The Truth about Teacher Pay

Andrew G. Biggs & Jason Richwine

Fall 2019

One of the most common beliefs about American education is that teaching is an "underpaid" profession. Think tanks purport to calculate the "teacher pay gap." The media run stories about teachers taking second and third jobs to pay the bills. Politicians call for across-the-board raises. They all see raising teacher pay as a matter of simple fairness, as well as a way to attract better teachers and improve educational outcomes.  

They are all misguided. The highly publicized "pay gap" that dominates news headlines is the product of a simplistic methodology that, when universally applied, suggests that nurses, firefighters, and other professionals are dramatically overpaid. Furthermore, predictions generated by the underpaid-teacher hypothesis — such as that teachers must have high quit rates, or that a large percentage of their income flows from second jobs — are not supported by the data. Teachers as a group are generally well compensated, and teacher pay and benefits have risen faster over time than compensation in private-sector jobs. Failure to recognize these facts can lead education reform down a blind alley.

Across-the-board raises, the usual solution to closing the teacher pay gap, come with high price tags. West Virginia's teacher walkout ended with the state legislature passing an across-the-board 5% salary increase. Arizona's teacher protests culminated in Governor Doug Ducey agreeing to a 20% salary increase over two years, a policy that will cost the state and schools over $450 million per year, in addition to higher pension costs based on the increased salaries. Democratic presidential candidate Kamala Harris has proposed to close the "gap" using federal funds, at an estimated 10-year cost of $315 billion. These are not costs that are incidental to government budgets.

Moreover, focusing on across-the-board raises distracts from less costly but more useful reforms, such as differential pay for hard-to-staff subjects, increased teacher mobility through experience credits and portable pensions, loosened tenure protections, and a reduction in non-teaching staff.

Finally, an inordinate focus on teacher salaries feeds unrealistic expectations for the profession. Although teacher quality certainly matters, most of the variance in student achievement is associated with factors outside the classroom. Just as current teachers should not be blamed for "failing schools," policymakers should not simply assume that investing heavily in teacher salaries is worth the political and economic costs. The most prudent course would be to implement modest structural reforms, while de-emphasizing the level of teacher pay as a focal point of education reform.


The union-affiliated Economic Policy Institute (EPI) issues an annual report on the "teacher pay gap" — the difference in salaries between teachers and similar private-sector workers. This year, EPI finds that the gap is 21% nationwide. At the state level, the gaps vary widely, from a salary penalty of just 0.2% in Wyoming to a high of 32.6% in Arizona. Whenever pay disputes arise in a state, the media treat EPI's findings as authoritative, rarely bothering to include a contrary view. After all, everyone knows that teachers are underpaid.

Reporters might be more skeptical if they realized that EPI's own pay-gap methodology leads to some other conclusions that are, to put it delicately, less intuitive. Using the same Census data and the same basic techniques that EPI applies to teachers, we find that registered nurses are "overpaid" by 29%. Meanwhile, telemarketers deserve a big raise, as they currently suffer a 26% salary penalty. Aerospace engineers are apparently overpaid by 38%, but "athletes, coaches, and umpires" are paid 21% less than their skills are worth. Photographers should consider going on strike, as they make 16% less than comparable workers. Firefighters are moochers by contrast, taking in 25% above their rightful salaries.

If all this sounds ridiculous, it's because EPI's method is so simplistic. To arrive at its 21% pay gap, EPI merely compares teacher salaries with the salaries of people who have roughly the same number of years of education and the same demographic characteristics. More specifically, EPI performs a regression analysis using Census Bureau survey data, in which respondents provide information on their salaries along with their age, education, region of residence, marital status, and other factors that are predictive of earnings. Included in this analysis is a "dummy variable" indicating whether the individual is a public-school teacher. The coefficient on the dummy variable represents the effect on salary of being a teacher after controlling for all of the other factors listed above.

While controlling for worker characteristics is the right idea, this model clearly cannot measure important differences among workers in different occupations. The argument breaks down when we drop the teacher dummy variable from EPI's regression and replace it with any other occupation, then observe the often-large salary premium or penalty associated with it. In fact, about four in every 10 occupations we analyzed show an alleged wage premium or penalty greater than the one EPI claims for teachers. We could write dozens of reports with titles such as "The Electrician Pay Premium" or "The Massage Therapist Penalty," all modeled directly on EPI's analysis of teachers. No one would publish these articles, however, because alleging significant pay penalties or premiums in nearly every occupation makes no sense. Such results say much more about the method of analysis than they do about America's labor market.

The control variables in EPI's statistical model obviously cannot explain the significant salary differences across occupations — otherwise there would not be so many occupations with large premiums or penalties. The model fails because salaries are determined by the supply and demand for specific skills that vary across occupations even after controlling for education. Therefore, in the context of comparing occupations, educational attainment is simply too imprecise as a skill measure. EPI's analysis treats every bachelor's degree as identical, and yet no one is surprised or upset that people with engineering degrees earn more on average than people with literature degrees, nor does anyone believe that every occupation requiring a college degree should be paid the same.

Yet these distinctions get muddled when comparing teacher salaries to those of private-sector employees. Around 95% of workers are in occupations where the average education level is below that of the typical teacher. Years of education (or highest degree attained) are good predictors of salaries for private-sector workers, and they're great predictors of pay for teachers, where salary schedules are often explicitly based on educational attainment. But in comparing teachers' salaries and private-sector pay, educational attainment is a skewed variable because it assumes that quantity of education equals quality of education.

Indeed, the problem is even more complicated: While teachers with more formal education earn more than less-educated teachers, more-educated teachers are not necessarily better at teaching. As the Urban Institute's Matthew Chingos puts it, "The fact that teachers with master's degrees are no more effective in the classroom, on average, than their colleagues without advanced degrees is one of the most consistent findings in education research."

Alternative measures of skill indicate how fragile EPI's results are. The Bureau of Labor Statistics assesses the skill requirements of hundreds of U.S. occupations, grading them on the General Schedule scale that is used in setting federal salaries. The BLS concludes that teaching positions demand skills that are roughly a GS-8 on the federal scale, with average GS levels ranging from 8.2 to 8.4 depending on the type of teacher.

The chart below shows the BLS grades of a variety of occupations graphed against the median annual wage paid in those occupations. To tighten the analysis further, only occupations with an average educational attainment of a bachelor's degree or more are included. Once again, there is significant variation in annual salaries even among jobs with similar skill requirements. Only half of occupations are paid within 10% of what their profession's BLS grade predicts. In other words, it's normal for occupational salaries to be above or below what occupational skill requirements might predict, and we could hardly assume that every deviation is evidence that an occupation is improperly paid. Unlike in the EPI analysis, however, teachers (the black dots on the chart) receive a salary premium of 9% once their shorter work year is accounted for. So not only are skill measures inadequate to explain pay variation across occupations, but the specific choice of skill measure can actually change the direction of the result.

As more variables are accounted for, the weakness of EPI's model becomes more evident. Texas A&M economist Lori Taylor finds that teachers are nearly twice as likely as other college graduates to live in rural areas, where both wages and the cost of living are lower. In Taylor's analysis, controlling tightly for geography reduces the teacher salary gap by nine percentage points, because it more accurately compares teacher salaries to those of other professionals in the areas where teachers live and work.

EPI also understates the critical role of fringe benefits, especially pensions. According to the National Income and Product Accounts (NIPA), which are the official ledger books of the U.S. economy, employees in public education received benefits — inclusive of the future pension benefit they accrue each year — equal to 45% of their annual wages. In the private sector, benefits averaged only 19% of wages. By itself, this benefit advantage is sufficient to negate a teacher salary penalty of up to 17%.

In sum, EPI's "pay gap" reports do not meaningfully inform the debate over teacher compensation. Unless one is willing to accept that nurses are overpaid, that telemarketers are underpaid, and that engineering majors should be paid the same as literature majors, then one cannot accept the claim that teachers suffer from a 21% salary gap.


Of course, just because EPI's teacher-pay methodology is weak does not mean we can dismiss its conclusions outright. Teachers may still in fact be facing a large and growing pay penalty. A better way to test a hypothesis is to evaluate whether it generates accurate real-world predictions. When we consider specific outcomes we might expect to see if public-school teachers suffered the pay gaps that EPI claims for them, we find the predictions are not borne out by the data.

For example, if teachers were underpaid, a large proportion might leave the profession because of low pay. But in fact, teachers quit their jobs at much lower rates than private-sector workers do. If quit rates are signals of an occupation's relative attractiveness, then teaching is far more attractive than most private-sector jobs. So why are there so many news stories about teachers quitting? Part of the confusion is that quit rates among education workers, including teachers, are historically high right now, according to data from the Job Openings and Labor Turnover Survey (JOLTS). But quit rates among private-sector workers are also high, for the simple reason that the labor market is tight. When job opportunities abound, employees switch jobs more often. The teacher quit rate runs roughly parallel to the private-sector quit rate — it goes down during recessions and up during expansions. The difference is that, at all times, the teacher quit rate is much lower.

Furthermore, a "quit" in the JOLTS data doesn't necessarily designate a teacher leaving the profession; it could mean moving between teaching jobs. According to the most recent School and Staffing Survey, just 7.5% of teachers are true "leavers" each year, a category that includes teachers who retire or who stay in the education sector in a non-teaching capacity, such as administration. By far the most common reason offered for leaving is "personal life factors," with only around 5% of leavers citing "salary and other job benefits" as their motivation.

Finally, if teachers were truly underpaid by 21% on average, we should expect large wage gains when they move to the private sector, especially since the teachers who would most gain from switching jobs are the most likely to actually switch. But the evidence shows nothing like these kinds of wage increases. Our analysis of the Survey of Income and Program Participation finds that teachers who switch to a different profession earn 3% less in their new jobs. Separate studies using administrative records in Florida, Missouri, and Georgia also find no average wage increase for ex-teachers.

Another consequence we would expect if teachers were underpaid is widespread teacher shortages. According to a trend-setting New York Times article from 2015, "districts are struggling with shortages of teachers...[as] a result of the layoffs of the recession years combined with an improving economy in which fewer people are training to be teachers." Today, an internet search of news media sources finds 262,000 references to "teacher shortages," often attributing such shortages to low teacher pay. And yet, as Kate Walsh of the National Council on Teacher Quality points out, there is "no real data — because neither states nor the federal government collect the data that's really needed to pronounce the onset of a teacher shortage." Instead, claims of a teacher shortage are heavy on anecdote and speculation.

Consider the 2016 report from the Learning Policy Institute's Linda Darling-Hammond and colleagues. They compared predicted teacher vacancies with the predicted number of graduates from teacher-training programs, concluding that the supply of new teachers would soon fall well short of demand. But this report is too speculative to be relied upon. For instance, the researchers assume that U.S. class sizes will fall and then calculate an insufficient number of teachers to staff them. If class sizes remain at today's levels, which are themselves much smaller than in the past, the report's teacher shortage disappears.

This is not the first such prediction. Darling-Hammond's 2016 report mirrors her earlier study, which proclaimed that "a general shortage of teachers is imminent." That report appeared in 1984, when she predicted teacher shortages beginning by 1985 and reaching 20% by the late 1980s. But the shortage never manifested. From the fall of 1987 through the fall of 2015, the number of public-school students increased by 20%, but the number of public-school teachers increased by 64%. More recently, in the four years leading up to the 2015-16 school year, teacher employment grew by 400,000, even as the number of students barely changed.

U.S. colleges continue to graduate more than enough prospective teachers to fill job vacancies as they occur. For instance, economist James Cowan and his co-authors find that in 2011, the most recent year for which Schools and Staffing Survey data were available, less than 100,000 teachers reported being in their first year of teaching. In that same year, the Education Department reports that states granted over 280,000 initial teaching credentials, suggesting that the pool of potential teachers was substantially larger than the number of positions that typically come open.

But are schools attracting new teachers simply by watering down standards? In fact, the quality of new teacher applicants has been rising. The University of Washington's Dan Goldhaber and Joe Walch find that "teacher applicants and new teachers in recent years have significantly higher SAT scores than their counterparts in the mid-1990s." The fact that schools have been recruiting better-qualified teachers simply by demanding them hints that teaching jobs were always in reasonably high demand among college graduates.

All of this occurred during a period in which — if EPI's figures are to be believed — teacher compensation became far less competitive, with the teacher wage penalty rising from 8% in 1987 to 21% in 2018. One would think that trend would degrade the skills of new teachers rather than enhance them. Yet again, a prediction generated by EPI's widely cited teacher pay gap claim is not borne out by the data.

This isn't to say teacher shortages don't exist in particular districts and in particular subject areas. They do. But isolated teacher shortages are not evidence of a broad teacher salary gap. Rather, shortages exist precisely where expected in a nationwide labor market that pays an increasing premium for STEM and other specialized skills. When researchers have examined the teaching vacancies that districts say they have trouble filling, they find that elementary, English, and social-studies teachers are not the problem. In fact, the Department of Education found that 20 states and the District of Columbia produced over twice as many elementary-education graduates as they had elementary-teaching positions to fill. At the same time, some districts struggle to fill STEM and special-education positions.

Data from the state of Connecticut, which has published unusually detailed data on teacher vacancies, tend to confirm this finding. Connecticut public schools received 113 applicants per open K-6 teaching position, with applicant quality ranked a four on a one-to-five scale. Connecticut schools received 85 applications for each open position in middle-school history. Overall, Connecticut received about 42 applicants for each open teaching position, with a typical applicant pool offering "many acceptable applicants." (This is a state where EPI says teachers suffer a wage penalty of 16.7%.) However, in middle- and high-school science, Connecticut schools received just 12 applicants per job opening, and the applicant quality ranked only a two on a one-to-five scale. This is what we would expect if teachers specializing in different areas had different job prospects outside of teaching.

Unfortunately, paying more for teachers in hard-to-staff subject areas is not something that most districts can do. Union contracts generally mandate that gym teachers must be paid the same as calculus teachers, with the predictable result of surpluses of gym teachers and shortages of calculus teachers. Differential pay would send a message to prospective teachers that certain fields are in greater demand than others. Across-the-board pay increases, by contrast, do little to address shortage areas because they provide no incentive for prospective teachers to specialize in areas of particular need. The result will be that shortages will continue, prompting even more calls for across-the-board salary increases.

Yet another assumption we could make given the widespread belief that teachers are underpaid is that their jobs might be especially demanding and stressful compared to similarly compensated work. Teaching is no doubt a challenging job. Many teachers have told us that we wouldn't last one day in the classroom, and we're inclined to agree. Nevertheless, it's important to remember that other occupations can be difficult too. For example, teachers frequently say they work long hours due to lesson planning or paper grading performed outside of the classroom, but other occupations also require long hours: nurses endure shift work; salesmen constantly travel; lawyers sleep in their offices. Rather than trade anecdotes, we need to look at the data. Participants in the American Time Use Survey (ATUS) list their activities in detailed time diaries. The diaries capture work time wherever and whenever it may occur. Using the ATUS, we find that teachers work an average of 40.6 hours during the work week, compared to 42.4 hours for private-sector professionals. Over a full calendar year including vacations, teachers work at most 83% as many hours as private-sector professionals.

Moreover, teaching jobs are not particularly stressful or unpleasant compared to other occupations. The BLS Occupational Information Network (O*NET) examines almost 1,000 occupations, providing details on multiple job characteristics that might plausibly be associated with pay. These include factors such as the consequences of error, the importance of being exact or accurate, the level of competition, time pressure, dealing with physically aggressive people, and responsibility for outcomes and results. On these and similar measures, teaching is a middle-of-the-pack occupation. Nursing, a profession that EPI's methodology implies is significantly overpaid, ranks much higher in these O*NET occupational requirements.

Job insecurity is another source of work-related stress. "Many of us have relatives or friends who were dismissed from their schools during the recession," reads one EPI report, which goes on to cite those potential layoffs as stressful for teachers. The reality, as shown in the JOLTS data, is that job losses in public education are infrequent relative to the private sector. Since 2000, the layoff and discharge rate in public education has often been less than one-third the private-sector rate. Even during the Great Recession, teachers' monthly layoff and discharge rate rose from 0.4% in 2007 to a peak of just 0.8% in September 2009 — a year in which the private-sector layoff rate peaked at 2.3%. Layoffs are much like quit rates — teacher advocates point to them as a downside of teaching, when in fact the problem is much greater in the private sector.

Likewise, a 2017 Brookings Institution report finds that wage inequality among public-school teachers is far lower than in other occupations. In other words, workers who pursue teaching can be much more confident in what their future salary will be, in contrast with other occupations where high pay for winners is offset by low salaries for losers. In almost all respects, teaching is a more stable occupation than alternatives in the private sector.


If teachers are as underpaid as EPI insists, we might expect to see a large proportion of them working second jobs. Indeed, news stories often feature sympathetic teachers who are forced to work second jobs in order to make ends meet. For instance, the New York Times Magazine reports,

Some teachers devote 60 hours a week to the classroom, then go to work elsewhere. The hours can be long, the labor physical, the pay close to minimum wage. Teachers across the country are now baristas, Amazon warehouse employees, movie-theater managers and fast-food grill cooks. They're entering the gig economy in off hours and struggling to stay awake during school days.

Since teaching is a profession whose compensation is weighted less toward salaries and more toward retirement benefits and a shorter work year, it's understandable for teachers to want some extra spending money, and summers provide a convenient opportunity to earn it.

The media's anecdotes, however, are inconsistent with the data and often less impressive than they seem. For instance, the Times writes of a 28-year-old single teacher living in Lexington, Kentucky who after five years of teaching earns $48,000 a year. She works extra jobs as a movie-theater manager and a sales representative at a Disney store, pulling in an additional $14,000 per year. Working three jobs sounds like desperation, until we consider that this teacher's $48,000 base salary is higher than roughly three-quarters of 28-year-old Kentucky residents working full time. In fact, barely one in 10 full-time Kentucky workers in that age range earns more than her $62,000 total. So while these examples may appear compelling to readers in high-cost, high-wage metropolitan areas, in a local context they are far less so.

Moreover, the vast majority of teachers do not have second jobs. The National Teacher and Principal Survey finds that just 18% of teachers have a second job, while the Current Population Survey (CPS) puts the number at 14% — barely more than the 11% of full-time non-teacher workers who have second jobs. Furthermore, only around 2% to 3% of total income earned by teachers (including those with only one job) comes from secondary work. As for super-long work weeks, only 7% of public-school teachers self-report working more than 60 hours per week in total, including any second jobs.

Digging a little deeper, we can use secondary incomes to test EPI's claims about a large and rising teacher pay gap. If the teacher pay gap pushed teachers to moonlight, we would expect to find that more teachers work second jobs in states where EPI says teachers suffer from the largest pay penalties. Likewise, as the teacher salary gap has supposedly increased over time, teachers nationally should now be more likely to work second jobs than they were in the past.

But the data do not support either prediction. We compared EPI's state-level teacher pay-gap figures to CPS data showing both the percentage of teachers with secondary income and the share of total income that teachers obtain from second jobs. There is essentially no relationship in either case.

We also compared EPI's nationwide pay gaps over time, ranging from 12.4% in 2003 to 21.4% in 2018, to the share of earnings that teachers receive from second jobs. In this case, the results are opposite to what EPI would predict — the share of teachers' total earnings coming from second jobs dropped from 2.8% to 2.2%, and the share of teachers with any second-job income fell from 17% to 13%. Once again, EPI's claimed 21% teacher salary gap fails to generate an accurate prediction.

Despite all the contrary data, the media cannot seem to resist sensationalized stories about teachers struggling. Last year, Time magazine published an article purporting to explain "What It's Like to Be a Teacher in America." In portraying teachers as barely able to put food on the table, Time took the teachers-are-underpaid claim to an absurd level. One teacher says she cannot afford to see her doctor; another works two extra jobs and donates blood "to pay the bills."

Once again, the data do not support the anecdotes. According to our calculations using Census data, the median teacher lives in a household with a total income (from all earners) of $106,000. Teachers have a poverty rate of 1.1%. Their unemployment rate is 0.7%. This is not the profile of people with a hand-to-mouth existence.

Similarly, the Federal Reserve's Survey of Household Economics and Decisionmaking inquires about financial well-being. Among teachers in the survey, 81% said they are either "doing okay" or "living comfortably," while only 2.7% of teachers describe their financial situation as "finding it hard to get by." Teachers are just as likely as other college graduates to say they are better off than their parents were at the same age, equally satisfied with the neighborhoods that they live in, and no more likely to be carrying student debt.

It is common to read that the average teacher salary in a state has barely grown over time, or has even shrunk once inflation is counted. One recent report from an education news outlet claims that the average teacher salary today is $1,000 lower than in 1989, after adjusting for inflation. In fact, data from the National Center for Education Statistics show that salaries, which are almost always based on years in the classroom, have grown substantially in almost every category of teacher experience. What has happened is that the teaching workforce has grown younger as Baby Boom-era teachers retire and the average teacher has fewer years of experience.

Moreover, data from the NIPA show that total compensation (including benefits) in the public-education sector has been growing relative to the private sector. In 1952, the average public-education employee — the largest category of which is K-12 school teachers — received total compensation that fell 1% below that of the average private-sector employee. By 2017, average public-education compensation exceeded private-sector levels by 22%, the highest compensation premium ever. These figures simply compare averages to averages; they do not attempt to account for changes in teachers' or private-sector workers' education or other earnings-related characteristics. Nevertheless, the data demonstrate that teachers have risen steadily in the overall income distribution. If teaching was a middle-class occupation in the past, it must be at least middle class today.


It is clear that the widely cited 21% teacher salary gap is a meaningless statistic. Furthermore, predictions generated by the underpaid-teacher hypothesis fail to be borne out by the data: Teachers rarely quit their jobs and, when they do, rarely cite low pay as the reason; only a tiny percentage of teachers' salaries come from second jobs, and that percentage has been falling over time; there is no generalized teacher shortage; most teachers live comfortable middle-class lives; and teaching is not more stressful or time-consuming than the average job. Over and over again, we fail to find evidence that teachers as a group are underpaid.

It is more likely that workers in public education are on average overpaid, in the sense that they could not earn as much in the private sector. Studies of teachers who switch jobs and comparisons of teaching to other occupations with the same BLS skill requirements suggest that the teacher wage penalty is close to zero. If that is true, then incorporating fringe benefits as measured in the NIPA would boost total teacher compensation about 18% above private-sector levels. This premium comes before adding the value of job security and the predictability of regular raises, which economic theory predicts would be offset by lower wages.

The most important policy lesson is that lawmakers should be wary of simple, across-the-board teacher pay raises. Arguments that across-the-board raises will improve teacher quality ignore the fact that, for many years to come, the vast majority of those salary increases will go to current teachers. If policymakers are displeased with the quality of the existing teaching workforce, raising pay for current teachers — and thus giving them less reason to quit or retire — would shrink the pool of new teachers entering the classroom.

New teacher quality is already rising. In fact, school districts should simply continue what they have been doing: improving applicant screening to identify those with skills most likely to lead to better student outcomes. Research from the 1990s found that schools often failed to hire better-qualified applicants even when presented with them, but the recent improvement in new-teacher qualifications appears to be tied to a more stringent application process.

Such considerations lead inevitably to the fraught question of whether teacher pay should be restructured to favor teachers who perform best on the job. "Merit pay" proposals sometimes cause teachers to bristle, and understandably so, as the implication is that teachers are to blame for educational problems. Too often policymakers respond to low test scores by condemning "failing schools," as if successful education is just a matter of finding better teachers and principals.

In reality, many outside factors affect learning, and not all of them have been trending in a positive direction. For example, in 1980, around 24% of school-aged children lived with a single parent, and just 9% lived in a household where a foreign language was spoken; today, 32% live with a single parent, and 22% are raised in households speaking a foreign language. These are the types of challenges that schools did not cause and that we cannot expect teachers to easily overcome on their own. Even more important, classroom achievement will always be constrained by the abilities of the students themselves. The psychologist Douglas Detterman finds that "the majority of the variance in educational outcomes is associated with students, probably as much as 90% in developed economies." From that perspective, blaming teachers for "failing schools" is unfair.

This is an argument that cuts both ways, however. One cannot assert that teachers are nearly powerless to overcome outside obstacles to student learning and simultaneously maintain that teachers require higher salaries for the vital work they do. Teachers' unions in particular suffer from this cognitive dissonance. "In the end, it's the students who pay the price for low teacher salaries," according to one National Education Association article. But a different NEA article declares that we must "stop blaming teachers" for the problems in education, inadvertently making the case against higher salaries.

Because school factors are associated with such a small fraction of the variance in student achievement, pushing for a sweeping, root-and-branch replacement of the teacher-compensation system is not necessarily the best path to reform. The rewards that would stem from a brand-new system may not be worth the political and economic investment that would be required to achieve it. In fact, advocates of recruiting "the best and brightest" to teaching rarely consider the opportunity cost of doing so. Are the best and brightest really going to be more productive in K-12 classrooms than they would be in the research and development departments at Google, Exxon, or DuPont?

Despite the exaggerations of their importance, teachers do matter. Their effectiveness looms large within the small percentage of the variance in achievement that is associated with schools. Therefore, there are some less ambitious (but achievable) compensation reforms that are worth pursuing. First, compensation needs to be more differentiated to adjust for differences in teacher supply across subject areas. The left-leaning Center for American Progress approvingly reports that many non-religious private schools and new charter schools do not base salaries on a strict age-and-experience schedule, with roughly half adjusting salaries based on performance and one-third paying more to teachers in hard-to-staff subject areas. In the traditional public-school system, a reform as simple as signing bonuses for teachers of hard-to-staff subjects may help alleviate hiring problems.

In addition, districts should end the practice of shortchanging experienced teachers who transfer into their systems. For example, Detroit — which complains of teacher shortages — offers only two years of salary credit to a transferring teacher regardless of the teacher's actual experience. A teacher with 15 years of experience moving into Detroit would be paid over $18,000 less than what he could earn when credited with his full work history. This policy discourages the mobility that is needed to address isolated teacher shortages.

Likewise, teacher pension systems discourage mobility, certainly across state lines and sometimes within states. For instance, a teacher who moved from the suburbs of St. Louis, Missouri, to the city would not merely encounter more difficult teaching positions but would also be forced to change pension systems, and the traditional "defined-benefit" pensions most teachers participate in can severely punish employees who fail to work a full career within the same plan. Portable retirement plans, whether a traditional pension variant called a cash-balance plan or a 401(k)-style defined-contribution plan, would not punish teachers who change jobs. Defined-contribution retirement plans also are substantially cheaper for school districts to run, making savings available to pay higher teacher salaries where needed.

There is evidence that certain schools have difficulty retaining teachers, in particular schools with a large number of low-income students. The school environment doubtless plays a role in teacher turnover. According to a 2019 Fordham Institute survey, 13% of teachers in high-poverty schools report being attacked by a student, versus only 4% in low-poverty schools. Thirty-two percent of teachers in high-poverty schools report physical fighting in their school on a daily or weekly basis, versus only 5% in low-poverty schools. Teachers in high-poverty schools simply experience very different working conditions from those teaching in more affluent areas.

As with hard-to-fill math and science positions, differential pay could help attract and retain teachers in schools serving low-income students. At the same time, such schools should work to improve classroom conditions. While violence is much greater in high-poverty schools, not every school experiences it, and districts should analyze best practices to maintain a safe and orderly environment for both teachers and students. Doing so will require a hard-headed approach to discipline, avoiding faddish notions that, for example, persistent trouble-makers must not be suspended. When discipline breaks down, students will suffer, and teachers will seek work elsewhere.

Due to the threat of politically motivated firings, public-sector job protections will probably always be stricter than those in the private sector. Nevertheless, firing poor-performing teachers has become difficult to the point of absurdity. New York City's $150 million annual budget for paying teachers who have been disqualified from teaching yet cannot be fired is only the most prominent example. Teachers justifiably ask to be treated as professionals, but we can think of no other professional occupation in which the risk of dismissal for poor performers is so low. Simply firing the worst of the worst — say, the bottom 5% of teachers — could move students into the classrooms of more effective teachers with minimal disruption. There is no perfect way to identify the bottom 5%, but that problem exists in every occupation. Lawyers cannot always choose their cases, nor doctors their patients, and yet managers in those occupations use their best judgment to promote performance.

Finally, districts should consider cutting back or at least stemming the growth of non-teaching staff. Since 1980, public schools have added 985,000 additional teachers, but also 461,000 instructional aides, 76,000 principals and assistant principals, 47,000 guidance counselors, and 669,000 miscellaneous support staff. The U.S. has twice as many non-teaching staff per teacher as the average OECD country, and the educational value of these hires is suspect. The savings could be redirected toward effective teachers in hard-to-staff subjects.

Even the number of actual teachers may be excessive. Had the pupil-to-teacher ratio stayed at 1970's 22.3-to-1 rather than declining to the 16.0-to-1 ratio reported by the National Center for Education Statistics for 2019, real teacher salaries could be 40% higher at the current level of spending. Instead, America's schools invested in smaller class sizes, a strategy that is popular with parents and teachers but has dubious educational merit. Indeed, research from the OECD finds that smaller class sizes bear little relationship to better educational outcomes.

The preceding suggestions, even if fully implemented, will not be transformative. They will not remake our schools, or cure inequality, or vault the U.S. into first place in the international test rankings. They can, however, improve the efficiency of an educational system in which teachers have a limited but important role. Once we move past the distraction of "underpaid" teachers, we can focus on moderate reforms that minimize teacher shortages, improve student outcomes at the margins, and target spending more efficiently.

Andrew G. Biggs is a resident scholar at the American Enterprise Institute.

Jason Richwine is a public-policy analyst in Washington, D.C.


from the


A weekly newsletter with free essays from past issues of National Affairs and The Public Interest that shed light on the week's pressing issues.


to your National Affairs subscriber account.

Already a subscriber? Activate your account.


Unlimited access to intelligent essays on the nation’s affairs.

Subscribe to National Affairs.