom 10% of parental income had a 5.6 percentile higher rank in adult income, which corresponds to an increase in adult earnings of 16.2%.

Another strand of research directly estimates the additional resources that would be needed to ensure that every public school student in the United States received an adequate education (Baker, Di Carlo, and Weber 2024)—defined as the level of funding needed in a district to ensure students reach an average level of student achievement.

This “adequacy gap” between what schools need to hit these metrics versus what they’re currently receiving in public resources is larger for districts where neighborhood poverty is higher. This is for two reasons. First, per-pupil spending is often lower in low-income districts. Second, students in low-income districts need more funding per pupil to compensate for the much lower level of out-of-school resources available to them. Higher-income parents can spend more on educational enrichment activities, can often spend more time helping children with schoolwork, and often have higher educational credentials themselves, placing them in a better position to guide children academically.

To understand how these differences in overall per-pupil spending affect different components of educational spending, Table 1 shows overall per-pupil spending by district income and spending for detailed expenditure categories. The table shows that schools in high-income districts spend nearly $2,000 much more on educational spending per pupil, compared with low-income districts. Children in poorer neighborhoods end up going to schools that are less resourced than those in well-off neighborhoods, leaving teachers and instructional staff the herculean task of providing similar education with fewer resources.

Table 1

The disparities in funding for teachers’ salaries and benefits highlight a particular struggle that low-income districts face in recruitment and retention. High-income districts spend $6,899 per pupil on salaries for teachers and support staff personnel, while low-income districts only spend $5,983. This trend is true for regular instruction, but also for special education instruction, with high-income districts outspending low-income districts by nearly 46% ($920 per pupil compared with $670). This trend can be particularly damaging because schools in high-poverty districts are likely to have a greater share of the student body with special education needs.

Student services refer to all the services schools provide beyond instruction, including transportation, pupil and instructional services, operations and maintenance, and business/central operations. High-income districts spend about $1,115 per pupil while low-income districts spend only $874 per pupil.

Again, this gap can be particularly damaging because students in low-income districts likely require more resources for pupil and student services beyond instructional time to compensate for the gap in out-of-school resources available to them relative to higher-income students.

One area where the funding allocation is actually higher in low-income districts than in high-income districts is food services. More students in low-income communities rely on free and reduced-cost lunch programs financed with public dollars, while food service is functionally privatized for moderate- and higher-income students. Finally, more is spent on adult education in low-income districts, because community members there are more likely to rely on public education for compensatory education (not shown Table 1).

Estimating what is actually required for universal adequacy

To assess the adequacy of school spending in high- and low- income districts, Figure H uses data from the School Finance Database to show the actual per-pupil spending in 2019 for low- and high-income districts, as well as the money that is required in those districts to provide an adequate education. We use spending in an earlier year (2019) to avoid any temporary spending from the COVID-19 Elementary and Secondary Spending programs that buoyed state spending in 2021 and better represent a baseline level of school funding available to most students.

Baker, Di Carlo, and Weber (2024) provide estimates on what level of spending is required for an adequate education, using a cost model to calculate adequate funding levels for most of the nation’s public-school districts. The cost model incorporates factors that affect students’ needs in a given district, such as Census child poverty rate, its labor costs, student characteristics, and teacher labor markets. These factors help illustrate the resources available outside the school that affect student learning outcomes. For example, students in high-poverty district will have fewer resources at home and thus require more resources in their school to achieve a given outcome goal, like a score on a standardized test, relative to students in more affluent areas. Similarly, teacher labor markets, a key education cost, vary across districts and thus have implications for student achievement. The model then provides, for each district, the cost amount of spending required to achieve an adequate education, which they determine to be national average math and reading scores in grades three to eight.

Some districts already spend more (some substantially more) than the adequate threshold of education spending, while others spend substantially less. This should not be a shock; there are hundreds if not thousands of well-resourced school districts in the United States providing a level of educational quality that would rival anywhere else in the world. The problem is that this is not the norm. Figure H shows that while low-poverty districts spend more than is necessarily required to meet educational standard, high-poverty districts are not spending enough to meet basic adequacy thresholds. In 2019, high-poverty districts underspent by $4,000 per student, while low-poverty districts overspent by $5,700 per student.

Figure H

Challenges to financing our schools fairly and adequately

This section highlights two among the many policy challenges to providing enough funding to guarantee an adequate public education for all U.S. children. First, there is the long-running structural challenge of so much public-school finance depending on local property taxes. The second challenge is a political one that stems from the fact that Republican-dominated state governments almost universally fail to provide enough public investment in K–12 education, with this problem becoming nationally acute in the 2010s as the number of Republican trifectas (control of both governorship and state legislatures) grew.

Local property taxes and disparities in public education funding

Educational funding is skewed toward high-income districts in part because public education in the U.S. relies heavily on revenue from local property tax revenues. Districts with higher-priced homes will be able to raise more revenue for a given tax rate, giving them a funding advantage over others. In 2019, for example, 46% of per-pupil education funding came from local revenue (overwhelmingly property taxes), 44% came from state revenue sources, and the remainder came from the federal government. This funding model in which local revenue is so important inevitably generates inequalities between districts.

Figure I shows the trends in average amount of per-pupil revenue spending, by type of government revenue, for high- and low- poverty districts over time. In 1995, local per-pupil revenue for high-poverty districts was $4,450 and $7,820 for low-poverty districts—a gap of $3,370. Yet by 2021, $12,110 of per-pupil revenue came in from local government in low-poverty districts while only $7,444 per pupil revenue was from local government in high-poverty districts—a gap of over $4,500 per pupil. The overall funding gap is closed a bit by revenue from federal and state sources in high-poverty districts exceeding that of low-poverty districts, but these state and federal resources are not nearly large enough to offset the inequality created by property tax financed local revenue.

Figure I

Republican-dominated states invest the least in public education

Between 2007 and the onset of the COVID-19 pandemic, there was extreme pressure on all sources of educational funding. The large decline in home prices in the years after 2007 led local property tax revenue to decline significantly. The American Recovery and Reinvestment Act (ARRA) of 2009 provided significant federal fiscal aid to state governments until the end of 2010. While this aid was not specifically dedicated to education, it was largely fungible and should have allowed states that wished to maintain resources to education help offset the property tax declines.

ARRA aid to states, however, ran out by the end of 2010. Further, the 2010 election cycle saw a large increase in the share of state governments dominated by Republican trifectas. These Republican trifectas passed extremely austere state budgets, with large cuts to education. This move toward austerity was mirrored at the federal level, with Republican control of Congress after 2010 leading to the Budget Control Act, which forced steep cuts across many areas of spending. These various funding pressures become extremely visible in the data starting in 2011.

Figure J shows overall per-pupil spending from 1995–2021, categorized by whether states did or did not have a Republican trifecta in 2011. States with a Republican trifecta in 2011 (Arizona, Florida, Georgia, Indiana, Idaho, Kansas, Louisiana, Maine, Mississippi, North Dakota, Ohio, Oklahoma, Pennsylvania, South Carolina, South Dakota, Tennessee, Texas, Utah, Wisconsin, and Wyoming) spent less than states without a Republican trifecta both before and after 2011, but the trend of spending between the two groups was identical leading up to the Great Recession. Between 1995 and 2008, the average gap in per-pupil spending was 13% (and was essentially  the same in each year). In the aftermath of the Great Recession, and particularly following the election cycles of 2010, many states significantly reduced education budgets, but this move was much more pronounced in Republican trifectas: The gap in spending grew to an average of 18% between 2011 and 2021.

Figure J

While Figure J illustrates simple averages in per-pupil spending for states with and without Republican trifectas in 2011, the two groups of states may differ in factors like the business cycle, or state- or year-specific factors that might also play a role in educational funding. For instance, all else being equal, education funding is likely to face downward pressure in periods when there is high unemployment.

Using a regression model to account for variation in state-specific unemployment rates can help account for this influence on education spending. Table 2 shows results from standard regression models that estimate per-pupil spending in states with a Republican governance trifecta relative to other states between 2011 and 2021. The table shows the difference in per-pupil spending between states that have and do not have a Republican trifecta, while removing the influence of state-specific unemployment rates. On average, states with a Republican trifecta spend $141 less per pupil than states without a Republican trifecta, controlling for the factors discussed above. For districts in high-poverty neighborhoods, being in a state with a Republican trifecta is associated with $244 less in per-pupil spending, while for districts in low-poverty neighborhoods, Republican trifectas are actually associated with an increase in student spending of $32 (although the estimate is not statistically significant).

Table 2

Recent threats to public education

More recently, Republican-led states have introduced voucher and tax credit programs that put even further fiscal pressure on resources available to public school students. Vouchers programs, which funnel public dollars to private schools, have been growing in recent years. While these programs have been around for decades—following the Brown v. Board of Education ruling in 1954, several Southern states used vouchers as a tool to undermine school integration efforts (Ford, Johnson, and Partelow 2017)—more recently, states have enacted voucher programs to attract a wider pool of students (called “universal voucher programs”) allowing any student to use public funds to pay for private education. In 2000, the number of students using vouchers stood at just 11,000, but today the number of students enrolled in voucher programs is over 600,000.

The beneficial effect of vouchers on student outcomes is weak both in theory (Wething and Bivens 2024) and in evidence. In Louisiana, Ohio, and Indiana, test scores for students in voucher programs declined following program enactment (Mills and Wolf 2017; Figlio and Karbownik 2016; Waddington and Berends 2018). Moreover, in places with mature voucher programs like Milwaukee, Wisconsin, nearly 40% of the private schools that took vouchers failed or closed within the program’s first 25 years, displacing students and creating uncertainty for students in need of a quality education and parents looking to send their kids to school. Studies show that students that experience a school closure are more likely to do worse on tests, attend school less, and drop out, relative to students who don’t experience school closure (Kim 2024). For private schools to have such a high closure rate suggests that the schooling alternative to public schools was tenuous and often not well-funded in the first place, and will likely harm students’ education prospects. 

Voucher programs directly threaten public-school funding in two main ways. First, they increase the cost to the state for students to attend private schools. If students had never intended to attend public school, these vouchers just pile on new fiscal obligations to the state. Any new fiscal obligation will obviously threaten to put downward pressure on education spending (one of the bigger parts of any state and local government’s budget). This pressure has been shown to be significant. In Arizona, only $33 million was allocated to for the first year of universal eligibility for their Empowerment Scholarship Account voucher program in 2022–2023,1 but it ended up costing $587 million in the first year and upwards of $708 million in fiscal year 2024 (Griffith and Burns 2024).

Second, vouchers also impose under-recognized indirect costs for students left in the traditional public-school system, as schools try to maintain the same level of education amid declining enrollment. This is because public schools—like nearly every other economic entity in the modern economy—require a mix of inputs, some of which are variable and some of which are fixed. Regardless of whether schools lose students due to voucher programs, some costs will have to be paid the same amount (e.g. electricity, heat, building maintenance, payments on debt, capital improvements). Costs that can change with enrollment take on a lion’s share of the downward adjustment when enrollment is reduced due to students leaving public schools for voucher programs. These adjustable costs include vital inputs into effective education like teachers’ salaries and textbooks. We refer to these indirect costs as the fiscal externality (a change in per-pupil spending on variable costs that come from sharp reductions in enrollment) and have created a calculator for districts to model fiscal externality costs in their districts (Wething 2024).

Conclusion

Universal public education supported the U.S. becoming one of the richest countries in the world in the 20th century (Goldin 2001). Getting there required consistent and stable funding to public schools and sufficient levels of spending to provide a strong education. Our research shows that public education funding in the U.S. has stalled out in recent years and, coupled with threats from voucher policies in states and at the national level, will harm future citizens by robbing them of the basic education and support they need to be capable citizens. Moreover, this policy decision, which was mostly in response to austerity measures by Republicans during the Great Recession, is not binding and could change with policies to fully fund public education. New evidence shows that money in education can make a huge difference in students’ educational attainment, such as improved test scores and their likelihood of finishing college, and their earnings well into their adulthood.

Notes

1. Ariz. H.R., HB 2853 (2022) Fiscal Note (June 16, 2022), https://www.azleg.gov/legtext/55leg/2R/fiscal/HB2853.DOCX.pdf

References

Allegretto, Sylvia. 2024. Teacher Pay Rises in 2023—but Not Enough to Shrink Pay Gap With Other College Graduates. Economic Policy Institute and Center for Economic and Policy Research, September 2024.

Baker, Bruce D., Matthew Di Carlo, and Mark Weber. 2024. The Adequacy and Fairness of State School Finance Systems. School Finance Indicators Database, January 2024.

Biasi, Barbara. 2019. “School Finance Equalization Increases Intergenerational Mobility: Evidence from a Simulated-Instruments.” Journal of Labor Economics 41, no. 1.

Biasi, Barbara, Lafortune, Julian, and David Schönholzer. 2025. “What Works and for Whom? Effectiveness and Efficiency of School Capital Investments Across the U.S.” The Quarterly Journal of Economics. https://doi.org/10.1093/qje/qjaf013.

Bureau of Economic Analysis. Various years. “Real Gross Domestic Product per Capita.” Public data series accessed through the Federal Reserve Economic Database. Accessed March 2025.

Bureau of Economic Analysis, National Income and Product Accounts. Various years. “Nominal Potential Gross Domestic Product.” Public data series accessed through the Federal Reserve Economic Database. Accessed March 2025.

Candelaria, Christopher A., and Kenneth A. Shores. 2019. “Court-Ordered Finance Reforms in the Adequacy Era: Heterogeneous Causal Effects and Sensitivity.” Education Finance and Policy 14, no. 1: 31–60. https://doi.org/10.1162/edfp_a_00236.

Cellini, Stephanie Riegg, Fernando Ferreira, and Jesse Rothstein. 2010. “The Value of School Facility Investments: Evidence from a Dynamic Regression Discontinuity Design.” Quarterly Journal of Economics 125, no. 1 (February 2010): 215–261. https://doi.org/10.1162/qjec.2010.125.1.215.

Davis, Carl. 2025. “Megabill Takes Cap Off Unprecedented Private School Voucher Tax Credit, Potentially Raising Cost by Tens of Billions Relative to Earlier Version.” Institute of Taxation and Economic Policy blog, July 2, 2025.

Dewey, Daniel C., Erin Fahle, Thomas J. Kane, Sean F. Reardon, and Douglas O. Staiger. 2025. Pivoting from Pandemic Recovery to Long-Term Reform: A District-Level Analysis. Center for Education Policy Research, February 2025.

Elbaum, Jack. 2023. “Test Scores Have Barely Risen Since 1970 Despite 245% Spending Increase.” Washington Examiner, June 22, 2023.

Ford, Chris, Stephenie Johnson, and Lisette Partelow. 2017. The Racist Origins of Private School Vouchers. Center for American Progress, July 2017. 

Figlio, David, and Krzysztof Karbownik. 2016. Evaluation of Ohio’s EdChoice Scholarship Program: Selection, Competition, and Performance Effects. Thomas B. Fordham Institute, July 2016.

Goldin, Claudia. 2001. “The Human-Capital Century and American Leadership: Virtues of the Past.” Journal of Economic History 61, no: 2 (June): 263–292.

Griffith, Michael, and Dion Burns. 2024. Understanding the Cost of Universal School Vouchers: An Analysis of Arizona’s Empowerment Scholarship Account Program. Learning Policy Institute, February 2024.

Hanushek, Eric A. 2006. “School Resources.” Handbook of the Economics of Education 2: 865–908, http://dx.doi.org/10.1016/S1574-0692(06)02014-9.

Hyman, Joshua. 2017. “Does Money Matter in the Long Run? Effects of School Spending on Educational Attainment” American Economic Journal: Economic Policy 9, no. 4 (November 2017): 256–280. 

Jackson, C. Kirabo, Rucker C. Johnson, and Claudia Persico. 2015. “The Effects of School Spending on Educational and Economic Outcomes: Evidence from School Finance Reforms.” Quarterly Journal of Economics 131, no. 1: 157–218. https://doi:10.1093/qje/qjv036.

Jackson, C. Kirabo, and Claire L. Mackevicius. 2024. “What Impacts Can We Expect from School Spending Policy? Evidence from Evaluations in the United States.” American Economic Journal: Applied Economics 16, no. 1 (January 2024): 412–446. https://doi.org/10.1257/app.20220279.

Kim, Jeonghyeok. 2024. “The Long Shadow of School Closures: Impacts on Students’ Educational and Labor Market Outcomes.” EdWorkingPaper no. 24-963 (Annenberg Institute at Brown University), May 2024. https://doi.org/10.26300/ax4m-3z14.

Kraft, Matthew A., and Melissa Arnold Lyon. 2024. “The Rise and Fall of the Teaching Profession: Prestige, Interest, Preparation, and Satisfaction over the Last Half Century.” EdWorkingPaper no. 22-679 (Annenberg Institute at Brown University), April 2024. https://doi.org/10.26300/7b1a-vk92.

Lafortune, Julien, Jesse Rothstein, and Diane Whitmore Schanzenbach. 2018. “School Finance Reform and the Distribution of Student Achievement.” American Economic Journal: Applied Economics 10, no. 2: 1–26. https://doi.org/10.1257/app.20160567.

Lafortune, Julien, and David Schönholzer. 2022. “The Impact of School Facility Investments on Students and Homeowners: Evidence from Los Angeles.” American Economic Journal: Applied Economics 14, no. 3 (July 2022): 254–289. https://doi.org/10.1257/app.20200467.

Lieberman, Mark. 2024. “How Much Does Special Education Truly Cost? Finally, an Answer Is on the Horizon.” Education Week, August 26, 2024.

Mills, Jonathan N., and Patrick J. Wolf. 2017. “Vouchers in the Bayou: The Effects of the Louisiana Scholarship Program on Student Achievement After 2 Years.” Education Evaluation and Policy Analysis 39, no. 3: 464–484. https://doi.org/10.3102/0162373717693108.

Mueller, Chris. 2024. “No, US Student Ranks Didn’t Plummet After Department of Education Creation.” USA Today, November 22, 2024.

National Center for Education Statistics. 2025a. “Table 236.60. Total and Current Expenditures Per Pupil in Fall Enrollment in Public Elementary and Secondary Schools, by Function and Subfunction: Selected School Years, 1990–1991 Through 2020–2021,” Digest of Education Statistics. Accessed February 2025.

National Center for Education Statistics. 2025b. “Table 236.55 Total and Current Expenditures per Pupil in Public Elementary and Secondary Schools: Selected School Years, 1919–1920 Through 2020–2021”, Digest of Education Statistics. Accessed February 2025.

National Center for Education Statistics, Common Core of Data. Various years. School District Finance Survey (F-33). Accessed March 2025.

National Center for Education Statistics, National Teacher and Principal Survey. Various years. “Public School Teacher and Private School Teacher Data Files.” Accessed March 2025.

Organisation for Economic Co-operation and Development. Various years. Indicators of Education Systems Programme. Accessed March 2025.

Rothstein, Jesse, and Diane Whitmore Schanzenbach. 2022. “Does Money Still Matter? Attainment and Earnings Effects of Post-1990 School Finance Reforms.” Journal of Labor Economics 40, no. 51.

Rothstein, Richard. 1997. Where’s the Money Going? Changes in the Level and Composition of Education Spending, 1991-96. Washington, D.C.: Economic Policy Institute.

Rothstein, Richard. 2024. “The Soft Bigotry of High Expectations.” Working Economics Blog (Economic Policy Institute), January 30, 2024.

Rothstein, Richard, and Karen Hawley. 1995. Where’s the Money Gone? Changes in the Level and Composition of Education Spending. Washington, D.C.: Economic Policy Institute.

U.S. Census Bureau, Small Area Income and Poverty Program Estimates. Various years. “School District Estimates for 2000.” Accessed March 2025.

U.S. Department of Agriculture (USDA). 2025.Child Nutrition Tables: National Level Monthly Data. Accessed May 2025.

Waddington, R. Joseph, and Mark Berends. 2018. “Impact of the Indiana Choice Scholarship Program: Achievement Effects for Students in Upper Elementary and Middle School.” Journal of Policy Analysis and Management 37, no. 4: 783–808. https://doi.org/10.1002/pam.22086.

Wething, Hilary. 2024. How Vouchers Harm Public Schools: Calculating the Cost of Voucher Programs to Public School Districts. Economic Policy Institute, December 2024. 

Wething, Hilary, and Josh Bivens. 2024 “Vouchers Undermine Efforts to Provide an Excellent Public Education for All.” Working Economics Blog (Economic Policy Institute), May 15, 2024.