The question, made precise

"How many teacher contracts are tied to student outcomes?" is several different questions in a trench coat. The narrowest reading — "are pay or evaluation explicitly indexed to student-growth scores?" — produces one set of numbers. The broader reading — "do contract provisions shape student outcomes in measurable ways?" — produces a much bigger set, because contracts also dictate how much time teachers spend with students and which teachers end up in front of which students.

The honest framing, then, is that CBAs influence outcomes through five distinct channels:

1. Compensation tied to outcomes. Does the contract pay teachers more (bonuses, accelerated step increases, performance-based base salary) on the basis of evaluation results that themselves include student-growth measures?
2. Evaluation tied to outcomes. Does the contract (or the state law it sits inside) require that student-growth measures be a component of the teacher's annual performance evaluation, regardless of whether that evaluation moves dollars?
3. Job security tied to outcomes. Does student-growth performance affect tenure, layoff order, dismissal procedures, or the standard of cause — the rules that determine whether a teacher keeps the job?
4. Instructional time. What does the contract say about the length of the teacher work day, the school day, the school year, the planning period, and the protected non-instructional time built into the calendar? More time with students raises outcomes; how time is allocated between teaching and planning shapes effectiveness.
5. Job-security protections. Independent of any outcome metric: how does the contract decide which teachers get tenure, get laid off, get transferred, get assigned to which schools? Even when the rules don't reference test scores, they govern who stands in front of which students — which is itself an enormous lever on outcomes.

The empirical answer is very different at each layer. About half of large-district contracts have something resembling (1)³. Three out of every four states have something resembling (2)². Almost no contract anywhere ties (3) directly to outcome data — and where districts have tried, courts and state legislatures have walked it back²⁰. Contracts (4) and (5) influence outcomes substantially without ever measuring them; that influence is the easiest part of the question to overlook and the hardest to undo.

There is no comprehensive census of all 13,000 U.S. teacher CBAs. The best systematic data covers about 150 of the country's largest districts.NCTQ Teacher Contract Database, 2024

The data we have (and don't)

The most important caveat in this brief is the missing denominator. The United States has roughly 13,000 K–12 public-school districts¹³; the great majority operate under negotiated collective bargaining agreements where the state allows or requires bargaining. No agency systematically collects all of those contracts. The U.S. Department of Education does not. The U.S. Census of Governments tracks bargaining structure but not contract content. State labor boards house local CBAs but do not standardize fields across districts.

What does exist is the NCTQ Teacher Contract Database, maintained by the National Council on Teacher Quality. It systematically codes the contracts and school-board policies of 147 school districts and 2 charter management organizations: the 100 largest districts in the country, the largest district in each state, and the member districts of the Council of Great City Schools¹. Together these districts educate a substantial share of U.S. public-school students — but they are large, urban-skewed, and not a probability sample of the universe.

For state-level law, the same organization's State of the States series tracks teacher-evaluation policy in all 50 states plus DC². Together, the two databases are the operative empirical foundation for any honest answer to this question.

The empirical answer

Compensation tied to outcomes

Of the 147 districts in NCTQ's Teacher Contract Database¹, NCTQ reports that roughly 48% — about 59 districts — offer some form of performance pay tied to a teacher's evaluation rating³. That figure has trended upward modestly since 2018 (the earlier NCTQ count was 51 districts)³.

The structure matters as much as the headline rate. NCTQ's sub-finding³: only about 19 districts in the database (roughly 13%) tie bonuses directly to evaluation outcomes — and of those, only 8 make the bonus available to all teachers regardless of subject or school assignment. The remainder are limited to particular schools or subjects (typically hard-to-staff or shortage areas), where the bonus's purpose is closer to recruitment than to performance differentiation. The "performance pay" headline figure of 48% includes, for example, contracts that grant accelerated step movement to highly-rated teachers; it is a much weaker form of outcome-tied pay than a real merit bonus.

Evaluation tied to outcomes

The story at the evaluation-design layer is the higher-level political fight, and it has moved sharply over the last decade and a half. The arc:

Where student growth is still required as a component, the weight varies dramatically: Louisiana law sets it at 50% of the evaluation (35% via VAM and 15% via student-learning targets)³²; Michigan caps it at 20%³³; South Carolina uses VAM at district discretion; Florida has eliminated the state-mandated component but allows districts to keep one²⁷. Texas's value-added system began rolling into the Teacher Incentive Allotment framework only in 2025–26³¹.

Job security tied to outcomes

Almost no major CBA in the country directly conditions tenure, layoff order, or dismissal on student-growth scores in a way that has survived legal challenge. The two big experiments in this layer — Houston's use of value-added scores as the basis for teacher termination²⁰, and similar uses in New Mexico³⁰ — both ended in litigation that the districts effectively lost. (See Administrative & case law.)

Calendar & instructional timeAdded

The most direct way a teacher contract influences student outcomes is the simplest: by deciding how many hours per day, and how many days per year, students and teachers are in the same room. Almost every American teacher CBA addresses three things in this domain — the length of the teacher work day, the length of the school year, and the protected non-instructional time within each day (the planning period, the prep period, lunch, restroom breaks). All three are negotiated; all three vary widely; and all three affect outcomes through the cleanest causal pathway in K–12 research.

The teacher work day

NCTQ's Every Minute Counts: How Districts Govern Teacher Time⁵ — the canonical empirical analysis of work-day provisions — finds that most teachers in the 145-district sample are contracted to work between 7.5 and 8 hours per day⁵, broadly aligning with the 40-hour work week. But the variance underneath that average matters:

• Over 20% of districts in NCTQ's sample do not address the length of the work day at all in their CBA⁵ — meaning the operative norm is whatever the school principal sets, with no contractual protection in either direction.
• Six districts require significantly shorter work days of 6.5 hours or less⁵.
• Stockton Unified School District (CA), an outlier, defines its teacher work day by instructional time alone — just over 4 hours per day for middle school and 4 hours per day for K–6 teachers, with everything else outside the contractual obligation⁵.

Planning time

The planning period is the single most heavily-negotiated calendar provision in U.S. teacher contracts. Public school teachers nationally average about 53 minutes of planning time per day, slightly above the median across NCTQ's sample⁵.

Source: NCTQ, Every Minute Counts: How Districts Govern Teacher Time.

The 30:1 ratio between Manatee County's elementary planning time and San Bernardino's is contractual, not budgetary — both are large urban districts — and translates directly into instructional time the kids do or do not receive. A district that bargains four hours of weekly planning instead of one shifts roughly three hours of teacher attention per week from in-front-of-students into preparation, which the literature suggests is worth doing up to a point but past which the marginal return of additional planning falls below the marginal return of additional contact.

The school year

The state-mandated minimum number of school days is most commonly 180 days¹, but the contractual school year for teachers is usually longer than the student year by anywhere from a handful to two-plus weeks of professional-development and pre-service days. The NCTQ database tracks both¹. The bargained school year is rarely shorter than the state minimum, but it is also rarely much longer — districts that have attempted to extend the year have generally had to bargain compensation increases proportional to the additional days.

Where state laws allow, contracts also typically restrict the district's ability to add days mid-year (e.g., for missed snow days or pandemic make-up) without negotiating compensation. This is why the COVID-era hybrid and remote schedules required reopening of contracts in many districts.

Why this matters for outcomes

About 25 percent of the variation in test outcomes between countries in reading and math could be attributed to how much time students spend learning.

— OECD, 2011, cited in NCTQ Every Minute Counts²⁸

The instructional-time channel is the strongest causal pathway between contract content and student outcomes that the research literature has identified. Cross-country variation in PISA scores tracks instructional time more reliably than any other observable system feature. Within the U.S., extended-day and extended-year experiments (charter schools, KIPP-style models, summer programs) have generally produced positive achievement effects, with the size of the effect roughly proportional to the time added.

This is why the planning-period and work-day provisions in a CBA are an outcome-relevant question even though no contract anywhere indexes them to test scores. The contractual decision to allocate 53 minutes per day to planning instead of 30 or 90 is, in expectation, an outcome decision — just one made through bargaining rather than measurement.

Teacher protections that shape who teaches whomAdded

The largest and least-discussed influence of CBAs on outcomes runs through what economists call "the workforce-allocation problem": even setting aside compensation and evaluation, the contract decides which teacher stands in front of which student. Five contractual provisions do most of that work.

Tenure procedures

The tenure decision — the moment at which a probationary teacher acquires "just cause" protection against dismissal — is the single most consequential personnel decision in U.S. K–12 employment. Contracts and state laws set the timing and the standard:

• 28 states set a 3-year probationary period before tenure⁶.
• 4 states (California, Mississippi, South Carolina, Vermont) have only a 2-year probation⁶.
• 19 states require 4 or more years⁶.
• 4 states (Florida, Kansas, North Carolina, Wisconsin) have functionally eliminated tenure entirely — teachers serve under recurring annual or multi-year contracts⁶.
• The District of Columbia and North Dakota have no statewide tenure policy; each district sets its own rules⁶.

A more demanding test: of the 45 states with tenure policies, only 22 have any explicit performance requirement a teacher must meet to earn tenure or non-probationary status⁶. In the other 23 states, tenure is effectively automatic at the end of the probationary period absent affirmative dismissal action by the district. Thirteen states allow the probationary period to be extended by 6 months to 2 years if a principal is uncertain — a softer alternative to denial⁶.

Tenure protections do not reference student outcomes anywhere on the front of the contract. But by setting a high procedural bar for dismissing veteran teachers, they govern which teachers stay in the classroom long-term and which leave during their early-career years.

Layoff order: the LIFO question

When a district reduces force, who goes first? In most American CBAs, the answer is last-in, first-out (LIFO) — reverse seniority. The data are stark, and they break by who is required to bargain layoff procedures:

Source: NCTQ Teacher Contract Database analysis of layoff procedures.

The most cited empirical work on LIFO is by labor economist Dan Goldhaber and colleagues, who have shown that:

Using a LIFO-layoff process rather than a performance-based system requires districts to lay off substantially more, and higher-quality, teachers — on the order of 20 to 26 percent of a standard deviation in student achievement — to reach equivalent budget-savings targets.

— Goldhaber et al., synthesis of LIFO research¹⁶

The mechanism is straightforward: junior teachers are paid less than senior ones, so cutting by seniority sheds the cheapest people on the roster — meaning more bodies must be cut to reach the same dollar target. Effectiveness research has found that, on average, junior teachers are not less effective than senior teachers in ways that justify cutting them first, particularly because the early-career years are when teacher quality variance is highest. The result is that LIFO predictably falls hardest on schools serving low-income and minority students — because high-poverty schools tend to have the youngest teachers.

Mutual consent vs. forced placement

When a teacher is excessed from one school (the position is eliminated, the program is closed, the teacher's certification no longer matches the assignment), where does that teacher go? Two contractual regimes answer differently:

• Forced placement (the older default). The central office assigns the excessed teacher to a vacancy at another school. The receiving principal has no veto. The receiving school may not have wanted that teacher; the teacher may not be a good fit for that school's program. Both sides absorb the mismatch.
• Mutual consent. A transfer requires the agreement of both the teacher and the receiving principal. If no principal wants to hire the teacher within a defined window, the teacher does not get placed — and the district must either find another role or stop paying.

Boston, New York, Chicago, and Newark have all moved partially or fully to mutual consent in the last 15 years⁸. The Boston experience is the cleanest natural experiment: Matthew Kraft and colleagues studied the BPS hiring reforms across 84 traditional schools and found:

The reforms moved up the median hire date by more than two months, reduced late hiring, increased new-hire diversity and retention, and improved student achievement by 0.09σ in math and 0.07σ in ELA — effect sizes that are larger than the average across 72 school-based experimental interventions and larger than 70% of effect sizes from 181 large-scale randomized control trials of education interventions.

— Kraft et al., on Boston Public Schools mutual-consent hiring reform¹⁷

Schools also hired approximately 110 more Black and Hispanic teachers in the four years after the reforms than they would have under the prior forced-placement regime¹⁷ — a finding that has shifted the equity politics of mutual consent in particular.

The "rubber room" / Absent Teacher Reserve

When a district adopts mutual consent, but the teacher's CBA still guarantees full salary even when no principal will hire them, the financial pressure has to go somewhere. The most prominent example is New York City's Absent Teacher Reserve (ATR) — teachers who have been excessed but remain on payroll, often rotating between schools as long-term substitutes:

Sources: Citizens Budget Commission; NYC Comptroller; Education Trust–New York.

And the equity wrinkle: ATR teachers permanently placed into classrooms have been disproportionately assigned to schools serving Black students, low-income students, and schools with significant academic needs.¹⁵ The ATR is the visible artifact of a contract that locks in salary obligation without locking in placement consent.

Dismissal procedures

The fifth layer is the standard for dismissing tenured teachers for cause. CBAs and state law together specify what counts as cause, what process the district must follow, what evidence is admissible, and what timeline must elapse before a final dismissal decision. The procedures themselves are not outcome-related, but their stringency affects how often districts actually dismiss low-performing teachers — which means it affects which teachers remain in classrooms long-term.

The empirical record is that formal dismissal of tenured teachers is rare in absolute terms in most districts. Most underperformance is addressed through informal counseling-out (the teacher leaves voluntarily after the principal makes the message clear), unsatisfactory-performance-rating processes, or buyouts. The contract architecture shapes which of those paths is available and how expensive each one is.

The state-level picture: a regulated retreat

The story between the 2015 peak and the 2022 trough is, mostly, ESSA. The Every Student Succeeds Act (signed December 10, 2015)⁹ reauthorized federal K–12 education law and removed the federal mandate that states tie teacher evaluations to student-achievement data — a condition that had been baked into the NCLB waivers granted by Secretary Arne Duncan's Department of Education and that had reached as far as the Race to the Top funding formula in 2009–10¹⁰.

What followed was not, in most places, an outright repeal. It was a quieter unwinding:

• Florida (HB 7069, 2017) eliminated the requirement to incorporate state-assessment data in evaluations, leaving the choice to districts²⁷.
• Michigan (SB 122) delayed a planned increase in test-score weight, holding it at 25%²⁶.
• Kentucky (2017) reversed a prior rule that had required student growth as a "significant factor" in evaluations²⁶.
• Arkansas, Alaska, Connecticut, North Carolina passed similar policies through state boards of education or state legislation²⁶.
• COVID-era pauses (2020–2021) caused at least seven additional states — Arizona, Delaware, Indiana, North Dakota, Oregon, South Carolina, West Virginia — to drop state assessments as a required or explicit measure of student growth².

The pattern is rarely "we no longer believe in measuring effectiveness." It is more often: lower the weight, broaden what counts as "growth" (so it is not solely standardized-test value-added), or leave the design choice to local districts. The result is a much more permissive regulatory floor than existed in 2015 — and a corresponding decline in the share of CBAs that need to make student-outcome ties pencil out under state law.

The federal arc: Race to the Top to ESSA

The reason the 2015 numbers were so high in the first place was federal money and federal pressure, not state preference. The arc:

Four districts, four endings

Most of what we know about how outcome-tied teacher contracts perform in practice comes from a small number of high-profile experiments, each of which arrived at a different equilibrium.

Washington, DC — IMPACT and IMPACTplus (2009–present)

The country's most ambitious surviving outcome-tied evaluation system is IMPACT, launched under DC Public Schools Chancellor Michelle Rhee in 2009²². The companion compensation overlay, IMPACTplus, pays a Highly Effective rating bonus and accelerates base-salary movement after multiple consecutive Highly Effective ratings. DCPS publishes the bonus structure: up to $25,000 annually, and as much as $3.7 million in additional career earnings for sustained top performance²².

IMPACT survived the change in administrations (it predates the federal RTTT push) and has remained operative through 2024–25 and 2025–26 with continuous refinements. The 2024–25 IMPACT Guidebook is published openly²². DCPS commissioned a multi-year review of the system starting in spring 2020 that produced calibration changes but did not retract the basic outcome-tied design. DC is the closest thing the country has to a stable, large-scale, outcome-tied teacher contract.

Denver — ProComp (1999–present, defanged)

Denver Public Schools launched the country's longest-running pay-for-performance pilot in 1999 and replaced it with the Professional Compensation System for Teachers (ProComp) in 2005. ProComp eliminated the single salary schedule and substituted a base salary plus four performance components — teacher knowledge and skills, professional evaluation, market incentives, and student growth²³.

The empirical record is not encouraging for the model's design ambitions. A Harvard Strategic Data Project study found that "ProComp does not establish a differentiated compensation system as it originally set out to do. Salaries under ProComp largely mirror salaries from traditional compensation systems." The average teacher earned $456 more under ProComp than they would have on the traditional schedule — effectively a small across-the-board bump, not differentiation by performance²³.

The Denver Classroom Teachers Association has been pushing in recent contract cycles to "create a more traditional salary schedule that enables stability across the district"²³ — an explicit retreat from the ProComp design even as the formal architecture remains in place.

Houston — ASPIRE and the VAM lawsuit (2007–2017, ended)

Houston Independent School District ran the country's most aggressive outcome-tied teacher discipline regime through its ASPIRE program: teacher value-added scores (calculated by a proprietary third-party algorithm called EVAAS) were used to inform termination decisions. The Houston Federation of Teachers sued in 2014. In May 2017, the Southern District of Texas ruled that plaintiff-teachers had stated a viable Fourteenth Amendment due-process claim — principally because the third-party VAM algorithm was a black box, the district could not produce the calculations on demand, and therefore teachers had no way to verify the basis on which they were being fired²⁰.

HISD settled later that year, paid $237,000 in plaintiffs' fees, and agreed to stop using value-added scores as a basis for termination²⁰. The settlement is the most consequential procedural-due-process precedent in U.S. teacher-evaluation litigation: not a holding that VAM is invalid, but a holding that opaque third-party VAM cannot be the basis for a terminatable evaluation. (See Administrative & case law for the legal architecture.)

Long Island, NY — Lederman v. King (2016)

One single-teacher case has done more than any other to shape the state-level retreat: Lederman v. King. Sheri Lederman, a veteran fourth-grade teacher in Great Neck on Long Island whose students consistently scored well above state standards, received a state-mandated growth-component score of 1 out of 20 in 2013–14 — "ineffective." Her score the previous year had been 14 of 20 ("effective"); the year after, with no curricular change, it returned to "effective."²¹

Lederman challenged the rating. In 2016 a New York Supreme Court judge held that her growth rating was "indisputably arbitrary and capricious" — specifically because the state's "bell curve" model required predetermined percentages of teachers in each rating category every year, regardless of underlying performance, and because the ceiling effect on Lederman's already-high-performing students made measurable improvement statistically improbable²¹. New York's APPR system has been substantially overhauled since.

Hearing transcripts & court findingsAdded

The cleanest verbatim record of how outcome-tied teacher contracts have been litigated and debated lives in three places: federal district-court opinions (where high-stakes VAM has been challenged on procedural-due-process grounds); state-court growth-score rulings; and state legislative floor records (where the bargaining-rights and tenure-elimination fights actually happened). Where transcripts are publicly available, they tell the story more vividly than any summary.

What follows is verbatim from the underlying records. Click any header to expand.

Administrative and case law

The legal architecture under teacher-evaluation contracts has three layers, each of which has been actively litigated.

Federal level

The Every Student Succeeds Act (Pub. L. 114–95, signed December 10, 2015) replaced the No Child Left Behind framework and explicitly returned authority over teacher-evaluation design to states. ESSA retained federal reporting requirements on teacher quality and equitable access but dropped the conditionality that had driven the 2009–2015 expansion. Federal regulations implementing ESSA's teacher provisions sit in 34 CFR Part 200, but the operative reality is that the federal floor on student-outcome-tied evaluation has been removed.

State administrative level

Where student-growth measures remain in evaluation systems, they live in state administrative codes promulgated by state education departments. Examples:

• Louisiana: Bulletin 130 of the Louisiana Administrative Code governs teacher evaluation; sets the 50%-professional-practice / 50%-student-growth division.
• Michigan: MCL 380.1249 (the state's educator-evaluation statute), implementing rules at the Michigan Department of Education.
• Texas: 19 Texas Administrative Code Chapter 150 (Teacher Appraisal); Teacher Incentive Allotment statutory authority at Texas Education Code § 21.5512; Texas Value-Added Model integration begins 2025–26.
• Florida: Section 1012.34, Florida Statutes (post-HB 7069), which now leaves student-growth methodology to local districts.

State CBAs operate inside this administrative floor. A district cannot bargain a contract that violates state-mandated evaluation rules, but where the state floor is permissive (Florida, post-HB 7069) the CBA is the operative source of law. The result: outcome-tied evaluation provisions are now far more often contractual than they are statutory.

Case law

The case-law pattern is narrow but consequential: courts have not held VAM-based evaluation to be substantively invalid. They have held that high-stakes uses (especially termination) require procedural due process — meaning the algorithm's outputs must be auditable, the data underlying them must be producible, and the standard of review must be intelligible to the affected teacher. Most school districts have read these cases as a deterrent against high-stakes outcome-tied job-security provisions, regardless of whether their own state mandates VAM.

Where the unions stand

The two national teachers' unions have meaningfully different postures, and the difference is reflected in local contract patterns.

NEA — opposed

The National Education Association (President Becky Pringle since 2020) has consistently opposed merit pay schemes and at its 2000 representative assembly amended its long-standing opposition to add a blanket rejection of "any other system of compensation based on an evaluation of an education employee's performance."²⁴ NEA officially permits bonuses for: (a) National Board for Professional Teaching Standards certification, (b) additional responsibilities like mentoring, and (c) hard-to-staff school assignments. It does not permit (in its policy resolutions) bonuses tied to evaluation ratings or student-growth scores²⁴.

AFT — permissive

The American Federation of Teachers (President Randi Weingarten since 2008) takes a meaningfully more open position. AFT's policy framework explicitly recognizes that the single salary schedule has "severe drawbacks" and "does not allow teachers to be compensated like other professionals in our society."²⁵ AFT urges its locals to explore alternative compensation, including "financial incentives to teachers who acquire additional knowledge and skills."²⁵ That difference shows up in negotiated contracts: AFT-affiliate locals have been more willing to bargain performance-pay structures than NEA-affiliate locals, particularly in the post-2009 wave of city-district reforms.

Nonetheless, AFT was the lead union plaintiff in Houston Federation of Teachers v. HISD²⁰. The union's distinction is not "outcome-tied compensation is bad" but "outcome-tied termination with a black-box algorithm is unconstitutional."

Union analysis: the money behind the positionsAdded

The policy positions described in the previous section — NEA's blanket opposition to outcome-tied compensation; AFT's more permissive but still litigation-leading posture — do not exist in a vacuum. They are downstream of a specific financial architecture that has been under sustained pressure since 2018. Understanding the financial machinery and the membership trends makes the policy positions easier to read.

The financial machinery

The two national teachers' unions are large, dues-funded organizations. Their headline financials for the most recent reporting periods⁴⁶:

Sources: NEA Strategic Framework & Budget; AFT LM-2 filings with U.S. Department of Labor; The 74 / Mike Antonucci analyses.

Per-member, the national portion of NEA dues for an active classroom teacher is $208 per year for the 2023–24 school year — before state and local affiliate dues, which typically multiply the total to roughly $700–$1,000 annually depending on the affiliate⁴⁶. These dues are most commonly collected via payroll deduction (“dues check-off”), which is itself a CBA-negotiated provision. The mechanism matters: payroll deduction makes the dues invisible to most members month-to-month, which historically suppressed turnover.

The Janus rupture

On June 27, 2018, the U.S. Supreme Court decided Janus v. AFSCME Council 31, 5–4, holding that public-sector unions could no longer collect “agency fees” from non-member employees they bargained on behalf of. Before Janus, in 22 states a teacher who declined to join the union still had to pay a reduced fee covering the union's bargaining costs. After Janus, those fees became unconstitutional — effectively making the entire United States a “right-to-work” jurisdiction for public employees⁴⁸.

The financial consequences have been substantial but not catastrophic. Across 66 public-sector unions in 22 non-RTW states, membership fell from 3,360,000 pre-Janus to 2,980,000 by 2022 — a 11.3% loss. The combined NEA + AFT + AFSCME total fell from 5,292,481 in 2018 to 4,743,397 by the most recent reporting — a 10.4% decline, or roughly 549,000 lost members. Aggregate annual union revenue is down an estimated $733 million, based on roughly 1.2 million public-sector workers who have resigned or declined membership since the ruling⁴⁸.

The state-level numbers are more dramatic in places that paired Janus with prior or simultaneous state-law changes:

The pattern: Janus by itself produces gradual attrition. Janus + a state-law change that eliminates dues check-off, scope of bargaining, or contract scope produces a cliff.

How the unions are responding financially

Three observable responses to the dues-revenue squeeze:

1. Higher per-member dues. Despite membership falling, NEA's total dues revenue keeps rising — from a 2023–24 collection of $379 million to FY24's $381 million⁴⁶. The math requires that the per-member rate has been raised to outpace the membership decline. The 74 has tracked this in repeat coverage: “NEA membership continued to drop in 2024 as revenue from dues hit $381 million.”
2. Member-retention investment. Bloomberg Law reported in 2020 that public-sector unions had “fend[ed] off membership exodus in 2 years since Janus ruling” — partly through aggressive recommitment campaigns and partly through legal challenges to opt-out procedures. NEA and AFT both spent significantly on internal organizing and member-services upgrades during 2018–2022⁴⁸.
3. Continued political-spending ramp. The NEA Advocacy Fund raised $27.9 million for the 2023–24 election cycle, including $2.5 million to Future Forward USA Action (pro-Harris super PAC) and $1.5 million each to the House Majority PAC and the Senate Majority PAC⁵⁰. The political ramp has not slowed proportionally to membership.

Why this matters for the CBA-and-outcomes question

The financial-incentive lens reframes several of the positions documented earlier in this brief:

NEA's blanket opposition to outcome-tied compensation

Outcome-tied pay creates within-member inequality: some teachers earn more than others under the same contract. That inequality is solvent for the employer but corrosive for collective member identification with the union — the “we're all in this together” framing that supports voluntary dues payment under Janus. An NEA local that bargains a merit-pay system is harder to organize than one that bargains a flat salary schedule. The financial incentive runs strongly against differentiated pay regardless of how the empirical research on its effectiveness turns out²⁴.

AFT's more permissive posture

AFT has a larger share of urban / mixed-profession membership and a smaller share of suburban classroom-only locals than NEA does. Its dues revenue base is more diverse and its members are more accustomed to differentiated pay structures (NYC HRA workers, hospital staff, etc., are AFT-affiliated). That permits AFT to take a slightly more open position on outcome-tied compensation without immediately threatening its dues base — though it remains the lead plaintiff in HFT v. HISD, the case that effectively ended VAM-based termination²⁵.

LIFO defense

Layoff-by-seniority is the single most defended contract provision in U.S. teacher CBAs — even where its disparate impact on high-poverty schools is well-documented¹⁶. The financial logic: senior members are the union's reliable dues-payers (longer track record, less likely to exit). LIFO protects them by definition. A performance-based layoff system would, in expectation, lay off a mix of newer and older teachers — reducing the “rational dues-payer” pool faster than LIFO does. The same Goldhaber finding that “LIFO is not very good for kids” is, from the union's revenue-maximization perspective, very good for the dues base.

The ATR and the “paid not to work” problem

The NYC Absent Teacher Reserve costs roughly $136 million per year. ATR members continue paying UFT dues during their time in the pool. From the union's perspective, a member in the ATR is financially equivalent to a member in a classroom; only the union dues column matters for revenue. The disposition of the ATR is therefore not, from the union's standpoint, a problem to be solved — it is a stable equilibrium to be defended¹⁴.

The pension dimension

The financial incentive also runs the other way: median U.S. teacher pension systems are only ~70% funded, and the national public-employee-pension shortfall stands at roughly $3 trillion⁵¹. Active members contribute the cash flow that keeps these systems running. A shrinking active membership directly threatens the actuarial assumptions of the pension funds — which is why unions oppose policy changes (vouchers, charter expansion, alternative certification) that would reduce district-employed-teacher headcount regardless of those changes' effect on student outcomes. AFT has reportedly used pension-board influence to apply pressure on fund managers it disfavors⁵¹, which extends the financial-incentive analysis into a less-visible but equally consequential domain.

What the union-decline trend implies for the next five years

Three plausible trajectories:

1. Continued slow attrition. 2018–2024 trend continues: ~1–2% membership loss per year in non-RTW states, larger in RTW or state-restricted states. National dues revenue stays flat or grows slightly via per-member-rate increases. Political spending unchanged. This is the modal scenario absent a major shock.
2. Acceleration via state-law changes. Tennessee enacted teacher-union restrictions in 2024 (the “Tampa Bay” pattern in Florida); other RTW Republican states could follow with bargaining-scope limits modeled on Wisconsin Act 10. Each new restrictive state adds a step-function loss to NEA / AFT revenue. Conditional on continued Republican gains in state legislatures.
3. Stabilization via a labor-coded shift. The unions reposition toward labor-political issues that resonate with younger workforce-entry teachers — AI-and-worker-displacement (see the New York chatbot bill, sponsored by AFT-aligned senators¹⁵); class-size limits; safety provisions. This is the trajectory implied by Hinchey's Workforce Stabilization Act and the broader 2026 AI package. Plausible if the 2026 election cycle reshapes the political map.

The bottom line: the policy positions documented in this brief are stable not because the empirical evidence supports them but because the financial machinery rewards them. The financial machinery has been weakened since 2018 but not broken. Until either membership declines steepen meaningfully or the dues check-off mechanism is undone in non-RTW states, the contract provisions that the empirical research most criticizes — LIFO, tenure-by-default, forced placement, ATR-style salary-without-placement — will remain the unions' default bargaining demands. The contract is not in the textbooks; it is on the dues spreadsheet.

Red states, blue states: the partisan mapAdded

Underneath the federal-arc story sits a sharper sub-pattern: the moves that make CBAs more outcome-tied (eliminating tenure, weakening collective bargaining, mandating VAM-based evaluation, enacting merit pay) cluster heavily in Republican-trifecta states; the moves that make CBAs less outcome-tied or that protect contract bargaining strength cluster in Democratic-trifecta states. The picture isn't pure — some red states have retreated from outcome-tied evaluation, and some blue states (notably DC) host the most ambitious surviving outcome-tied compensation system in the country — but the partisan signal is real.

How states sort, as of April 2026

The starting point is the underlying political map. As of April 2026, U.S. state governments split into roughly three categories³⁴:

• 23 Republican trifectas (governor + both legislative chambers): the bulk of the South (FL, TX, GA, AL, MS, LA, SC, TN, AR, OK), most of the Mountain West (ID, WY, MT, UT), and the Plains (KS, NE, ND, SD, IA, MO, IN, OH, WV, AK, NH).
• 16 Democratic trifectas: the Northeast (NY, NJ, MA, CT, RI, DE, MD), Mid-Atlantic plus DC, the West Coast (CA, OR, WA), Hawaii, Illinois, Colorado, New Mexico, Minnesota (lost trifecta in 2024 elections; included with reservations), and Virginia (gained trifecta 2024).
• 11 divided governments: AZ, KY, MI, NV, NC, PA, VT, WI, plus a few others where the legislature and governor split. These are the "purple" or "ambiguous" category for present purposes.

That partisan map maps imperfectly but recognizably onto the teacher-CBA terrain.

Collective bargaining itself

The starkest red/blue split is at the most fundamental layer: whether public-school teachers can bargain at all. Approximately 46 states permit teacher collective bargaining³⁵; the small set that prohibits or heavily restricts it is dominated by red Southern states. North Carolina and South Carolina expressly prohibit public-sector collective bargaining; Georgia and Texas restrict it to "meet and confer" arrangements that have no force of contract. Virginia repealed its prohibition in 2020 under a brief Democratic trifecta. Wisconsin's Act 10 (2011, signed by Republican Gov. Scott Walker) didn't formally prohibit teacher CB but gutted it — teachers can bargain only over base-wage increases capped at inflation; dues check-off is banned; unions must recertify annually³⁶. A Dane County circuit court struck down parts of Act 10 in December 2024 as a violation of the state constitution's equal-protection clause; the appeal is pending as of 2026³⁶.

None of these CB-restrictive moves were taken in a Democratic-trifecta state.

Tenure elimination is a red-state move

All four states that have functionally eliminated tenure for teachers acted under Republican control:

In addition, the District of Columbia (Democratic) and North Dakota (Republican) have no statewide tenure policy — but for different reasons. DC opted out because its school governance was reorganized under mayoral control with the IMPACT-style reform agenda (a blue-jurisdiction outcome-tied model). North Dakota left it to local school boards in a low-density, low-conflict policy environment⁶.

The 2024 Republican Party platform explicitly endorses ending teacher tenure nationwide; no comparable Democratic Party position exists at the national level.

Where strong outcome-tied evaluation lives

Among the surviving high-weight VAM systems, the partisan pattern is also clear:

• Louisiana (R-trifecta): 50% of evaluation is student growth; 35% via VAM³².
• Texas (R-trifecta): Texas VAM rolling into the Teacher Incentive Allotment in 2025–26³¹.
• Tennessee (R-trifecta): TVAAS (Tennessee Value-Added Assessment System) is among the country's longest-running VAM systems and remains operative.
• Florida (R-trifecta): had high VAM weights pre-2017; HB 7069 (R-trifecta legislation) eliminated the state-mandated component, leaving the choice to districts²⁷.
• Michigan (purple, divided government as of 2024): 25% test-score weight, holding pattern under SB 122²⁶.

The states that retreated from VAM-based evaluation since ESSA include both red retrenchment (AR, KS, OK, KY, AK, WY, NC) and blue concession (NM after the AFT lawsuit, DC). The mechanism is different: red retreats happen through legislative action under the "ESSA flexibility" framing; blue retreats happen through litigation pressure or quiet rule-change.

Where performance pay actually exists

The 67 teacher-pay bills introduced in 26 states in 2023, of which 11 became law in 9 states, sort by state with the same partisan signal³⁸:

• Republican-trifecta states have been the most active sources of performance-pay legislation: Utah's SB 173 (2023, R-trifecta) created a five-year pilot paying $10,000 bonuses to the top 5% of Utah teachers and $2,000–$5,000 to the next 20%³⁸. Tennessee, Arkansas, Florida, and Oklahoma all enacted teacher-pay measures in 2023 with performance components.
• Democratic-trifecta states have generally focused on across-the-board base-pay increases, not differentiated performance pay. Maryland and Washington enacted 2023 measures focused on raising minimum salaries.
• The major outlier: District of Columbia. DC is a Democratic jurisdiction but hosts IMPACT/IMPACTplus — the country's most generous outcome-tied compensation system — with bonuses up to $25,000 and career-earning differentials up to $3.7 million²². DC's exception traces to its unique mayoral-control governance and the legacy of the Rhee-era reform window in 2007–2010.

Public opinion is partisan, modestly

Public-opinion data on outcome-tied teacher pay shows a real but moderate partisan gap:

Even the lower number among Democrats — 47% — is higher than the 3.5% of districts that have actually adopted merit pay. The constraint isn't public opinion; it's the bargaining table.

Litigation venues split predictably

The major teacher-evaluation lawsuits have come out of red states and blue states for different reasons:

• Red-state venues: Houston (TX) — HFT v. HISD²⁰; Florida and Tennessee multiple cases²⁹. The pattern: aggressive outcome-tied regimes provoke procedural-due-process challenges.
• Blue-state venues: New York (Lederman)²¹; New Mexico (AFT-NM)³⁰. The pattern: state mandates (often holdover RTTT-era requirements) get challenged on the science of the underlying VAM, with state governments more willing to settle than fight.

The honest synthesis

The pattern is not "Republicans want outcome-tied teacher contracts and Democrats don't." Both parties have been involved in the federal push for outcome-tied evaluation (Race to the Top was a Democratic-administration program; ESSA's pull-back was bipartisan). The real partisan pattern is at the state-implementation level:

1. The most aggressive moves to make CBAs more outcome-tied (eliminating tenure, weakening bargaining, enacting merit pay, mandating high-weight VAM) are concentrated in Republican-trifecta states.
2. The most aggressive moves to preserve traditional CBA structures (statutory protections, mandatory bargaining, restrictions on RIF criteria) are concentrated in Democratic-trifecta states.
3. The retreat from outcome-tied evaluation since ESSA has happened in both red and blue states — but for different reasons. Red retreats are framed as deregulation; blue retreats are framed as research-driven course corrections.
4. The country's strongest working example of outcome-tied compensation (DC IMPACT) sits in a Democratic jurisdiction operating outside the standard state-CBA framework. There is no pure Democratic-trifecta-state IMPACT analog.

What the research consensus says

The empirical literature is not flattering to the strong form of outcome-tied teacher pay, and the professional research associations have said so on the record.

• American Statistical Association, April 2014: position statement cautioning that VAM should not be used as the primary or sole means of evaluating teachers, citing instability across years, sensitivity to test choice, and the small share of student-outcome variance that is attributable to teacher effects¹⁹.
• American Educational Research Association, November 2015: formal statement holding that VAM use in high-stakes employment decisions carries "considerable risks of misclassification and misinterpretation." AERA explicitly does not recommend a single methodology but lists eight conditions that must be met before VAM-based evaluation is defensible¹⁸.
• Amrein-Beardsley et al., 2020: "Houston, We Have a Lawsuit: A Cautionary Tale for the Implementation of Value-Added Models for High-Stakes Employment Decisions." Educational Researcher, 49(5)²⁹. Synthesizes the legal and methodological objections that produced the HFT settlement.
• Hanushek, 2014: Stanford economist Eric Hanushek's research on teacher quality finds large teacher-effect differences in student achievement, providing the strongest published empirical case for outcome-tied compensation. His work is the principal counterweight to the methodological-skepticism school.

The professional consensus is genuinely contested at the level of "how big is teacher effect, and how stably can it be measured?" but is more unified at the level of "whatever the answer, the algorithms commercially deployed in 2010–2015 were not stable enough to be the basis for firing people."

Where things stand now

As of the 2025–26 school year, the headline state of play is one of regulated retreat with localized exceptions:

• Compensation tier: Roughly half of large-district CBAs (NCTQ's 147-district sample) include some form of evaluation-tied performance pay³, but the bonus amounts are usually modest ($100–$5,000 universal range) and the strongest forms (multi-thousand-dollar bonuses with sustained-rating accelerators) are concentrated in a handful of districts — DC²², a small set of TIA-participating Texas districts³¹, and a few city outliers.
• Evaluation tier: 34 states still require objective student-growth measures, down from 43 in 2015². Where it persists, the weight has typically been lowered (Michigan 25%²⁶; many states 20% or "informational only"). Texas's 2025–26 VAM rollout is the most significant new entrant in years³¹.
• Job-security tier: Effectively absent at scale. Houston Federation of Teachers v. HISD is the deterrent precedent²⁰; New Mexico's settlement is the regional confirming case³⁰. Districts that retain VAM in evaluation almost universally do not use it as the sole basis for termination.
• Federal pressure tier: Removed in 2015 and not restored⁹. Neither the first nor second Trump administration nor the Biden administration restored the conditionality that had driven the 2009–2015 expansion. There is no current federal proposal to do so.

Why the share has fallen

Four overlapping reasons explain the trajectory:

1. Methodological problems with VAM proved durable. The 2014–2017 wave of measurement skepticism (ASA, AERA, peer-reviewed work in Educational Researcher) was not refuted; it was absorbed.
2. Litigation made high-stakes use risky. Houston, New Mexico, and the Lederman case taught districts that any opaque algorithm used for adverse employment action invites Fourteenth Amendment exposure.
3. Federal money stopped flowing. ESSA's removal of the conditionality made the political costs of pushing outcome-tied evaluation through state legislatures harder to justify locally.
4. COVID disruption created a permission structure for unwinding. Many states paused evaluations during the 2020–21 school year; the pause was politically convenient for restructuring or quietly retiring components that had been controversial since 2010.

What we still don't know

• The true national share. Without a comprehensive CBA database covering small districts, the NCTQ subsample is the floor. The actual share of all U.S. CBAs with any outcome ties may be substantially lower than 48% (small districts are typically less likely to have formalized performance-pay structures).
• Whether the surviving regimes work. Quasi-experimental evaluations of DC IMPACT have shown positive effects; ProComp's evaluations have been more mixed; Denver and Houston's outcomes have been deeply confounded by simultaneous policy changes. There is no clean cross-district comparative evaluation of outcome-tied vs. traditional compensation as currently implemented.
• Whether the Texas 2025–26 VAM rollout will scale or repeat the cycle. Texas's TIA framework is the most ambitious new experiment; its early-implementation data should be available by 2027.
• How the Janus v. AFSCME (2018) reduction in mandatory union dues has shifted local contract-bargaining dynamics. Public-sector union finances and political capacity have changed materially since 2018; whether that has affected willingness to bargain outcome-tied compensation is not yet well-documented.

Prior Manhattan Institute workAdded

The Manhattan Institute has been one of the most consistent voices on teacher-contract reform for two decades. Its publishing has run across the Institute's main reports, City Journal, and (more recently) The Bigger Apple. The Institute's general posture: skeptical of seniority-based personnel rules (LIFO, tenure-by-default), favorable to performance-based compensation, and consistently critical of contract provisions that protect underperforming teachers in ways that fall hardest on disadvantaged students.

Tenure, layoffs, and LIFO

• “Tackling NY Teacher Tenure” — Manhattan Institute. Direct argument for raising the tenure bar in New York, with explicit attention to the “due process so burdensome and with so little probability of success that most schools don't even attempt to remove the worst” problem this brief covers in the dismissal-procedures sub-section. manhattan.institute →
• “One Small Strike Against Teacher Seniority” — City Journal. Coverage of the Los Angeles court ruling against seniority-only layoff policies; argues the precedent could constrain LIFO contracts elsewhere. city-journal.org →
• “Life After Tenure” — City Journal. Analysis of Vergara v. California, the Los Angeles Superior Court decision that voided California's teacher tenure, seniority, and dismissal statutes (later reversed on appeal). The Institute reads the case as the constitutional-challenge model for LIFO-and-tenure reform. city-journal.org →
• “There's One Thing Worse Than Paying Bad Teachers Not to Work” — Manhattan Institute. The Institute's most-cited piece on NYC's Absent Teacher Reserve — the same $136M-per-year artifact this brief covers in the teacher-protections section. manhattan.institute →

Performance pay and outcome-tied compensation

• “The NYC Teacher Pay-for-Performance Program: Early Evidence from a Randomized Trial” — Manhattan Institute report. Empirical evaluation of the NYC SPBP (School-wide Performance Bonus Program); finds limited first-year effects on proficiency or school environment. Methodologically the closest analogue in the MI corpus to the Goldhaber LIFO work cited elsewhere in this brief. manhattan.institute →
• “Good Teachers Matter, So Pay the Best More” — Manhattan Institute. The Institute's affirmative case for differentiated compensation tied to student-outcome measures — the same theoretical position Eric Hanushek (footnoted in this brief) developed empirically. manhattan.institute →
• “Merit Pay: A Tool for Teaching” — Manhattan Institute. Survey of merit-pay-program designs (Cincinnati, TAP, others). manhattan.institute →
• “Merit Pay Will Improve N.Y.'s Schools” — Manhattan Institute (op-ed). The shorter affirmative case for NY-specific adoption. manhattan.institute →

Teacher quality, credentials, and measurement

• “Measuring Teacher Effectiveness: Credentials Unrelated to Student Achievement” — Manhattan Institute. Finding: external teacher credentials predict next to nothing about classroom performance — the foundational claim that motivates outcome-tied evaluation in MI's framing. manhattan.institute →
• “Rewarding Experienced Teachers: How Much Do Schools Really Pay?” — Manhattan Institute. Quantifies the experience-based pay premium in conventional salary schedules. manhattan.institute →
• “Better Pay, Fairer Pensions: Reforming Teacher Compensation” — Manhattan Institute. Pairs salary-schedule critique with pension-design critique. manhattan.institute →
• “The Teacher-Pay Myth” — Manhattan Institute. The standing rebuttal to the “teachers are underpaid” framing that dominates union messaging. manhattan.institute →

Union power and political economy

• “The Union That Devoured Education Reform” — City Journal. The Institute's most-cited historical piece on the UFT's contract pattern and how Bloomberg-era reforms (mutual consent in hiring, end of seniority transfers) were rolled back. city-journal.org →
• “How Teachers' Unions Handcuff Schools” — City Journal. Maps the contract provisions (uniform pay, seniority, work-day limits) back to the 1961 UFT contract that established them. city-journal.org →
• “The Fight over Flunked-Out Teachers” — City Journal. The reporting that established the ATR-as-political-artifact narrative this brief uses. city-journal.org →
• “Students vs. Status Quo” — City Journal. Source for the “high-poverty schools lose 30% more teachers in seniority-based layoffs” empirical claim that aligns with Goldhaber's LIFO research cited in this brief. city-journal.org →
• “New York's Teachers' Unions Are So Progressive, They're Even Alienating Teachers” — City Journal. Recent piece on intra-union political tension. city-journal.org →

The throughline: MI's argument is that contract provisions that are facially about teacher protection — LIFO, tenure-by-default, ATR-style paid-but-unplaced pools — have disparate impact on the students least able to absorb the cost. The empirical case in this brief (Goldhaber, Kraft, NCTQ) and the Institute's longer-running case point at the same conclusion through different methodologies.