The Netherlands Wants Inclusive Schools. Should Separate Ones Survive?

In September 2024, a United Nations disability-rights committee told the Netherlands that its education system should move toward ending separate special education. In May 2026, the Dutch government described its own destination for 2035: ordinary schools capable of educating far more children with disabilities, supported by specialist expertise, but with separate specialist facilities still available where necessary.

These are different destinations. The UN committee’s recommendation is to end specialized education as a distinct system. The Dutch government’s plan would retain some separate settings, potentially indefinitely. Neither side is proposing to abandon children who need intensive support - they disagree over whether keeping a separate educational pathway is itself an obstacle to inclusion or a legitimate way to provide it.

The Netherlands should aim to make separate placement an increasingly rare, individually justified exception rather than preserve it as a permanent, parallel default. But closing a school or moving a child into an ordinary classroom before equivalent support exists would be a poor way to pursue that ambition. The test of reform should be what children can actually learn, experience, and choose, not how successfully officials redraw the school map.

Two destinations for inclusive education

The legal starting point is Article 24 of the UN Convention on the Rights of Persons with Disabilities. It commits states to an inclusive education system, access to general education without discrimination on the basis of disability, reasonable accommodation, and individualized support directed toward full inclusion. The treaty is binding on the Netherlands. That does not settle, without further legal interpretation, whether every existing Dutch specialist school is unlawful or whether an individual family can compel a particular mainstream placement.

The Committee on the Rights of Persons with Disabilities takes a more specific position on institutional design. In its September 27, 2024 concluding observations, paragraphs 53 and 54, it criticized the continuation of what it called a dual education system. Paragraph 54(a) recommended a fully inclusive system that “ends specialized education”. The committee also called for immediate investment in ordinary schools, better teacher preparation, accommodations, changes to exemptions from compulsory schooling, measurable timelines, and participation by learners with disabilities and their representative organizations. Its vision is considerably more demanding than transferring pupils to buildings that remain unable to teach them.

Those concluding observations are recommendations from a treaty-monitoring body, not an enforceable Dutch court judgment ordering special schools to close. Their assessment concerned the implementation agenda as it stood in 2024, not policies announced almost two years later.

The government’s May 20, 2026 letter, Van passend naar inclusief onderwijs, is more ambitious than a defense of the existing arrangement. It envisages every school offering inclusive education by 2035, with specialist teachers and services increasingly working in ordinary schools. It describes neighborhood schools supported by regional networks, alongside specialist classes, departments, and potentially separate locations for children with more intensive needs. Its formula is zo kort als mogelijk en zo lang als noodzakelijk: for as short a time as possible, and as long as necessary.

The letter specifies progress reviews in 2029, 2032, and 2035. Its first phase concentrates on improving support for pupils and school teams; later phases are intended to expand and consolidate inclusive provision. It also acknowledges that laws, funding flows, teacher preparation, school buildings, and supervision will need attention. These are plans and milestones, not evidence that the intended system already exists.

The Tweede Kamer votes on October 6, 2026 brought that tension into view. The lower house adopted a motion calling for sufficient special education after 2035, another opposing reductions in appropriate specialist provision before suitable support exists elsewhere, and an amended motion seeking sufficient temporary specialist capacity in every region. Motions to abandon the move toward inclusion or its 2035 ambition (31 497-525, -528, and -530) were rejected. The adopted motions request government action; they neither enact individual placement rights nor demonstrate that staff and places are available. Protecting support against premature withdrawal is already a parliamentary demand.

The unresolved issue lies in that word necessary. Necessary compared with today’s understaffed mainstream school, or necessary even after an ordinary school has the expertise, facilities, and resources it ought to have? The two standards could produce very different numbers of children in separate settings by 2035.

How the separate pathway works

Dutch passend onderwijs, roughly education adapted to pupils’ needs, has operated through its current regional support architecture since 2014. An ordinary school is expected to provide support within defined legal duties, but the Netherlands also maintains distinct forms of more specialized schooling. Speciaal basisonderwijs (sbo) is a form of special primary education; speciaal onderwijs (so) serves children needing particular forms of specialist primary-level provision; and voortgezet speciaal onderwijs (vso) provides specialist secondary education. These are different school types, with different pupils and rules. Treating their results or admission routes as interchangeable would make the debate easier and less accurate.

The samenwerkingsverband passend onderwijs, a regional partnership of school boards, coordinates the regional support network and allocates additional resources. For many sbo, so, and vso placements, it also determines eligibility through a toelaatbaarheidsverklaring (TLV), an admissibility certificate. Arrangements for pupils with visual, hearing, or certain communication disabilities follow different specialist routes. A TLV establishes the relevant eligibility; it does not by itself oblige a particular school to produce an available, staffed place. Admission, funding, educational support, and transport are related but distinct decisions. The government’s explanation of specialist education outlines these differences.

There is also a real legal duty on ordinary schools. Under zorgplicht, the board responsible for a school must, in defined admission circumstances involving a pupil who needs extra support, arrange an appropriate alternative willing to admit the child before refusing admission. The Education Inspectorate describes this as a result obligation, subject to the statutory scope, not an unrestricted right to attend any school a family chooses. Separately, the Wet gelijke behandeling op grond van handicap of chronische ziekte, the disability and chronic-illness equal-treatment law, requires education providers to investigate and make effective disability-related adjustments unless doing so would impose a disproportionate burden.

These obligations do not disappear when a regional partnership controls some of the resources that might make a placement possible. Nor can the partnership simply substitute for the individual school’s admission duties. Parents may pursue different complaints, equality opinions, objections, or judicial remedies depending on the decision. Yet a legally correct placement decision and a classroom with the trained staff and equipment to implement it remain two different things.

The same specialist placement can reflect very different degrees of choice. A family may value a specialist school for its teaching, environment, or peers. Another may agree to the same placement because the local ordinary school cannot provide support. Both appear in an administrative record as children attending specialist education, although only one reflects a choice between genuinely workable alternatives. A policy that promises parents more options needs to measure the alternatives actually available to them.

Why a permanent dual system is hard to defend

The strongest argument against separate specialist schools begins with the obligations imposed on ordinary schools. If a child is routinely directed elsewhere when specialized teaching or accommodations are needed, the existence of a separate institution can become an alternative to improving the neighborhood school. The accumulated result may be fewer shared classrooms, school activities, and neighborhood relationships.

This is a structural argument, not proof that every separate school is harmful. Longer travel, difficult transitions, and lower expectations are plausible costs of separation, but vary by child and setting. The rights question is whether disability becomes the reason a child cannot receive an effective education alongside others, particularly when adjustments could have made that possible.

Separate schooling has a cost for nondisabled pupils, too: fewer opportunities to know disabled classmates as peers. A 2017 review of 12 controlled school-based studies found that direct-contact interventions improved children’s attitudes toward disability (average effect size d = 0.55). In a small study comparing two elementary classrooms, hearing pupils who learned alongside deaf and hard-of-hearing classmates showed more favorable attitudes toward deafness and better sign-language skills. Neither result establishes the long-term effects of inclusive schooling.

Merely sharing a room does not guarantee inclusion. A 2020 systematic review of 55 studies by University of Groningen researchers found more consistent attitude improvements when contact was combined with information about disability, while effects on friendships, belonging, and lasting social participation were less certain. Disabled children are not there to teach their peers tolerance; their own education and well-being come first. But a system that routinely separates children also limits opportunities for familiarity, shared experience, and learning across disability differences.

Furthermore, the United Nations disability-rights committee’s criticism was not abstract. Its 2024 observations identified waiting lists for specialist schools, compulsory-education exemptions, insufficient preparation of mainstream teachers, and inadequate support and accommodations, including for deaf and hard-of-hearing pupils. Those are findings and concerns attributed to the committee, not a verified census of every school or family. Its strongest point is that weak provision in ordinary schools should be improved rather than used to justify a permanent separate track.

A 2024 Education Inspectorate investigation found significant weaknesses in how additional support was planned, delivered, and evaluated across primary, secondary, and specialist education. The shortcomings appeared across school types. Calling one side inclusive and the other specialist obscures considerable variation in what schools actually provide.

Individual equality cases also complicate sweeping judgments. In a 2023 opinion, the Dutch Institute for Human Rights found that a school had inadequately investigated adjustments in an individual dispute. In a different 2024 case, it found no prohibited discrimination on the particular evidence. These are nonbinding opinions about specific facts, not competing estimates of how often schools discriminate. They nevertheless illustrate that accommodation duties require an actual assessment, not an automatic presumption that either the school or the family must be right.

Most importantly, inclusive education has to be defined by participation and learning as well as location. A pupil sitting in an ordinary classroom without accessible materials, communication support, a manageable sensory environment, or properly trained assistance has gained a mainstream address. Whether that pupil has gained a better education is a separate question. This is why the committee’s call for resources, training, and the participation of disabled learners belongs at the center of its argument, rather than in a footnote to a closure demand.

The specialist-school argument deserves more than a disclaimer

Some children need a combination of expertise, facilities, communication methods, predictable routines, intensive teaching, or physical assistance that an ordinary school cannot currently supply. A well-run specialist school may provide smaller groups, experienced teachers, suitable equipment, and peers with whom a pupil feels understood. For some pupils and families, those features are valuable in themselves. Making an ordinary school theoretically available would be a poor exchange if the child lost effective teaching, confidence, or safety in practice.

Classmates and teachers also have legitimate interests in a functioning and safe learning environment. They should not be asked to compensate for an inclusion reform that provides neither additional staff nor appropriate training. Likewise, a specialist professional cannot necessarily divide a demanding caseload among several ordinary schools without losing the concentration of expertise that made the original service work. New duties on paper do not instantly create skilled workers, suitable buildings, or specialized equipment.

The present specialist system cannot simply be dismissed as inherently inferior. In its 2026 sector report, the Education Inspectorate reported that 81% of 185 inspected so and vso schools met its overall sufficient-quality standard in sample-based investigations during the 2023–24 and 2024–25 school years. The same report noted weaknesses requiring improvement even at schools assessed as sufficient. This is meaningful inspection evidence about schools. It does not mean that 81% of children in specialist education are thriving, or establish that a separate placement caused better academic or emotional outcomes than an adequately supported ordinary school would have.

The comparison is genuinely difficult. In a 2024 study of different educational placements, the Inspectorate found results suggesting that, for some pupils near a placement boundary, less separate educational routes may be associated with better subsequent education or labor-market prospects. Those comparisons included sbo versus so and, at the secondary level, regular secondary education (vo), practical education (praktijkonderwijs, pro), and vso. Some analyses grouped vo and pro against vso, while others compared vo with pro and vso combined; they were not a single mainstream-versus-special-school experiment. Crucially, pupils are not randomly allocated. Differences in support needs can influence both placement and later results; the study’s attempts to address selection did not establish an unambiguous general causal effect. It also could not tell us whether pupils felt safer, more accepted, or better supported during their school years.

This uncertainty cuts in both directions. It prevents a confident claim that moving children out of specialist education will improve their outcomes. It equally prevents the inspection ratings or positive specialist-school experiences from proving that the Netherlands needs to maintain a permanent, separately organized national track. That is a choice about how to arrange expertise, access, and rights across the whole system, not merely whether a good school deserves to exist.

The practical risks of transition fall unevenly. An ordinary school might gain additional staff and facilities over several years. A child who needs intensive support next September cannot wait for the 2035 destination to become reality. If a specialist placement is the only appropriately staffed option now, preserving it protects the child. It does not follow that the same separation should remain the default option for future pupils once ordinary provision has changed.

The test the Netherlands should set for 2035

The May 2026 roadmap already acknowledges several necessary ingredients: more specialist expertise inside ordinary schools, changes to law and funding, better teacher preparation, accessible facilities, and staged reviews in 2029, 2032, and 2035. It also describes existing experiments in which pupils registered in specialist education can receive teaching in ordinary schools with specialist support. The open question is how to turn those intentions into a standard that can actually determine when separate education remains justified.

The Netherlands should start by specifying a credible minimum offer in ordinary schools and the regional teams supporting them. Individual teachers need not become experts in every disability; a school should know which specialized teaching, communication assistance, accessible materials, quiet spaces, and behavioral or physical support it can obtain, on what timetable, and from whom. The Ministerie van Onderwijs, Cultuur en Wetenschap (OCW), the Dutch Ministry of Education, Culture and Science, should define and finance the relevant national expectations; school boards should be accountable for delivery; and regional partnerships and successor networks should make specialist capacity available rather than merely provide referrals. These allocations would need to be translated into workable funding and legal rules, not assumed to be fully settled already.

Transition financing is unavoidable. For a time, the system may have to support both competent separate schools and the expansion of ordinary-school capacity, while protecting specialists from being scattered so thinly that neither setting works well. The government’s letter promises to investigate funding redistribution and transition costs. It should specify money, staffing, and sequence before making separate-placement reductions a target. One cannot responsibly close a staffed place and count the resulting enrollment elsewhere as progress without checking what education the pupil receives there.

Placement decisions should begin with a presumption in favor of an inclusive neighborhood placement, subject to an individual, reviewable assessment of actual needs and available adjustments. Where a separate specialist setting is recommended, the decision should explain what the accessible mainstream alternative cannot yet provide, whether the obstacle is temporary, and how the pupil and parents were involved. Plans should include a meaningful review and a possible route back, without making a child repeatedly prove the need for support that is plainly enduring. Families who prefer a specialist environment deserve to be heard, although a choice offered only because ordinary provision is inadequate should not be mistaken for proof that separation is necessary.

The checkpoints could make that presumption measurable. By 2029, government and school networks should be able to show which support is reliably available in ordinary settings, where specialist vacancies persist, and who finances each missing service. By 2032, an evaluation should examine the experiences and progress of pupils who have moved between settings, with attention to differences in needs rather than a crude comparison of average test scores. By 2035, the central question should be whether separate placement is a genuinely exceptional, child-specific necessity or still the routine destination for needs that ordinary schools have never been equipped to accommodate. These are proposed tests for the existing checkpoints, not targets already adopted in the May 2026 letter.

A useful scorecard would track the time until appropriate support begins, actual teaching received, learning and qualifications suited to pupils’ goals, safety, pupils’ own assessments of belonging, travel burdens, and whether decisions are reconsidered when circumstances change. It should also record when an attempted mainstream placement fails and what happens next. Enrollment reductions can conceal unmet needs, while unchanged specialist-school totals can conceal better mainstream access. Child-level monitoring also needs privacy protections.

My preferred long-term arrangement is therefore closer to the committee’s direction than to treating two institutional tracks as equally permanent. Ordinary schools should progressively acquire responsibility and capacity for pupils they currently refer away, while specialist expertise becomes part of a common public education offer. But I would not impose an unconditional deadline to close every separate site; a small number may remain necessary for particular pupils even in a substantially stronger inclusive system, especially if careful evidence demonstrates advantages that cannot reasonably be replicated elsewhere. This is less categorical than the committee’s recommendation, but more demanding than the government’s still-developing necessity principle because it calls for a reviewable child-specific comparison with a genuinely supported inclusive option.

What could change that judgment? Reliable evidence that children with identifiable intensive needs consistently learn better, feel safer, and prefer a separate specialist setting even when a genuinely well-resourced inclusive option is available would strengthen the case for an enduring specialist option. Evidence that separate placement chiefly reflects unsupported neighborhood schools, and that appropriately supported pupils do as well or better in them, would strengthen the case for ending it more completely. Existing inspection data cannot settle that comparison.

By 2035, the Netherlands should no longer accept a separate school as the routine answer to an ordinary school’s lack of capacity. It should also refuse to make individual children pay for an institutional ambition by removing support before a credible alternative exists. A separate placement must be justified by what the child needs after reasonable inclusive options have been made real, not by what the system has found convenient to keep separate.