How do local authorities improve SEND outcomes?

You are reading this against a deadline. A civil servant has to sign off your Local SEND reform plan before the High Needs Stability Grant and the 90% deficit write-off land, and the cabinet wants to know what you are actually going to do that a decade of rising spend did not. The honest answer is that the question has changed under you. It is no longer “how do we stabilise the budget”. It is “which of the levers we control move outcomes, and which only move money”. This is the difference, set out the way the plan has to read.

What changed in February 2026, and why outcomes are now the test

The February 2026 settlement turned a budget problem into an outcomes problem. Government will absorb 90% of eligible historic Dedicated Schools Grant deficits, with your authority carrying the remaining 10%, and the replacement money flows as the High Needs Stability Grant from autumn 2026-27. There is one condition. You only get it with an approved Local SEND reform plan (DfE explanatory note, 9 February 2026). The plan is the gate, not the paperwork after the gate.

Get the timeline right, because the old framing is wrong now. The statutory override, the accounting mechanism that keeps these deficits off your general fund, was due to fall away in March 2026. It has been extended to the end of 2027-28, so to March 2028 (Tes, 2025). From 2028-29 high needs spend sits inside the overall budget. So the cliff did not arrive in March 2026. It moved. If your plan still talks about a March 2026 sunset, it is two years out of date before anyone reads it. The fuller mechanics are in our note on the DSG statutory override.

The Safety Valve programme is winding down. The reason matters for the room: the new settlement is more generous than most Safety Valve deals would have been, existing agreements with instalments running to 2031-32 are honoured, and no authority is left worse off for having signed one (National Education Union, 2026). Nobody is being punished for the path they took to get here.

Here is why the gate is built the way it is. The Isos Partnership review for the County Councils Network and the LGA found that high needs spend roughly trebled, from about £4bn to a projected £12bn by 2026, while attainment for children with SEND did not improve and some measures went backwards (Isos Partnership for CCN and LGA, 2024). Treasury read that and drew the obvious conclusion. They will clear the debt. The price is a credible plan to improve outcomes, not another plan to spend more carefully.

Which outcomes you are actually accountable for

Split the word “outcomes” into two piles before you write a target, because you control one pile directly and only influence the other. Process metrics sit in the first pile: EHC needs-assessment and plan timeliness, whether annual reviews are current, dispute and appeal rates. Population outcomes sit in the second: attainment, attendance, suspensions and exclusions, post-16 destinations, parental confidence. Mixing them up is how plans end up promising things the authority cannot deliver alone.

Your most visible process lever is the one that is failing. Only 46.4% of new EHC plans were issued inside the 20-week duty in 2024, the lowest figure on record, and the spread is brutal: 12 authorities under 10%, 21 over 90% (DfE, EHC plans in England, 2025). That is not a demand story. Two authorities facing similar pressure can sit at opposite ends of that range, which means a large part of it is in your gift.

The volume underneath explains the strain without excusing it. There were 638,700 children and young people with an EHC plan in January 2025, up 10.8% in a year, with 97,700 new plans issued in 2024 and 154,500 assessment requests (DfE, 2025). Demand is real. It is also not a defence in an inspection.

The accountability frame is wider than your own service, and your plan has to reflect that. Area SEND inspections, run jointly by Ofsted and CQC under the framework that came in from January 2023, judge the lived experiences and outcomes of children across the whole local area partnership: your SEND service, integrated care board health services and social care, together (Ofsted and CQC, 2023). Outcomes are a joint product. You cannot subcontract the bits you find awkward.

Lever 1: inclusive mainstream capacity

Most outcomes are won or lost in ordinary classrooms, before anyone writes a plan. The cross-system consensus is unusually settled on this. The Isos review, the Education Committee's 2025 report Solving the SEND crisis, and the 2026 Schools White Paper all land in the same place: when more needs are met in mainstream through ordinarily available provision, the drift toward EHCPs and special-school placements slows, because much of that drift is mainstream running out of capacity rather than children getting more complex.

The levers here are yours to pull. A clear, enforced offer of ordinarily available provision in the Local Offer, so schools and families both know what every mainstream setting should provide without a plan. SEND training and the graduated approach (the assess, plan, do, review cycle in the SEND Code of Practice 2015, paras 6.36 to 6.39) embedded across schools rather than printed in a policy. Targeted help in the early years, before or without a diagnosis, where it is cheapest and works best. We set out the authority-side version of this in how LAs build inclusive mainstream capacity.

The teaching assistant question, answered honestly

Teaching assistant deployment is the single most evidence-rich classroom lever you can champion, and it is the one most authorities get backwards. The Education Endowment Foundation's guidance Making Best Use of Teaching Assistants is blunt about the failure mode: the “Velcro TA”, where the lowest-attaining children are routinely taught by the least-qualified adult in the room and learn to wait for a prompt rather than to think (Education Endowment Foundation). Deploy TAs to add to the teacher, not to replace them, and to build independence through scaffolding and least-to-most prompting.

The positive case is just as specific. When TAs deliver structured, well-trained one-to-one and small-group interventions, the EEF Toolkit records on average around four months of additional progress (EEF Toolkit). Read the condition carefully: it is structure and training that buys the four months, not headcount. An authority that funds 200 more TAs without changing how they are used has spent money and bought nothing. That is exactly the trap the reform plan exists to avoid. Our piece on building independence rather than TA dependence carries the classroom detail.

The money is now lining up behind this. The 2026 Schools White Paper Every Child Achieving and Thriving backs inclusive mainstream with around £1.8bn over three years for an “Experts at Hand” service, pooling educational psychologists, occupational therapists, speech and language therapists and specialist SEND teachers so schools can reach them with or without an EHCP. It adds over £3.7bn of capital for 60,000 new specialist places and £7bn more for SEND support in 2028-29 than in 2025-26 (DfE Schools White Paper, 2026). Your reform plan is where you show how that capital buys capacity rather than absorbing demand you could have met earlier and cheaper.

Lever 2: timely, lawful, consistent process

The 20-week duty is not negotiable and it is not a stretch target. The obligation to assess and issue inside 20 weeks runs from the Children and Families Act 2014 (s.36) and the SEND Regulations 2014, and sustained failure is both a legal liability and a recurring inspection finding. The clock is the floor, not the achievement.

Read your appeal data as a quality signal, not just a demand signal. A high overturn rate at the SEND Tribunal usually points upstream, to decisions that were not consistent or evidence-led in the first place, not simply to pushy parents. Avoidable disputes burn officer time twice, once making the original decision and again defending it, and they corrode the parental confidence inspectors ask families about directly. The practical version of fixing this sits in how LAs reduce SEND tribunal losses.

Annual reviews are the quiet one. Kept current and acted on (CFA 2014 s.44 and the SEND Regulations 2014), they are where a plan stays matched to a changing child. Left to lapse, they are one of the most reliable drivers of poor outcomes nobody puts in a board paper, because a plan that is three years stale is funding last year's need. If you carry a backlog, clearing it is itself an outcomes intervention, not just housekeeping; the method is in how to clear an annual review backlog.

Co-production is a duty and a shortcut at once. The law requires you to have regard to the views, wishes and feelings of the child and parent (CFA 2014 s.19). Beyond the duty, plans built with families tend to specify provision that actually fits, which means fewer amendments, fewer appeals and fewer plans that name something nobody can deliver.

One decode for whoever signs the plan off. When a service reports “we are improving timeliness”, ask what it traded to get there. The fastest way to lift a 20-week figure is to clear a backlog by issuing thinner plans faster, vaguer outcomes, looser provision, less quantified support. That moves the clock and weakens the plan, and it is precisely what a tribunal or an inspector exposes. Timeliness and quality are not a trade. The plan has to hold both.

Lever 3: the local area partnership

Inspections judge the whole partnership, and the first two years say it is the hardest thing to get right. Of the first 26 partnerships inspected under the 2023 framework, only 7 were positive, 11 were inconsistent, and 8 had significant concerns (Ofsted and CQC, 2025). That is the bar the next inspection sets against you, and most authorities are not clearing it.

Health is the recurrent weak link. The Education Committee's 2025 report Solving the SEND crisis called explicitly for health services to step up and for far better education-health collaboration (House of Commons Education Committee, 2025). Integrated care boards are co-accountable for SEND outcomes inside the partnership, which means the therapies named in plans are a shared delivery problem, not yours to chase alone. We unpack the interface in how LAs work with health partners on EHCPs.

Name the law instead of saying “partnership”. The duty to commission education, health and care provision jointly sits in the Children and Families Act 2014 (s.26), and the broader duty on bodies to cooperate sits in the Children Act 1989 (s.27). Those two sections are the legal spine of joint working. A reform plan that describes collaboration without anchoring it to them reads as aspiration; one that cites them reads as a system that knows what it owes. The mechanics are in what joint commissioning for SEND means.

Then the concrete things an inspector looks for. A functioning designated medical or clinical officer interface, so health has a named route into your processes. Joint commissioning arrangements that exist in practice, not just in a memorandum. Shared data, so the partnership sees the same children the same way.

The Local Offer is the public test of all of it (CFA 2014 s.30). It is the duty to publish and keep current what support is available, and co-produced with families it doubles as a signal of whether the partnership's provision is real and findable or notional. A thin, stale Local Offer tells an inspector more than a polished self-evaluation; keeping it live is covered in how LAs keep the Local Offer up to date.

Lever 4: a reform plan that survives contact with reality

The reform plan is now what releases the money, so write it as a genuine outcomes and sufficiency plan, not a deficit-reduction spreadsheet with a narrative bolted on. A civil servant signing it off is checking for the same thing the next inspection will: targets that move children, owned by a partnership that can deliver them. A plan that only balances the high needs block answers a question nobody is asking any more.

Write it toward the system that is coming, not the one you are leaving. The 2026 Schools White Paper and the Education for All Bill announced in the 2026 King's Speech set a phased direction: an Individual Support Plan for every child with SEND in every school, nursery and college, becoming statutory from September 2029; National Inclusion Standards in place by 2028; EHCPs retained for the most complex needs, with no change to EHCP support before September 2030 and reform phased through to around 2035 (DfE Schools White Paper, 2026; House of Commons Library, 2026). Build the plan so it bends toward that without overclaiming a settled endpoint, because the detail will move.

Carry the transitional protections into your messaging from day one, because staff and families will ask. The White Paper is explicit that no child currently getting effective support loses it, every pupil with a special-school place in September 2029 can stay, and no child is moved from a special school against their wishes (DfE Schools White Paper, 2026). Say that plainly and early. Anxiety is a delivery risk, and reassurance is cheap.

What the plan has to contain to pass

Treat the following as the load-bearing structure, ordered the way a sign-off reads it: sufficiency first, then workforce, then accountable targets.

1. Sufficiency. Forecast specialist-place need against the 60,000 new places and £3.7bn capital, so you are building to a number, not bidding against a queue.
2. Workforce. Build the “Experts at Hand” model (around £1.8bn over three years) into early help, so an educational psychologist or speech and language therapist is reachable with or without an EHCP, which is what actually reduces plan demand.
3. Accountable targets. Set measurable outcome targets and let the partnership own them jointly: 20-week timeliness with a quality check, mainstream inclusion rates, the suspension and exclusion gap for children with SEND, parental confidence. Those are the exact things the next Area SEND inspection and the DfE sign-off will test.

If you do one thing this quarter, make it the targets. Most reform plans fail sign-off not on ambition but on accountability: outcomes nobody owns and numbers nobody is tracking. A plan with four jointly owned, measurable outcome targets, each tied to a lever in this article, is one a civil servant can sign and a partnership can be held to. The runbook for the inspection that checks it is in how LAs prepare for an Area SEND inspection.

References

Explanatory note on the government's approach to Dedicated Schools Grant deficits (DfE / GOV.UK, 9 February 2026). SEND funding: government extends councils' statutory override (Tes, 2025). Government to write off 90% of councils' SEND deficits (Schools Week, 2026). Safety Valve and Delivering Better Value programmes (National Education Union, 2026). Educational outcomes for SEND pupils have failed to improve over the last decade despite costs (Isos Partnership for the County Councils Network and LGA, 2024). Education, health and care plans in England 2025 (DfE / Explore Education Statistics, 2025). Area SEND framework: findings from the first two years of inspections (Ofsted, CQC and DfE, 2025). Area SEND inspections framework and handbook (Ofsted and CQC, 2023). Making Best Use of Teaching Assistants, and the Teaching and Learning Toolkit (Education Endowment Foundation). Schools White Paper: Every Child Achieving and Thriving (DfE / The Education Hub, 2026). Research briefing on the 2026 Schools White Paper and Education for All Bill (House of Commons Library, 2026). Solving the SEND crisis (House of Commons Education Committee, 2025). Children and Families Act 2014, sections 19, 26, 30, 36 and 44; Children Act 1989, section 27; SEND Regulations 2014; SEND Code of Practice 2015, paras 6.36 to 6.39 (legislation.gov.uk and GOV.UK).

This article is general information for local authority officers, not legal or financial advice on your authority's specific position. Figures and timelines here carry a year for a reason; before the plan goes for sign-off, re-check the live ones against the current DfE and Treasury guidance, because in this field the statute moves slowly and the settlement detail moves every few months.