Politics & media · Brexit: what changed

What was promised, and what happened

Campaign claims checked against their wording, timeframe and evidence — with observed outcomes separated from modelled alternatives.

View shareable visual summariesBrexit sectionLeave claimsRemain claims

Five rules this database follows

  1. 1

    Both sides, same treatment

    Claims from Leave and Remain are checked against their original wording, timeframe and the available evidence.

  2. 2

    The claim as made, not as remembered

    Recorded in the wording actually used. The bus did not say “we will spend £350m a week on the NHS”, and the difference is the whole argument.

  3. 3

    “Not yet measurable” is a real verdict

    A prediction with a 2030 horizon cannot be marked wrong in 2026. Forcing every claim into right or wrong would be the fastest way to make this dishonest.

  4. 4

    Some claims were never testable

    “Britain will thrive outside the EU” has no failure condition. It is recorded and labelled, not scored, because scoring it would be scoring a mood.

  5. 5

    Counterfactuals are flagged

    Where a verdict depends on comparing with a world that did not happen, the entry carries a counterfactual marker rather than pretending the comparison was observed.

What each verdict means

  • SupportedThe thing claimed has happened, on the measure the claim itself used.
  • Partly supportedPart of the claim holds and part does not, or it holds on one measure and not another.
  • Not supportedThe thing claimed has not happened, measured as the claim framed it.
  • Not yet measurableThe horizon has not arrived, or no measurement of the right thing exists yet.
  • Not falsifiableThe claim has no condition under which it could be shown wrong. Recorded, not scored.

Where the claims land

Leave claims (8)

  • Supported1
  • Partly supported3
  • Not supported3
  • Not falsifiable1

Remain claims (8)

  • Supported3
  • Partly supported1
  • Not supported2
  • Not yet measurable2

Do not read these as a score. This is a selection, not a census. It covers the campaign propositions that were most prominent in the coverage, most specific, or most often cited since — and prominence is a judgement. It is not a scoreboard: counting verdicts and declaring a winner would treat "Turkey is joining the EU" and "sterling will fall 12%" as equal units, which they are not. Read the entries, not the totals.

Leave campaign claims

LeaveEU budgetPartly supported
“We send the EU £350 million a week — let’s fund our NHS instead.”

Vote Leave (the designated lead campaign) · Throughout the campaign, April–June 2016

Source of the claim: Vote Leave campaign bus and campaign literature; recorded in KCL’s content analysis, which found the £350m argument in 147 articles and the NHS version of it in 77 ↗

What happened

The £350m figure was the UK’s gross theoretical contribution before the rebate, which was deducted before any money was sent, and before EU spending in the UK. The UK Statistics Authority publicly criticised its continued use as misleading. The UK did stop paying contributions, and NHS England funding did subsequently rise substantially — but the 2018 announcement of £20.5bn a year more for the NHS was funded by tax and borrowing, and the IFS found there was no Brexit dividend available to pay for it.

Source: UK Statistics Authority correspondence; Institute for Fiscal Studies analysis of the 2018 NHS funding settlement

The honest caveat. The wording is doing a lot of work. "We send" was not true of £350m; "let’s fund our NHS instead" was a suggestion, not a pledge, and Vote Leave was not a government able to make one. Both the claim that it was a lie and the claim that it was honestly delivered overstate their case.

Partly supported: Part of the claim holds and part does not, or it holds on one measure and not another.

LeaveTradeSupported
Outside the EU the UK would be free to negotiate its own trade agreements with the rest of the world.

Vote Leave · 2016

Source of the claim: Vote Leave campaign material

What happened

Unambiguously delivered as a matter of legal capability. The UK has an independent trade policy, has signed new agreements including with Australia and New Zealand, and acceded to the CPTPP. Their legal existence does not establish the size of any economic benefit.

Source: Department for Business and Trade, UK trade agreements ↗

The honest caveat. Capability and benefit are different claims. This claim was about capability, and on that basis it holds.

Supported: The thing claimed has happened, on the measure the claim itself used.

LeaveImmigrationPartly supported
“Take back control of our borders” — ending free movement so that Britain decides who comes.

Vote Leave · 2016

Source of the claim: Vote Leave campaign material; the campaign’s central slogan

What happened

EU free movement ended for new arrivals, with continuing Irish and Withdrawal Agreement rights. EU+ net migration was −42k in the year to December 2025. Non-EU+ migration rose and then fell under later visa policies. These observations do not establish that all voters understood control to mean a particular numerical target.

Source: ONS, Long-term international migration, provisional: year ending December 2025 (published 21 May 2026) ↗

The honest caveat. The gap between the two readings is the whole immigration argument after Brexit. Control was delivered and used to admit more people, not fewer — a policy choice by successive governments, not a failure of the mechanism.

Partly supported: Part of the claim holds and part does not, or it holds on one measure and not another.

LeaveImmigrationNot supported
“Turkey (population 76 million) is joining the EU” — with the implication of free movement rights for Turkish citizens.

Vote Leave poster campaign · May–June 2016

Source of the claim: Vote Leave poster and campaign material

What happened

Turkey has not joined the EU. Accession negotiations, formally opened in 2005, were effectively frozen by 2018 and remain so. No date for Turkish accession existed in 2016 and none exists now.

Source: European Commission enlargement policy

Not supported: The thing claimed has not happened, measured as the claim framed it.

LeaveEconomyNot supported
“There will be no downside to Brexit, only a considerable upside.”

David Davis MP, later Secretary of State for Exiting the European Union · 2016

Source of the claim: Public remarks during and after the campaign

What happened

Stated as an absolute — no downside at all — it is contradicted by the government’s own subsequent forecaster. The OBR assumes the post-Brexit trading relationship reduces long-run productivity by 4% relative to remaining. Whatever the upsides, "no downside" is not a description of the official position.

Source: Office for Budget Responsibility, Brexit analysis ↗

Not supported: The thing claimed has not happened, measured as the claim framed it.

LeaveTradeNot supported
A UK–EU deal could be negotiated quickly and easily, because the UK held the stronger hand.

Multiple Leave campaigners, including Michael Gove’s “we hold all the cards” · 2016

Source of the claim: Campaign speeches and interviews

What happened

Four and a half years elapsed between the referendum and the Trade and Cooperation Agreement taking effect. In between: three extensions to Article 50, three parliamentary defeats of the Withdrawal Agreement, two changes of Prime Minister and a general election. The Northern Ireland arrangements then required renegotiation in 2023.

Source: Trade and Cooperation Agreement, in force 1 January 2021; Windsor Framework, February 2023

Not supported: The thing claimed has not happened, measured as the claim framed it.

LeaveTradePartly supported
Leaving the Common Fisheries Policy would return control of UK waters and benefit the UK fishing industry.

Vote Leave and fishing industry campaigners · 2016

Source of the claim: Campaign material

What happened

The UK left the Common Fisheries Policy and now negotiates annual quotas as an independent coastal state, with a phased increase in UK quota shares under the TCA. But EU vessels retained substantial access during the adjustment period, the quota gains were far smaller than the industry had campaigned for, and exporters gained new certification and border costs on a trade that goes overwhelmingly to the EU.

Source: Trade and Cooperation Agreement fisheries provisions; Defra fisheries management plans ↗

The honest caveat. The June 2025 fisheries decision extends reciprocal access to 30 June 2038. These legal and administrative changes do not establish a net economic effect for the whole industry.

Partly supported: Part of the claim holds and part does not, or it holds on one measure and not another.

LeaveEconomyNot falsifiable
Britain will “prosper” and “thrive” outside the European Union.

Multiple Leave campaigners — recorded by KCL as appearing in 150 articles · 2016

Source of the claim: KCL content analysis, Table 8 ↗

What happened

No threshold was attached, so no outcome could disconfirm it. Recorded here because it was one of the most frequently made economic arguments of the campaign, and because unfalsifiable claims doing heavy campaign work is itself a finding.

Source: Not applicable

Not falsifiable: The claim has no condition under which it could be shown wrong. Recorded, not scored.

Remain campaign claims

This half of the page is the one usually missing. The Treasury’s immediate-impact analysis predicted a recession and 500,000 job losses that did not happen, and the damage that did to the credibility of official economic forecasting is still being paid for in every argument about the OBR’s current estimates.

RemainEconomyNot supported
A vote to leave would tip the UK into a year-long recession, with around 500,000 jobs lost and GDP around 3.6% lower after two years, in the Treasury’s “shock” scenario.

HM Treasury, under Chancellor George Osborne · 23 May 2016

Source of the claim: HM Treasury analysis: the immediate economic impact of leaving the EU (Cm 9292) ↗

What happened

The predicted immediate recession did not occur: annual real GDP grew in 2016, 2017 and 2018. That tests the recession prediction, but observed growth alone cannot test how much higher GDP would have been in the model’s remain baseline.

Source: ONS national accounts and labour market statistics, 2016–2018

The honest caveat. The Treasury’s analysis was conditional on its policy and uncertainty assumptions. The actual exit timetable and monetary-policy response differed. Its recession forecast and its estimated difference from a remain baseline are distinct claims.

Not supported: The thing claimed has not happened, measured as the claim framed it.

RemainEconomy CounterfactualNot yet measurable
Leaving the EU would leave UK households £4,300 a year worse off by 2030.

HM Treasury, under Chancellor George Osborne · 18 April 2016

Source of the claim: HM Treasury analysis: the long-term economic impact of EU membership and the alternatives. KCL recorded it in 365 articles — more coverage than the £350m claim received ↗

What happened

The stated horizon is 2030, which has not arrived. The direction of the underlying estimate is broadly consistent with the OBR’s current assumption of a 4% long-run productivity reduction, but the two are different models measuring different things over different periods.

Source: OBR Brexit analysis, updated 20 July 2026 ↗

The honest caveat. The presentation was criticised at the time and the criticism was fair: the figure was GDP loss divided by number of households, which is not the same as households being £4,300 poorer in their own pockets. A defensible number was framed in an indefensible way.

Not yet measurable: The horizon has not arrived, or no measurement of the right thing exists yet.

RemainEconomySupported
Sterling would fall sharply — by around 12% in the Treasury’s shock scenario.

HM Treasury; also warned by the Bank of England · 23 May 2016

Source of the claim: HM Treasury analysis: the immediate economic impact of leaving the EU

What happened

The Bank of England’s August 2016 Inflation Report recorded a 9% fall in sterling’s effective exchange-rate index since 23 June. This supports the predicted direction, but is not an exact test of the Treasury’s 12% scenario over its full horizon.

Source: Bank of England, Inflation Report August 2016, Section 1 ↗

Supported: The thing claimed has happened, on the measure the claim itself used.

RemainEconomyNot supported
An emergency budget would be needed within weeks of a Leave vote, raising taxes and cutting spending.

George Osborne, Chancellor of the Exchequer · 15 June 2016

Source of the claim: Public statements during the campaign

What happened

The proposed emergency budget was not delivered after the vote. Osborne left office in July 2016; the subsequent fiscal policy differed from the campaign proposal.

Source: HM Treasury; no emergency budget was delivered in 2016

Not supported: The thing claimed has not happened, measured as the claim framed it.

RemainTrade CounterfactualSupported
Leaving the single market and customs union would create new barriers and reduce UK–EU trade.

Remain campaign, HM Treasury, and a broad range of economic institutions · 2016

Source of the claim: Campaign material and official analysis

What happened

UK goods exports to the EU changed −14% between 2019 and 2025 in the stored ONS volume series, excluding precious metals. New customs and regulatory processes are documented; the OBR’s 15% long-run trade-intensity reduction remains a model assumption.

Source: House of Commons Library, Statistics on UK trade with the EU; OBR Brexit analysis ↗

The honest caveat. Goods exports to non-EU countries changed −8%, while services exports to all partners changed +26%. The EU goods series has a January 2021 recording break. Services are not split by partner here because the comparable split is in current prices, not volumes.

Supported: The thing claimed has happened, on the measure the claim itself used.

RemainNorthern IrelandSupported
Leaving would create serious problems for the Irish border and the Good Friday Agreement settlement.

Remain campaigners, including former Prime Ministers John Major and Tony Blair · 2016

Source of the claim: Campaign speeches, June 2016

What happened

The Irish border was the central obstacle to the Withdrawal Agreement and caused its rejection three times. It produced the Northern Ireland Protocol, a trade border in the Irish Sea, the collapse of power-sharing at Stormont for two years, and a renegotiation in the form of the Windsor Framework in 2023.

Source: Windsor Framework; Northern Ireland Assembly restoration, February 2024 ↗

Supported: The thing claimed has happened, on the measure the claim itself used.

RemainEconomy CounterfactualPartly supported
Business investment would be damaged by uncertainty and reduced market access.

Remain campaign and multiple economic institutions · 2016

Source of the claim: Campaign material and official analysis

What happened

Business investment stagnated between the referendum and the pandemic, in a period when it was growing in comparable economies, and the OBR judges that around two-fifths of its estimated 4% productivity effect had already occurred before the TCA even came into force — through uncertainty weighing on investment. Investment has since recovered; the linked Economy page computes the latest position from the current ONS series.

Source: ONS Business investment (current revised series); OBR Brexit analysis ↗

Partly supported: Part of the claim holds and part does not, or it holds on one measure and not another.

RemainEconomy CounterfactualNot yet measurable
Brexit would reduce house prices — in the Treasury’s shock scenario, by around 10%.

HM Treasury; recorded by KCL as an argument appearing in 216 articles · May 2016

Source of the claim: HM Treasury analysis: the immediate economic impact of leaving the EU

What happened

The Treasury’s 10% was a modelled difference from a remain baseline after two years, not a simple forecast of a 10% fall from the pre-vote price level. Rising observed prices alone cannot disprove that counterfactual estimate; this record has no matched causal estimate to validate its size.

Source: ONS/HM Land Registry UK House Price Index

The honest caveat. The comparison is with an unobserved alternative path, not with the starting price. The original two-year horizon has passed, but that does not make the alternative path observable.

Not yet measurable: The horizon has not arrived, or no measurement of the right thing exists yet.

How these claims were selected

Claims were selected where three conditions held: the claim was made by a designated campaign, a government department or a senior named figure; it was recorded in a citable contemporaneous source; and it was specific enough that some evidence could bear on it. Claims meeting the first two but not the third are included and marked "not falsifiable" rather than dropped, because leaving them out would flatter both campaigns.

This page follows the same editorial rules as the site’s fact-check methodology: verdicts are never silently revised, and where a figure is disputed the dispute is shown rather than resolved by assertion.

Next: the economic evidence behind several of these verdicts, and the scoreboard.

Reviewed 11 October 2026. Trade and migration figures use the current stored ONS series. The house-price entry now distinguishes a modelled shortfall from an observed fall. Original scenario: HM Treasury’s May 2016 analysis. Fisheries: June 2025 access decision.