Teacher shortages and turnover hit under-resourced districts hardest

The claim

The United States faces a persistent teacher shortage, but that shortage is not a uniform national phenomenon — it is a concentrated structural failure in specific types of schools. High-poverty urban districts, rural communities, and schools serving predominantly low-income students of color consistently report vacancy rates, out-of-field teaching assignments, and annual turnover rates that are two to four times those observed in affluent suburban districts. Because teacher quality is among the most powerful school-based determinants of student learning, and because these shortages translate directly into instructional instability and lower-qualified instruction, the unequal distribution of teacher supply compounds rather than merely reflects the unequal distribution of school funding. Structural analysts argue this is not a coincidence of individual teacher preferences but the predictable output of a teacher labor market in which compensation, working conditions, and professional status are locally determined and inversely correlated with student need.

The mechanism

The teacher labor market in the United States operates through a set of structural mechanisms that predictably direct teachers away from the highest-need schools.

First, teacher salaries are locally determined and track district wealth. Because school funding flows primarily from local property taxes, wealthier districts can offer substantially higher starting and experienced-teacher salaries without raising mill rates. A teacher choosing between a position in an affluent suburban district and a position in a neighboring high-poverty urban or rural district confronts not only an income differential that compounds over a career but also differences in working conditions — class sizes, administrative support, physical plant quality, student behavioral and mental health needs, and professional development resources. The rational response to this structure is to accept positions in more favorable districts when possible, which is what the data show.

Second, the entry-level composition of the teacher pipeline concentrates new and uncredentialed teachers in shortage schools. When high-poverty districts cannot fill vacancies competitively, they hire emergency-permitted teachers, out-of-field teachers, and long-term substitutes at higher rates. This means students in the highest-need schools are more likely to receive instruction from teachers who are themselves still learning the craft — the inverse of the distribution that would be optimal for closing achievement gaps.

Third, turnover in high-poverty schools is not simply a function of teacher dissatisfaction; it is structurally reproduced by the conditions that generate it. High turnover disrupts school culture, prevents teams from developing shared instructional practices, and imposes recurring administrative costs that further constrain the resources available for instruction. The turnover itself becomes a working condition that accelerates the next round of turnover — a self-reinforcing cycle.

The evidence

NCES and NTPS vacancy data by school poverty level

The National Teacher and Principal Survey (NTPS), conducted by NCES, is the most comprehensive nationally representative source on teacher labor markets by school characteristics. The 2017–18 NTPS found that high-poverty schools — those with 75% or more of students eligible for free or reduced-price lunch — were approximately three times more likely to report at least one teaching vacancy that was difficult to fill than low-poverty schools (fewer than 25% FRPL-eligible). Rural schools showed elevated vacancy rates comparable to high-poverty urban schools, often in different subject shortages (science and mathematics versus special education and bilingual education in urban contexts). State-level data from California, Texas, and New York corroborate the national picture: teaching vacancies are geographically concentrated in the poorest districts and in rural regions where salary competition from adjacent labor markets is weakest.

The experience and qualification distribution: Goldhaber, Lavery, and Theobald (2015)

Dan Goldhaber, Lesley Lavery, and Roddy Theobald’s analysis of administrative data linking teachers to students in Washington State provides the most methodologically rigorous documentation of how teacher experience and qualifications are distributed across schools by poverty level. Using teacher–student linked records, they found that students in the highest-poverty quintile of schools were three to four times more likely to be taught by a first-year teacher and substantially more likely to be taught by an out-of-field teacher (one teaching outside their primary certification area) than students in the lowest-poverty quintile. These are not marginal differences. The effect of being taught by a first-year versus a third-year teacher on student learning is roughly comparable in magnitude to the effect of a $2,000–$3,000 per-pupil spending increase — meaning the maldistribution of teacher experience effectively imposes an additional resource deficit on top of the documented funding gap. Goldhaber et al.’s findings have been replicated in studies using administrative data from multiple states, including North Carolina (Clotfelter, Ladd, and Vigdor, 2005) and New York.

Turnover rates and the Ingersoll longitudinal evidence

Richard Ingersoll’s longitudinal research using the Schools and Staffing Survey (SASS) and its companion Teacher Follow-up Survey (TFS) documents teacher turnover rates across decades and school types. His 2018 report with Merrill, Stuckey, and Collins, covering the period from the mid-1980s through the early 2010s, shows that annual teacher attrition from high-poverty schools has consistently run approximately 20% — roughly twice the rate in low-poverty schools. This means that the average high-poverty school replaces one in five of its teachers every year. The instructional cost is not only in the quality of replacements but in the destruction of institutional knowledge, mentoring relationships, and team coherence that accumulate over years in stable faculties. Ingersoll’s analysis also distinguishes migration from attrition: many teachers leaving high-poverty schools do not leave teaching — they transfer to lower-poverty schools within the same district or to wealthier neighboring districts. This transfer pattern is directly responsive to salary differentials and working conditions, confirming a structural rather than purely individual explanation.

Salary gaps and their spatial concentration

Allegretto and Mishel’s Economic Policy Institute analyses of teacher compensation document that within states — controlling for regional cost of living — starting teacher salaries in high-poverty districts average $11,000–$17,000 less than in the highest-spending suburban districts. Over a 30-year career, this salary gap compounds to several hundred thousand dollars in foregone income and reduced pension accumulation. Teachers with outside options — those with strong subject-matter credentials in STEM fields or with the credentials to teach in nearby affluent districts — respond to this differential, widening the qualification distribution between district types. The salary gap also affects the composition of teacher preparation program enrollees: lower lifetime earnings expectations reduce recruitment from higher-ability candidates with competitive outside labor market options, a dynamic Darling-Hammond and others have documented through admissions and employment data comparisons with high-performing countries.

The Teach For America model and the permanent shortage critique

Teach For America (TFA) and similar alternative pipeline programs were designed to address high-poverty school shortages by recruiting high-achieving college graduates for two-year placements. Research on TFA teacher effectiveness finds mixed results — some studies find small positive effects in mathematics for secondary students; others find neutral or negative effects compared to certified career teachers. The structural critique of TFA is distinct from the effectiveness question: a model built around two-year placements structurally accepts and perpetuates high turnover as the norm in the schools it serves. Linda Darling-Hammond and others have argued that TFA and short-term alternative certifications treat shortage as a fixed condition to be managed rather than a structural failure to be corrected. The business model of alternative certification vendors depends on continued shortage; their political advocacy, accordingly, has consistently opposed the salary increases and working condition improvements that would most directly address the structural drivers. Some high-poverty districts have become dependent on TFA cohorts as a de facto staffing strategy, with TFA functionally serving as a staffing agency for roles that should be filled by career teachers with permanent community ties.

Cross-national comparison: teaching as a selective profession

The teacher shortage pattern observed in high-poverty US schools is not universal among high-income nations, and the differences trace directly to structural features of teacher labor markets. Singapore recruits teachers from the top one-third of each graduating university cohort, pays salaries competitive with engineering and law, manages workforce supply through the National Institute of Education, and assigns teachers based on national workforce planning rather than local labor market competition. The result is that no Singaporean school faces a structural shortage of qualified teachers, and the distribution of teacher quality across schools by student socioeconomic status is far more compressed than in the United States. Finland recruits from the top 10% of university graduates into selective teacher preparation programs, with fully subsidized training and salaries that reflect professional status. Finnish principals report rejection rates of more than 90% for teacher preparation program applicants. South Korea and Germany operate similar nationally managed pipelines with competitive compensation and status that do not depend on local property tax bases.

The contrast with the United States is structural, not cultural. In these systems, the salary and prestige of teaching are set at the national level and are not sensitive to local property values. The consequence is that teacher quality is not inversely distributed by school poverty level — the distribution problem the United States has is largely an artifact of decentralized, property-tax-dependent compensation.

Who benefits

Charter management organizations (CMOs) with high teacher turnover models — including some of the largest national networks — benefit from a labor market in which replacement teachers are abundant, unions are weak or absent, and short-tenure teachers cost less in salary scale advancement and benefits than career teachers would. High turnover suppresses average compensation costs and, in non-union schools, reduces bargaining leverage. Several major CMOs have been documented by Ingersoll and others to have teacher attrition rates exceeding 30% per year — substantially higher than the already-elevated rates in traditional public schools serving similar populations.

Teach For America, Inc., and the network of alternative certification providers — including TNTP (The New Teacher Project), Relay Graduate School of Education, and state-level equivalents — have direct financial interests in continued structural shortage. Their revenue depends on school districts purchasing placement and training services that are unnecessary in adequately staffed systems. These organizations have lobbied against collective bargaining protections, advocated for tenure reform that reduces job security, and opposed salary schedule restructuring that would compete for the same college-graduate labor pool they recruit from.

Employers and donor networks associated with the Walton Family Foundation, Arnold Ventures, and aligned education reform philanthropy have funded the alternative certification and CMO ecosystem on ideological grounds that favor market-oriented school reform. These funders have also supported policy campaigns opposing teacher pay increases tied to experience and advanced degrees — the very features of compensation structures that reduce turnover in the long run.

State governments that fund teacher preparation programs inadequately and then purchase short-term alternative certification as a cheaper substitute effectively externalize the cost of chronic shortage onto the students in affected schools.

The counter

The most substantive counterargument is that teacher distribution is partly the product of individual teacher preference, not just salary and working conditions. Survey research on teacher preferences consistently shows that some teachers actively seek high-need urban and rural placements for intrinsic reasons — mission alignment, community connection, and the professional satisfaction of working with underserved students. A non-trivial share of the workforce demonstrates that teachers can and do choose high-poverty placements, and programs that train and support teachers specifically for these contexts (the Urban Teacher Residency model, for example) show better retention than standard preparation pathways. This suggests the problem is not insoluble through labor market reform alone — recruitment and preparation culture matter.

There is also a genuine empirical debate about what drives turnover: compensation differentials, working conditions (discipline, administrative support, class size), or the characteristics of the student population itself. Hanushek, Kain, and Rivkin (2004) found, using Texas administrative data, that a large share of teacher mobility was attributable to student characteristics including behavioral challenges rather than salaries alone. This does not exonerate the structural factors — difficult working conditions are themselves partly a function of under-resourcing, inadequate mental health support for students, and large class sizes — but it cautions against salary-only interventions as a complete solution.

The international comparisons, while compelling, carry a caveat: Singapore and Finland operate in substantially different institutional and cultural contexts. Singapore’s centralized workforce management depends on a size and institutional capacity that is difficult to replicate across fifty US state systems. Finland’s low poverty rates (child poverty is consistently under 4%, versus the US rate of approximately 14%) mean that the distribution of student need across schools is less acute, reducing the structural pressure that generates differential vacancies. These comparators illuminate structural possibilities rather than directly transferable policy blueprints.

The structural argument does not require dismissing these counterpoints. It requires showing that after controlling for individual preferences, school type, and regional labor market conditions, the poverty level of the school remains an independent predictor of vacancy rates, turnover, and teacher qualification levels. The Goldhaber et al. administrative data and the Ingersoll longitudinal surveys both show this, and the policy responsiveness evidence — salary increases reducing vacancies and turnover — confirms that the structural factors are operative and addressable, not merely correlated artifacts.

References

Goldhaber, D., Lavery, L., & Theobald, R. (2015). Uneven playing field? Assessing the teacher quality gap between advantaged and disadvantaged students. Educational Researcher, 44(5), 293–307. https://doi.org/10.3102/0013189X15580116

Ingersoll, R., Merrill, E., Stuckey, D., & Collins, G. (2018). Seven trends: The transformation of the teaching force (CPRE Research Report #RR-2018-2). Consortium for Policy Research in Education, University of Pennsylvania. https://repository.upenn.edu/cpre_researchreports/91/

Clotfelter, C. T., Ladd, H. F., & Vigdor, J. (2005). Who teaches whom? Race and the distribution of novice teachers. Economics of Education Review, 24(4), 377–392. https://doi.org/10.1016/j.econedurev.2004.06.008

Darling-Hammond, L. (2017). Teacher education around the world: What can we learn from international practice? European Journal of Teacher Education, 40(3), 291–309. https://doi.org/10.1080/02619768.2017.1315399

Hanushek, E. A., Kain, J. F., & Rivkin, S. G. (2004). Why public schools lose teachers. Journal of Human Resources, 39(2), 326–354. https://doi.org/10.2307/3559017

Allegretto, S., & Mishel, L. (2020). Teacher pay penalty dips but persists in 2019. Economic Policy Institute. https://www.epi.org/publication/teacher-pay-penalty-dips-but-persists-in-2019/

National Center for Education Statistics. (2020). National Teacher and Principal Survey, public school teacher and principal data file, 2017–18 (NCES 2020-142). U.S. Department of Education. https://nces.ed.gov/surveys/ntps/

Sutcher, L., Darling-Hammond, L., & Carver-Thomas, D. (2016). A coming crisis in teaching? Teacher supply, demand, and shortages in the U.S. Learning Policy Institute. https://learningpolicyinstitute.org/product/coming-crisis-teaching

Ronfeldt, M., Loeb, S., & Wyckoff, J. (2013). How teacher turnover harms student achievement. American Educational Research Journal, 50(1), 4–36. https://doi.org/10.3102/0002831212463813

OECD. (2023). Education at a glance 2023: OECD indicators. OECD Publishing. https://doi.org/10.1787/e13bef63-en