Brexit At Ten
The Audit

The most useful Brexit number is no longer the loudest one. In 2016, the Treasury warned of an immediate recession. It never came. The Centre for European Reform built a widely cited model that compared Britain to an imaginary "twin" version of itself that stayed in the EU. Its author quietly retired the model in 2023 after the 2022 energy crisis distorted the comparison — the imaginary twin was made up of countries that took an even harder hit from gas prices, which exaggerated the gap. That does not make the Brexit cost vanish. It makes the honest range narrower, slower and more boring than either side spent a decade arguing about.

The serious independent estimate now belongs to the National Institute of Economic and Social Research, the country's oldest economics think-tank. Its model, updated quarterly, puts UK output 2.5 per cent lower in 2023 than it would have been under continued EU membership, with wages 5.8 per cent lower, productivity 1.3 per cent lower and investment 12.4 per cent lower (NIESR, Revisiting the Effect of Brexit, November 2023). The Treasury's independent forecaster, the Office for Budget Responsibility, assumes a long-run 4 per cent productivity drag from a 15 per cent fall in how much Britain trades. That number is the assumption inside the official forecast — not a measured outcome. The distinction matters whenever a politician quotes it.

The strongest dissenting voices have more substance than they are usually given credit for. Briefings for Britain, a pro-Leave research outlet, compiled OECD growth data and pointed out that between mid-2016 and mid-2022 the UK economy grew 6.8 per cent — slower than France at 7.6 per cent, but faster than Germany at 5.5 per cent and Italy at 4.0 per cent. The UK was in the middle of its peer group, not visibly behind it (Gudgin, Jessop and Western, October 2022). The free-market Institute of Economic Affairs added that UK trade did not collapse: goods exports to the EU rose 13.5 per cent in cash terms between 2019 and 2022, while exports to the rest of the world rose 14.3 per cent (McBride, November 2023). Treat these as critiques of overclaim, not as evidence Brexit was free. The basic economic logic still applies: a mid-sized economy made trade with its nearest large market more expensive, and trade with that market grew more slowly than it would have done.

The migration promise was not false in the narrow sense. The UK ended free movement on 31 December 2020 and built a new visa system that treats EU citizens the same as everyone else. Net migration from the EU collapsed: from a peak of plus 253,000 in 2016 to minus 42,000 in the year to December 2025 (ONS, Long-term international migration, May 2026). That is a real sovereign choice, and a real result. What happened next was not.

Total net migration peaked at 891,000 in 2022 and was still 848,000 in 2023 — more than three times the pre-Brexit average. The new system did not deliver lower numbers. It delivered a different mix. The reason is that the rules collided with three unrelated events.

The first was the British National (Overseas) visa, opened in January 2021 for Hong Kong residents after Beijing imposed a sweeping national security law on the territory in June 2020. The route was designed as a moral response to a foreign policy crisis; it has brought over 150,000 people. The second was Ukraine: after Russia's invasion in February 2022, the UK opened humanitarian schemes that brought several hundred thousand more. Neither of these was a labour market decision. Both ran on top of the points-based system rather than through it.

The third was where the new system itself was deliberately loosened. The 2021 rules lowered the salary and skill thresholds for the Skilled Worker visa from previous Home Office benchmarks, abolished the cap on annual numbers, and removed the requirement for employers to advertise locally before sponsoring a foreign worker. The Health and Care Worker visa, opened in August 2020, was then expanded in February 2022 to include lower-paid social-care roles after acute NHS and care-home staffing shortages — many of which had themselves been worsened by EU workers leaving during the pandemic. The Graduate route, opened in July 2021, allowed any international student finishing a UK degree to stay for two years (three for doctorates). Until early 2024, most postgraduate students could bring family with them, which is why work and study together drove most of the surge.

When the political cost of those numbers became unmanageable, the rules tightened. In 2024 and 2025, dependants on most study and care routes were restricted, salary thresholds were raised, and in May 2025 the care-worker route was closed to new overseas recruitment. Net migration fell to 171,000 by the end of 2025. The Migration Advisory Committee — the government's own expert panel — has said that without further policy changes, long-run net migration is likely to settle around 300,000 per year (Migration Advisory Committee, May 2025).

What this did to the economy is harder to see but more important. NIESR estimated in early 2023 that the UK labour market was short of roughly 330,000 workers as a direct result of Brexit, concentrated in lower-paid sectors that had relied on European workers — social care, hospitality, agriculture, food processing, warehousing, and the heavy goods drivers whose 2021 absence caused the petrol shortages people queued for. Skilled European workers had previously been able to take a UK job on the same day they decided to. Now they need an employer sponsorship licence, a visa fee, a health surcharge and a several-month wait. The friction is small at the margin and large at the aggregate.

The clearest answer to "what was Brexit actually for?" is the rewrite of how British insurance companies are allowed to invest. Insurers — the firms that pay annuity pensions and life-insurance contracts — sit on enormous pools of long-term money. The previous EU rulebook required them to hold that money in a tightly defined set of safe assets, mostly government bonds. The new UK regime, called Solvency UK, was designed to let insurers put more of that money into the things the country needs to fund: infrastructure, housing, utilities and productive private companies.

The Bank of England's prudential arm, which supervises insurers, finalised the new framework in two stages. In June 2024, it widened the range of assets insurers could use to back annuity promises, with a cap on the riskier categories and a personal sign-off — called an attestation — by senior executives. Then in October 2025, the regulator added a fast-track approval system so that firms with existing permission could deploy new assets quickly, with twenty-four months to clean up the paperwork. This is sovereignty as a working tool, not a flag in a window.

The caution is in the numbers. The insurance industry's own trade body says its members put £10.9bn into UK productive assets in 2024 — £3.8bn into property, £2.7bn into utilities, £1bn into transport and £3.4bn into other sectors (Association of British Insurers, Industry pledge progress report, Summer 2025). That is real deployment. But the same report carries an honest caveat: the direct impact of the rule change cannot yet be quantified, because the reforms are bedding in and investment decisions have many drivers. The three largest UK annuity writers — Phoenix, Legal & General and Aviva — are putting some released capital into productive assets and some into other balance-sheet uses. The mix matters more than the headline pledge.

The May 2025 Mansion House Accord was a parallel push on workplace pensions: seventeen pension providers, covering roughly 90 per cent of active UK savers in defined-contribution schemes, pledged to put 10 per cent of their portfolios into private and productive assets by 2030, with at least half of that ringfenced for the UK. The Treasury estimates that could unlock up to £50bn, of which £25bn would land directly in the UK economy. The policy is serious. The payoff depends on the pipeline of investable projects, on planning permissions, on infrastructure delivery and on regulators who can move at the speed of capital without pretending risk has gone away.

The most revealing sovereignty choices are the ones Britain did not make. On banking rules, the UK had freedom to set its own capital standards but ended up broadly aligned with the EU's new framework, because banks compete globally and being an outlier hurts. The implementation timetable then slipped to 2027 anyway. On product safety, Britain was supposed to replace the EU's "CE" safety mark with its own domestic version. After years of business pushback and quiet extensions, the government accepted in August 2023 that it would recognise the EU mark indefinitely; the British alternative effectively shelved.

On data protection, the UK has the legal power to diverge from EU rules, but the EU's "adequacy" decision — its judgement that British data law is good enough to allow data to flow freely from Europe to Britain — is worth more than divergence. The Data (Use and Access) Act 2025 was deliberately written narrowly to preserve that. On working hours, ministers under previous governments floated rewriting EU-derived employment law more sharply. Employers preferred stability. On state subsidies, the UK replaced the EU's state-aid rules with its own framework, but the trade deal itself still contains rules limiting what either side can subsidise.

This is not failure. It is the discovery that sovereignty is not the same as maximum difference. It is the capacity to choose difference where the benefit clearly beats the cost. Britain can — and does — diverge in insurance, in listings, in farm subsidies, in gene-editing rules, in public procurement and in parts of data law. Each act of divergence consumes regulatory bandwidth and creates a second compliance question for firms that also trade with Europe. Choosing not to diverge is not evidence that Brexit was an illusion. It is evidence that autonomy has a price tag, and that a serious government uses the pen sparingly.

The most truthful thing to say about Brexit ten years on is that both sides got something they recognised, and both sides got something they did not expect.

Brexiteers were right that sovereignty was real — the UK now writes its own immigration rules, designs its own insurance regime, runs its own subsidy system, and chooses whether to track EU regulations or not. They were also right that the catastrophe forecasts of 2016 did not materialise. The UK did not enter immediate recession. London did not collapse as a financial centre. Goods trade with the EU did not disappear. The economy carried on functioning.

Remainers were right that the bill was real, and larger than the campaign acknowledged. The country produces less than it would have done. The trade hit to goods is structural rather than temporary. Smaller and less-resourced firms have absorbed more of the cost than their larger competitors. The migration promise of lower numbers was delivered only when the government reversed parts of its own visa policy in 2024 and 2025. The City survived by building a parallel office in Frankfurt and Paris.

Both sides also got what they did not bargain for. The "control" the Brexit campaign promised turned out to require ongoing political effort and Whitehall capacity to operate. The "softer Brexit" of an embedded trading relationship that some Remainers hoped for turned out to require accepting the same external rules that Brexit was supposed to escape. The 2025–2026 reset is the most honest part of the settlement: both sides negotiating, item by item, the price of friction reduction.

The country did take back control. What it bought with that control turned out to be a series of operating decisions — and the line that was missing from the side of the bus was the operating budget.