The legal-aid reforms did what government said they would do in one important sense.

They cut the legal-aid bill.

Civil legal-aid spending fell sharply after the changes introduced around the Legal Aid, Sentencing and Punishment of Offenders Act 2012 — LASPO — and the associated fee and eligibility reforms.

That part is not seriously in doubt.

The harder question is the one government never managed to answer properly:

How much did the country save once the costs that moved somewhere else were counted too?

After examining the original savings models, court effects, mediation, provider capacity, housing, debt, family cases, immigration, health, household burdens and later government corrections, the evidence supports a narrower conclusion than either side of the political argument may want.

The budget saving was real. Cost displacement was real too. But the evidence does not support a defensible whole-system net figure — and it does not prove that LASPO cost more than it saved.

That measurement failure is the story.

The saving was real

There is no need to pretend the legal-aid budget did not fall.

The National Audit Office’s 2014 examination estimated an eventual annual reduction of about £300 million in civil Legal Aid Agency spending using actual 2013/14 matter starts.

But that number needs a warning label.

It was a model, not an audited annual cash saving. It also covered the effect of LASPO scope changes and the earlier 10% provider-fee reduction. It is therefore not a clean figure for “LASPO alone”.

Other prominent numbers describe different things again.

The Government’s £350 million figure was a wider legal-aid reform forecast for 2014/15. The long-run Impact Assessment components used different time bases. The observed fall in civil legal-aid expenditure is real accounting evidence, but payment lags, pre-existing trends, fee changes and case mix mean it cannot simply be relabelled as a causal LASPO saving.

The safe conclusion is still substantial: public expenditure on civil legal aid was reduced materially.

The problem starts when that reduction is treated as if it automatically describes the effect on the public purse as a whole.

Some of the saving moved immediately

The 2014 NAO report contains two unusually clear examples of why a departmental saving is not necessarily a government-wide saving.

First, lower payments by the Legal Aid Agency also meant lower VAT receipts for HM Treasury.

The NAO estimated that this reduced tax revenue by about £60 million in 2013/14, roughly 20% of the Agency saving in its model.

That gives one bounded calculation that can be made without mixing unrelated evidence:

about £300m gross Agency spending reduction minus about £60m VAT-revenue loss = about £240m before other displaced costs.

That £240m is not the final answer.

It is not audited Treasury cash saved by LASPO. It is not LASPO-scope-only. It is not a whole-system net. It is simply the part of one NAO model left after applying an offset that the same audit explicitly connected to the Agency saving.

The second example is court fees.

When legal aid paid a person’s court fee, that payment appeared in the legal-aid account. Once legal aid was removed, some people instead became eligible for court-fee waivers.

The Ministry estimated the resulting cost to the courts at £10 million.

The NAO’s description matters: it said this was a transfer of costs within the Ministry of Justice, not an overall Ministry saving.

That is direct evidence of cost displacement.

But it still should not be casually subtracted from the £240m figure. The evidence available to this investigation does not match its exact period and overlap closely enough to justify that arithmetic.

The famous court-cost estimate is weaker than it looks

Another number often associated with LASPO is roughly £3 million in additional family-court costs linked to more litigants in person.

The method can be reconstructed.

The estimate used 11,144 additional family cases in which neither party was represented, an average family-court case cost of £547, and an assumption that cases involving litigants in person used 50% more court time.

The arithmetic produces just over £3 million.

The weakness is the 50% assumption.

Later parliamentary scrutiny found that the Ministry did not have reliable actual hearing-length data and did not accept the anecdotal 50% figure as established fact. Later analyses still did not produce the clean causal series needed to update the estimate across subsequent years.

So the £3m figure is useful as an early model of a possible operational burden.

It is not a number that can responsibly be multiplied by thirteen years and added to a bill for LASPO.

Mediation did not replace lawyers as expected

One of the reform’s intended substitutions was family mediation.

The expectation was that removing legal aid from much private-family litigation would push more disputes toward mediation.

Initially, the opposite happened.

The NAO recorded family mediation assessments falling from 30,668 in 2012/13 to 13,423 in 2013/14 — a fall of about 56%. Mediation starts also fell.

The mechanism was not mysterious. Solicitors had previously been an important route into mediation, and removing funded legal consultation also removed part of the referral route.

Self-representation increased at the same time.

That establishes a real implementation effect.

It does not establish that every later family-court delay, difficult hearing or safeguarding problem was caused by LASPO. Case complexity, court procedure and later policy changes matter too.

The distinction is important because the strongest version of the story is not the one with the largest speculative number. It is the one that survives competing explanations.

Wider costs exist, but most cannot be priced honestly

The investigation tested six wider routes: housing; debt and welfare; family and safeguarding; non-asylum immigration; health and wellbeing; and unpaid household or informal-support work.

Across those areas, the same pattern appeared repeatedly.

There is credible evidence that early legal advice can prevent or reduce escalation in some cases. There is credible evidence that people without representation absorb more navigation, preparation, form-filling, travel and support work. There is credible evidence of provider shortages and people being turned away. There are plausible routes from unresolved legal problems into homelessness processes, enforcement, family-court difficulties and poorer wellbeing.

But plausible mechanism is not the same thing as a national monetary total.

Government itself tried to close part of this gap through the Early Legal Advice Pilot. The pilot was designed to test whether earlier housing, debt and welfare advice produced better outcomes and downstream public benefits.

It did not produce the robust treatment-effect evidence needed for that calculation because uptake was too low for the intended trial to proceed.

The resulting model contains useful unit-cost assumptions and scenarios.

It does not tell us how many pounds LASPO transferred into homelessness services, health care or other budgets.

The same boundary applies elsewhere.

A lawful welfare payment secured after advice is not a “cost caused by advice”. A household’s unpaid time is a real burden but not Treasury cash. A QALY is not an NHS cash saving. Current asylum accommodation spending cannot be assigned to LASPO when asylum legal aid remained in scope. Current homelessness expenditure cannot simply be multiplied by an assumed legal-advice effect.

Those shortcuts would manufacture precision the evidence does not contain.

Provider strain is real — and not all of it is LASPO

Civil legal aid also depends on there being someone able and willing to provide it.

The provider evidence shows a system under strain.

Official and parliamentary evidence records falling provider numbers in many categories, excess demand, turn-away evidence and geographic gaps in face-to-face provision. In June 2026, the Justice Committee reported 21 areas without face-to-face housing and debt provision, although the Legal Aid Agency said outreach covered nine of them. Nine family areas were reported below minimum provision standards.

Those are practical access problems, not merely points on a map.

But present provider weakness has several causes: fee levels, recruitment and retention, contract design, administration, later policy changes and the wider labour market, as well as the long-run effects of scope reduction.

Treating every provider shortage observed in 2026 as a clean LASPO effect would be another form of bad accounting.

Government did respond

A fair account also has to include the evidence that cuts against a simple failure narrative.

Government did not leave the 2013 settlement completely untouched.

Legal aid was restored for the non-asylum immigration and citizenship matters of separated migrant children. The mandatory telephone gateway was removed. Exceptional Case Funding remained available. The Housing Loss Prevention Advice Service created an early-advice route for people at risk of losing their homes. Targeted housing and immigration fee increases have been introduced, with the controlled-work element in force from December 2025.

These are real corrections and real spending decisions.

They do not prove that LASPO was wrong.

Nor do they prove that the system is now adequate.

They show something more useful: governments themselves have repeatedly adjusted the system where access, evidence or provider sustainability required it.

The question is whether those corrections have been evaluated well enough to tell us what they achieved.

Often, they have not.

Thirteen years later, the central accounting gap remains

The most important finding does not come from an advocacy group opposing LASPO.

It comes from the National Audit Office.

In 2024, the NAO concluded that the Ministry of Justice had achieved its objective of significantly reducing legal-aid spending. But it also said the Ministry remained aware that costs may have shifted elsewhere in government and did not know their scale.

The consequence was stark: the Ministry could not demonstrate how much the reforms represented a spending reduction for the public purse overall.

The Public Accounts Committee reached a similar conclusion and pressed for cross-government work on cost-shunting.

Government later recorded that recommendation as implemented through engagement, analysis and new work.

But by July 2026 the Justice Committee was still reporting that attempts to understand the costs of not providing legal advice had not produced convincing results because of poor methodology.

That is not evidence that the hidden costs must be larger than the saving.

It is evidence that government still did not know.

What we can say — and what we cannot

The evidence now supports four conclusions.

First, the civil legal-aid budget saving was substantial. Any account that denies that is weaker than the evidence.

Second, some cost displacement is directly established. The VAT-revenue offset and court-fee transfer are not theories about what might have happened; they were identified in the government’s own accounting environment and the NAO audit.

Third, wider displacement is supported but mostly not monetisable. Courts, providers, households and other services can carry work or pressure that no longer appears in the legal-aid budget, but the incidence and accounting boundaries are not good enough for a national total.

Fourth, a false-economy claim is not proven. There is no matched evidence showing that the displaced public costs exceeded the legal-aid saving.

Those conclusions fit together.

The mistake would be to force them into a simpler political story.

LASPO exposes a recurring weakness in public-sector accounting.

A department can often measure the money it stops spending.

It is much harder to measure the work transferred to another service, the tax revenue that disappears, the case that becomes harder to resolve, the person who spends days navigating a process alone, or the crisis that appears later in a different budget.

That does not mean every cut is a false economy.

It means a saving in one ledger is not proof of a saving to the country.

For a reform sold partly through value for money, government should be able to answer two questions with comparable evidence:

What did we stop spending?

And what happened to the cost after we stopped spending it?

On civil legal aid, the first question was answered much better than the second.

What to watch next

The next useful evidence is not another headline total.

It is whether the government finally produces a repeatable cross-department method for measuring what happens when eligible people cannot obtain early legal advice.

That means watching the response to the Justice Committee’s July 2026 report, the delayed means-test reforms, the remaining civil fee changes, provider-demand evidence and any evaluation that links early legal help to actual downstream service use rather than modelled assumptions.

Until then, the most defensible conclusion remains the least convenient one.

LASPO saved a substantial amount from the legal-aid budget. Some of the cost moved elsewhere. Government still cannot say, on comparable evidence, what the country saved overall.