Texas study suggests special ed formulas miss much of district spending

A new EdWorkingPaper from University of Texas at Austin researchers lands at an awkwardly useful moment for Texas school systems: just as the state begins replacing its old placement-based special education funding approach, the paper suggests that the kinds of proxies finance formulas often rely on — especially educational placement and disability mix — do matter, but still explain only a modest share of what districts actually spend. In other words, the problem may not be that formulas pick the wrong expensive category. It may be that the categories themselves are too blunt for real-world budgeting. (edworkingpapers.com)

That matters beyond one state. Texas Education Agency guidance says the new funding system takes effect on September 1, 2026, with districts moving from an instructional-arrangement model toward funding based on service intensity and service groups. But TEA has also warned districts that for the 2026-27 school year they must report students under both the old and new systems, and that final funding levels under the new approach will not be clear until the end of the school year. For chief financial officers, special education directors, and superintendents, that leaves a familiar problem in place: they still have to build budgets now, even when the formula may not line up neatly with actual staffing and service delivery. (tea.texas.gov)

What the paper actually found

In the study, David E. DeMatthews, Jinseok Shin, and Pedro Reyes analyzed 6,116 Texas district-year observations from 2019-20 through 2024-25. Their descriptive findings are straightforward and important: special education spending varied substantially across districts, the variation persisted over time, and it was especially large among smaller districts. Their regression analyses also found that districts serving larger shares of students in highly self-contained placements, and districts with higher concentrations of students with autism, spent more per student receiving special education services. (edworkingpapers.com)

But the headline result is the limit, not the relationship. Placement, disability composition, and other observable district characteristics explained only a modest share of overall spending variation, according to the paper. That gives some support to the basic logic behind placement- and disability-based formulas: those categories are not meaningless. It also undercuts the idea that a state can understand district need simply by counting students in a few administratively convenient buckets. (edworkingpapers.com)

Readers should treat the findings as credible but still preliminary. EdWorkingPapers says its papers are circulated for comment and discussion and have not gone through a rigorous peer-review process. And this study is district-level and observational: it does not show that a particular placement or disability category causes spending to rise, and it does not tell readers whether higher spending in any district reflects stronger services, higher costs, inefficient delivery, or some combination of those factors.

Why formulas miss what districts feel

Texas is a particularly useful place to test the question because the state has been operating under a predominantly placement-based system — the very kind of formula the paper is scrutinizing. TEA says the replacement model is designed to shift attention away from where services are delivered and toward the intensity of specially designed instruction and related supports documented in a student’s IEP. Under the new framework, districts will classify students across five domains — curriculum and instruction, behavior, communication, independent functioning, and personal care/health — and may receive additional funding through service groups layered on top of intensity tiers. (tea.texas.gov)

The paper helps explain why Texas is making that move, while also suggesting that even a more refined formula may not solve the full problem. District budgets are shaped by forces that simple state proxies often do not capture well: local labor markets, therapist and contractor availability, transportation demands, cooperative arrangements, sparse enrollment, and the fixed cost of keeping specialized staff and programs available even when only a small number of students need them. The authors describe these as organizational, fiscal, and contextual factors outside traditional formulas. Separately, a 2026 GAO report found that district and school officials commonly cited resource availability, facilities, school culture, and parental involvement as factors shaping educational settings for students with disabilities, while a 2024 GAO report found personnel shortages were a key obstacle across the districts and organizations it reviewed. Taken together, those findings point to a classroom reality formulas struggle to see: districts do not purchase “placement categories”; they assemble people, time, expertise, and coverage. (edworkingpapers.com)

That has practical consequences for inclusion debates. Federal IDEA rules require educational placement decisions to be based on the student’s IEP and made in conformity with least-restrictive-environment requirements, not on a district’s reimbursement preferences. Yet if a state formula tracks some settings or categories better than it tracks the cost of supports delivered inside general education classrooms, district leaders can end up trying to explain a mismatch to school boards and families: the system may be asking them to be more inclusive while funding them through older, rougher signals of need. That is an inference from the study and the policy context, not a causal finding of the paper itself. (sites.ed.gov)

What districts should not overread

The study is not evidence that students are misidentified, that autism is being used as a budget category, or that more restrictive placements are inherently wasteful. It also does not settle how much special education should cost. What it does do is sharpen a harder finance question: whether states are serious about funding actual services, or whether they are still using administratively tidy stand-ins for those services and hoping the averages come out close enough. (edworkingpapers.com)

That question is becoming national. The Institute of Education Sciences says the best national estimates of special education spending are more than 20 years old, and its National Study of Special Education Spending is moving into full data collection in the 2026-27 school year; a federal notice published in August said collection activities will take place during the 2026-27 and 2027-28 school years. If Texas — a state actively rebuilding its formula right now — still shows large unexplained differences between spending and common funding proxies, the coming federal work may be most useful not for naming one perfect formula, but for revealing how much of special education finance still sits outside the categories states currently reward.

For district leaders, the immediate takeaway is less ideological than operational. If your board keeps asking why special education costs are rising, this paper suggests the answer may not be visible in disability labels or placements alone. Districts will need stronger internal cost accounting tied to actual service patterns — staffing, contracted providers, transportation, and program design — if they want to explain spending clearly or argue that state formulas are still missing the mark. As Texas implements its new intensity-based system this school year, the next development worth watching is whether better student-level service data narrows that gap, or merely measures more precisely how much formula funding still cannot see. (spedsupport.tea.texas.gov)