Research library

Published estimates of what leaving the EU changed

14 estimates from 9 published studies, grouped by what they measure. Each study sits next to the critiques and alternative estimates found for it, and each links to its source.

What a counterfactual study is

Nobody can observe the UK economy in a world where it stayed in the EU. A counterfactual study builds an estimate of that world and compares it with what happened. A figure of "6% lower" means 6% below the study's own no-Brexit path, not 6% below some earlier level of GDP.

Estimates differ for reasons that are in the open. Some compare the UK with a weighted basket of other countries (synthetic control); some use firm surveys; some are models of trade and productivity that were run before the exit terms were known; some are assumptions made by forecasters. The studies cover different dates, different measures and different scenarios, and each has its own critics. Pre-exit projections and after-the-fact comparisons answer different questions, so check the horizon on each row before comparing two numbers.

Inclusion criteria

An estimate is in the library when it meets all of these:

It has been published, with a date.
It has a named author or organisation.
It states an estimate, or a clearly stated finding about the size or presence of an effect.
The source is reachable at a link, and the entry was checked against the original where possible (the "Verification" note on each row says what it was checked against).

The distribution of publication dates reflects the wider body of research; the library includes all verifiable estimates that meet our inclusion criteria, whatever effect they find. Pieces that state a finding without a number are listed but not plotted.

The range of estimates, by measure

Each panel has its own axis: % versus the study's own counterfactual, where negative means lower and positive means higher. Estimates for different measures are never drawn on the same axis. A whisker runs from the study's lowest to its highest published figure.

Peer-reviewed paper (circle)
Working paper (square)
Official forecaster (diamond)
Think tank (triangle)
Campaign network (hexagon)
Hollow: midpoint of a published range, no central estimate given

GDP (6 estimates plotted)

Not plotted, no numerical estimate published: Minford and Zhu 2023.

Investment (2 estimates plotted)

Productivity (2 estimates plotted)

Not plotted, no numerical estimate published: Briefings for Britain 2025.

Trade (2 estimates plotted)

Source: studies listed below, last checked 6 November 2026. Counts by source type:peer-reviewed paper 1, working paper 5, official forecaster 2, think tank 4, campaign network 2. The library is not a sample of all work on the subject, and the number of studies on each side of a question says nothing about which is right.

All studies, with critiques and alternatives beside them

"Verification" on each row records the date and what the entry was checked against. Where the original document could not be reached, the row says so and the figures come from a secondary report.

MeasureSource typeYear

Showing all 14 estimates.

Published counterfactual estimates with critiques and alternative estimates beside each
Study	Measure and estimate	Method	Critiques and alternative estimates
The Macro-Economic Impact of Brexit: Using the CBR Macro-Economic Model of the UK Economy (UKMOD) Graham Gudgin, Ken Coutts, Neil Gibson, Jordan Buchanan Centre for Business Research, University of Cambridge, Working Paper no. 483 (with Ulster University Economic Policy Centre) · May 2017 Working paper (not peer-reviewed)	GDP · by 2025 (pre-exit projection) −1.5% to −1% % of GDP vs a pre-referendum forecast, 2025, in the authors' milder (baseline) Brexit scenario, which they call more realistic but still pessimistic. The abstract gives a minor loss 'around 1%' by 2025; the results section gives 'some 1.5% below' the pre-referendum forecast in 2025, so both are shown as a range. Per capita GDP is about unchanged. Their scenario built on Treasury trade assumptions is much worse (abstract: 'nearer 4%'; results section: 6%) and is not shown.	Scenarios run on the CBR macro-economic model (UKMOD), compared with a pre-referendum forecast; trade losses informed by a replicated gravity-model analysis; pre-exit projection Verification Checked 2026-11-06 against: original paper. Read the paper itself: Centre for Business Research (University of Cambridge) Working Paper no. 483, May 2017, a revised version of a working paper first published in December 2016, by Graham Gudgin and Ken Coutts (CBR) and Neil Gibson and Jordan Buchanan (Ulster University Economic Policy Centre). The earlier library figure ('worst case about 1%') came from a blog interview and was wrong: in the paper about 1% is the authors' milder baseline scenario, not a worst case, and the horizon is 2025, not the early 2020s. The paper's abstract and results section differ slightly on both scenarios (about 1% against 1.5%; nearer 4% against 6%), which is why a range is shown. The 2017 blog interview is no longer the linked source. Pre-exit projection, not an outturn measurement.	Critiques Simon Wren-Lewis, Mainly Macro blog. 19 March 2017, replying to a Prime Economics piece by Gudgin, Coutts and Gibson. Concedes an error in a trend figure he used but argues that the authors' assumed upward trend in UK export penetration makes little sense and appears to make the effect of EU membership look small, a point he says they did not answer. Does not address the gravity-model estimates directly. No reply from LSE Centre for Economic Performance authors (for example Thomas Sampson) to Gudgin et al. was found. Alternative estimates Born et al. (2019), GDP 1.7-2.5% lower by end-2018. Outturn estimate for the effect of the vote by end-2018. Bloom et al. (2025), GDP 6-8% lower by 2025. Outturn estimate to 2025.
From Project Fear to Project Prosperity: An Introduction Patrick Minford (Chair, Economists for Free Trade) Economists for Free Trade (formerly Economists for Brexit) · August 2017 Campaign network (not peer-reviewed) Stance as published: Its Chair, Patrick Minford, writes that the group was formed in March 2016 as Economists for Brexit, 'to present what we saw as the economics of the EU, our membership of it and the economics of leaving', and was renamed Economists for Free Trade; it sets out the case for free trade 'properly understood as global free trade and not simply free trade with the rest of the EU' (document introduction).	GDP · long run (pre-exit scenario) central +6.8% % higher GDP under a 'hard Brexit' (free trade outside the Single Market and Customs Union). The document gives about 4% from free trade, 2% from improved regulation (a total of 6%), plus 0.6% from regaining the net EU budget contribution and 0.2% from removing the taxpayer subsidy to unskilled immigration: 6.8% in all. Press and secondary summaries round this to about 7% and about GBP 135bn a year; that rounding does not appear in the document.	Calibrated 'Classical' trade model (Cardiff); pre-exit forecast of a scenario with unilateral free trade, abolition of the CAP and lower regulation Verification Checked 2026-11-06 against: original report (third-party copy). Read a PDF copy hosted on a personal blog (paulbacon.wordpress.com), not the publisher's own site: the Economists for Free Trade website returned a certificate error and its web-archive copy is not reachable from this tool, and neither was worked around, so no official copy was read. The document is the 9-page Introduction, dated 15 August 2017 (from the file name; the text itself carries no date), to the then forthcoming full report, From Project Fear to Project Prosperity, and is by Patrick Minford as Chair. It gives the components of the gain (4% + 2% + 0.6% + 0.2% = 6.8%) but not the figures '7%' or 'GBP 135bn', which come from City A.M. and a Kent Matthews summary (LSE EUROPP, 6 September 2017); the earlier entry showed 7.0 and listed 0.8% for the EU contribution, which the document gives as 0.6%. The full report was not retrieved. A forecast of a policy package that was not adopted, so not a measurement of the UK's actual outcome.	Critiques Economists for Remain (as summarised by Kent Matthews), LSE EUROPP blog. Trade follows a gravity pattern so lost EU trade is not fully replaced; demand constraints and non-tariff barriers matter; devaluation gains are offset by imported inputs. Matthews says the gap comes from different theoretical assumptions and calls for empirical testing of both models. Helena Vieira, summarising Thomas Sampson, Swati Dhingra, Gianmarco Ottaviano and John Van Reenen (LSE Centre for Economic Performance), LSE Business Review. 27 May 2016, on the 2016 Economists for Brexit version of this model. Disputes the assumed 10% fall in trade costs against average UK tariffs of about 3%, the unilateral free-trade scenario, the treatment of EU regulation and the model's neglect of the gravity relationship.
The Costs of Economic Nationalism: Evidence from the Brexit Experiment Benjamin Born, Gernot Muller, Moritz Schularick, Petr Sedlacek The Economic Journal, vol. 129, issue 623, pp. 2722-2744 · October 2019 Peer-reviewed paper	GDP · by end-2018 −2.5% to −1.7% % of UK output vs synthetic UK, end-2018 (the effect of the vote, before any change in trading terms).	Synthetic control on the referendum result as a natural experiment; expectations-augmented VAR for the channels Verification Checked 2026-11-06 against: original paper. Record page read (title, journal, volume, pages, 1.7-2.5%). The month of publication is not given on the record; October 2019 is the issue date of volume 129 issue 623 as commonly cited and should be confirmed. Earlier drafts (2017, 2018) gave smaller figures. Measures the effect of the vote up to 2018, so it is not comparable in horizon with later estimates. Critiques added 2026-11-06: no Gudgin, Coutts or Briefings for Britain piece naming Born et al. was found, so the two critiques listed are of the doppelganger / synthetic comparison method in general and say so.	Critiques Graham Gudgin, Policy Exchange. 17 June 2022. A critique of the CER 'doppelganger' method, a close relative of synthetic control; it does not mention Born et al. Argues that a comparison group attributes all post-2016 divergence to Brexit although cyclical and other causes may contribute, that the comparison countries change between reports, and that Covid-era data distort the comparison. Julian Jessop, Institute of Economic Affairs (press release, 14 September 2018). Written about the 2.1% estimates from UBS, CER and others, not about Born et al. by name. Says the 'doppelganger' method is unreliable over short periods and that the post-vote weakness may be partly temporary or due to other causes such as sterling's fall; puts the likely effect nearer 1%. Alternative estimates Bloom et al. (2025), GDP 6-8% lower by 2025. Longer horizon (to 2025), includes effects after the 2020 exit terms took effect. CER (Dec 2022), GDP 5.5% lower by Q2 2022. Different horizon and comparison basket.
The cost of Brexit to June 2022 John Springford Centre for European Reform (CER) · December 2022 Think tank (not peer-reviewed)	GDP · by Q2 2022 central −5.5% % of real GDP vs the doppelganger UK.	Synthetic control ('doppelganger'): a weighted basket of 22 advanced economies matched to the UK on 2009 to 2016 data, compared with actual UK data Verification Checked 2026-11-06 against: original paper. Author, date (21 December 2022), method and the 5.5% figure confirmed on the CER page. SPEC linked a mailings.cer.eu copy; the cer.eu page was used. Gudgin's critique (17 June 2022) answers the earlier June 2022 update (5.2%), not this December edition, so it is shown as a critique of the method. No CER update after December 2022 was found.	Critiques Graham Gudgin, Policy Exchange. 17 June 2022, on the 5.2% update. Argues the comparison countries change between measures and reports, cyclical and other causes are not separated, and Covid-era data are unreliable; finds UK growth near the G7 average since 2016 on per-head GDP. Julian Jessop, Independent economist (Substack). 3 May 2026. Objects that doppelganger comparisons cannot separate Brexit from other shocks.
Modelling the effects of Brexit on the British economy Patrick Minford, Zheyi Zhu Cardiff Business School, Cardiff Economics Working Paper E2023/19 · July 2023 Working paper (not peer-reviewed)	GDP · short run No numerical estimate published No numerical estimate in the abstract. Describes temporary, slightly negative effects on GDP, exports and imports, and slightly positive effects on inflation and interest rates; the paper also reviews models that find long-run gains.	Structural VAR with Brexit effects identified by event dates (referendum, exit from the single market) Verification Checked 2026-11-06 against: original paper. Only the abstract page was read; it gives direction and relative size, no figures. Listed in the table but not plotted. Full text should be read by hand before any number is quoted. Critiques added 2026-11-06: no published response to this paper itself was found; the UK in a Changing Europe and LSE items listed predate it and concern Minford's earlier modelling, and the City AM item is about regulatory divergence generally. Each says so.	Critiques Iain Begg, UK in a Changing Europe / LSE. 4 May 2016, on the Economists for Brexit analysis, not on the 2023 paper. Says its 4% GDP gain rests on assumptions, including a deregulation gain modelled as a 2% cut in employer National Insurance without explanation, and offers limited evidence. This is the closest UK in a Changing Europe piece found on Minford's modelling assumptions. Helena Vieira, summarising Sampson, Dhingra, Ottaviano and Van Reenen (LSE Centre for Economic Performance), LSE Business Review. 27 May 2016, on the earlier long-run model, not the 2023 paper. Questions the trade-cost, unilateral free-trade and EU-regulation assumptions behind its long-run gains. Paul Johnson and Robert Johnson, City AM (extract from The Brexit Effect 2016-2026, ed. Anthony Seldon, Cambridge University Press). 9 July 2026. Does not discuss Minford and Zhu's modelling. Reports that the UK has so far diverged very little from EU rules, which bears on the regulatory-divergence channel in long-run models of gains. Alternative estimates Bloom et al. (2025), GDP 6-8% lower by 2025. Finds larger and persistent effects over a longer horizon. Born et al. (2019), GDP 1.7-2.5% lower by end-2018. Peer-reviewed estimate of the effect of the vote.
The Economic Impact of Brexit Nicholas Bloom, Philip Bunn, Paul Mizen, Pawel Smietanka, Gregory Thwaites, Sasha Abrahams NBER (National Bureau of Economic Research), working paper 34459 · November 2025 Working paper (not peer-reviewed)	GDP · by 2025 −8% to −6% % of GDP vs no-Brexit counterfactual. The NBER digest reports the micro estimate as 6% and the macro estimate as 8% (GDP per capita).	Macro estimate (UK against 33 comparable countries) and micro estimate (Decision Maker Panel firm survey, firms with higher and lower EU exposure) Verification Checked 2026-11-06 against: original paper. SPEC listed the econpapers record; the NBER page was used instead (issued November 2025, revised June 2026). Range 6-8% confirmed in the full text of the June 2026 revision (macro 8% on GDP per capita, micro 6%). The NBER page lists it as a working paper; no journal publication found.	Critiques Julian Jessop, Independent economist (Substack). 3 May 2026. Argues doppelganger and synthetic comparisons cannot separate Brexit from Covid, the energy shock, US stimulus and other shocks, that catch-up growth in comparison countries flatters the baseline, and that the survey behind the bottom-up estimate may be unrepresentative.
Brexit and the UK Economy Ten Years On: Stocktake and Future Options Benjamin Caswell, Peter Dixon National Institute of Economic and Social Research (NIESR), policy brief · June 2026 Think tank (not peer-reviewed)	GDP · by end-2025 −9% to −4.9% % UK underperformance of GDP relative to the synthetic control by end-2025: 4.9% (22-country panel) to 9.0% (31-country panel). The authors say they do not claim to identify how much is due to Brexit.	Synthetic control: weighted baskets of comparable economies matched to UK GDP growth from 1996 Q1 to 2016 Q2; two panels (31 countries; 22 countries excluding Ireland and similar) Verification Checked 2026-11-06 against: original paper. Read from the NIESR PDF. These are the figures the document states; they are a position at end-2025 relative to a synthetic control, not a stated long-term estimate. The paper cautions that synthetic controls cannot identify the causal effect of Brexit, are sensitive to the period and country panel, and that the Covid-period gap is partly a measurement artefact of how UK public-sector output is counted. It also cites the OBR's 4% and Bloom et al.'s 6-8% without giving a separate NIESR long-run figure. Earlier NIESR model-based projections (Autumn 2023 feature: GDP 2-3% lower three years after the transition, 5-6% lower by 2035) were seen on NIESR's site but not added.	Critiques Graham Gudgin, Policy Exchange. 17 June 2022. A critique of the CER 'doppelganger' method, a close relative of synthetic control, written before and not naming this NIESR paper. Argues that cyclical and other causes are not separated from Brexit and that Covid-era data are unreliable. Alternative estimates Bloom et al. (2025), GDP 6-8% lower by 2025. Macro and firm-survey estimates over the same horizon. CER (Dec 2022), GDP 5.5% lower by Q2 2022. Synthetic-control estimate over an earlier horizon.
The cost of Brexit to June 2022 John Springford Centre for European Reform (CER) · December 2022 Think tank (not peer-reviewed)	Investment · by Q2 2022 central −11% % of investment vs the doppelganger UK.	Synthetic control ('doppelganger') from 22 advanced economies Verification Checked 2026-11-06 against: original paper. Confirmed on the CER page.	Critiques Graham Gudgin, Policy Exchange. 17 June 2022. Critique of the doppelganger baseline (see the GDP entry).
The Economic Impact of Brexit Nicholas Bloom, Philip Bunn, Paul Mizen, Pawel Smietanka, Gregory Thwaites, Sasha Abrahams NBER (National Bureau of Economic Research), working paper 34459 · November 2025 Working paper (not peer-reviewed)	Investment · by 2025 −13% to −12% % of business investment vs no-Brexit counterfactual. Micro estimate 12%, macro estimate 13% (macro: average across five weighting metrics, range 8% to 22%, the largest from the synthetic control).	Macro estimate (UK against 33 comparable countries) and micro estimate (Decision Maker Panel firm survey) Verification Checked 2026-11-06 against: original paper. Corrected on 2026-11-06 from the full text of the latest NBER version (issued November 2025, revised June 2026): the abstract, introduction and conclusion all state investment reduced by 12% to 13% (micro 12%, macro 13%). The earlier 12-18% range (SPEC and an earlier NBER digest) does not appear in the revised paper and is no longer used; some press coverage still repeats it.	Critiques Julian Jessop, Independent economist (Substack). 3 May 2026. Questions how far the survey-based and country-comparison estimates isolate Brexit from other shocks (see the GDP entry).
Brexit analysis: long-run productivity assumption Office for Budget Responsibility Office for Budget Responsibility (OBR) · November 2016 Official forecaster (not peer-reviewed) Stance as published: Independent fiscal watchdog funded by the Treasury.	Productivity · long run (about 15 years) central −4% % lower long-run productivity than if the UK had remained in the EU. A single OBR assumption, not a range.	Assumption in the OBR's forecasts since November 2016: the OBR says its trade-intensity assumption drives it Verification Checked 2026-11-06 against: official document. Wording confirmed in the OBR's Economic and fiscal outlook, October 2021 (Box 2.5), as published on gov.uk: the 15% trade reduction 'drives the 4 per cent reduction in long-run potential productivity we assume will eventually result from our departure from the EU'. The same box says forecasts have assumed this since the first post-referendum forecast in November 2016, so the date is corrected from the March 2021 date inferred earlier. The OBR's own Brexit analysis hub (obr.uk) returns HTTP 403 to automated fetches and was not used. A later check looked for the March 2023 Economic and fiscal outlook on gov.uk (assets.publishing.service.gov.uk) to confirm the 'broadly on track' wording and the 15-year horizon; no gov.uk copy of that edition was found (the gov.uk publication pages and the guessed asset address returned 404), so those two points stay unconfirmed. Search-engine extracts of that hub say the OBR saw no case for altering the 4% view in its 2023 review; that was not confirmed from a primary copy. The 'about 15 years' horizon comes from the same extracts, not from the October 2021 box. A November 2025 report of a separate cut to the OBR's general productivity assumption was seen only in a search summary and is not included.	Critiques The Critic (author not retrieved), The Critic magazine. Argues the 4% figure is an average of 13 earlier external studies rather than OBR modelling, and that most of those studies were below 4%. The article itself returned HTTP 403, so this summary comes from a search-engine extract. Graham Gudgin, Cambridge Centre for Business Research (interview, 31 March 2017). Earlier than the 2021 OBR wording: says OBR GDP forecasts rest mainly on a productivity assumption he considers questionable given weak productivity growth over the preceding decade. Briefings for Britain (no individual author named), Briefings for Britain. 22 November 2025. Says the OBR's 4% long-run productivity hit rests on an average of studies with questionable assumptions and that the link from trade to productivity is weak. Argues post-2016 productivity growth matches the slower pace seen since the financial crisis. Offers no alternative figure. Alternative estimates Bloom et al. (2025), productivity 3-4% lower by 2025. Estimated from outcomes to 2025 rather than assumed in advance; reaches a similar size.
There is no evidence that Brexit has damaged UK productivity growth Briefings for Britain (no individual author named) Briefings for Britain · November 2025 Campaign network (not peer-reviewed) Stance as published: Describes itself as a volunteer group, originally of academics, formed to supply fact-based material to inform public debate on Brexit; says it has no party links and is funded only by supporters. (Briefings for Britain 'About' page, read 2026-11-06.)	Productivity · 2016 to 2025 No numerical estimate published No numerical estimate of a Brexit effect. Stated finding: no measurable Brexit-driven slowdown in UK productivity growth; the post-2016 trend matches the slower pace seen since the financial crisis (the page cites about 2.2% a year before the crisis and about 0.6% after).	Comparison of UK productivity, manufacturing, trade openness and investment trends before and after 2016; no counterfactual model Verification Checked 2026-11-06 against: original paper. Page dated 22 November 2025 as retrieved; the browser title is 'Brexit and Productivity: No Evidence of a Slowdown'. Added as the one 2024-26 piece found that clearly states a small-or-zero-effect finding; it gives no estimate, so it is listed but not plotted. It responds directly to the OBR's 4% productivity assumption. No 2024-26 IEA or Briefings for Britain piece giving a GDP estimate was found, and no author named Harry Western could be found.	Alternative estimates OBR long-run productivity assumption, 4% lower. An assumption, not an outturn estimate; the page above is written in response to it. Bloom et al. (2025), productivity 3-4% lower by 2025. Estimated from outcomes to 2025; reaches a different conclusion. No published critique of this study was found when the library was checked; alternative estimates are listed instead.
The Economic Impact of Brexit Nicholas Bloom, Philip Bunn, Paul Mizen, Pawel Smietanka, Gregory Thwaites, Sasha Abrahams NBER (National Bureau of Economic Research), working paper 34459 · November 2025 Working paper (not peer-reviewed)	Productivity · by 2025 −4% to −3% % of productivity vs no-Brexit counterfactual. The two approaches agree closely (NBER digest).	Macro estimate (UK against 33 comparable countries) and micro estimate (Decision Maker Panel firm survey) Verification Checked 2026-11-06 against: original paper. Confirmed 3-4% on the NBER page. The NBER page also reports employment 3-4% lower, not shown separately here.	Critiques Julian Jessop, Independent economist (Substack). 3 May 2026. General objection to attributing post-2016 gaps to Brexit alone (see the GDP entry).
Brexit analysis: long-run trade intensity assumption Office for Budget Responsibility Office for Budget Responsibility (OBR) · November 2016 Official forecaster (not peer-reviewed) Stance as published: Independent fiscal watchdog funded by the Treasury.	Trade · long run central −15% Trade intensity (trade relative to GDP), not total trade volume: the OBR assumes UK imports and exports will each eventually be 15% lower, relative to GDP, than had the UK stayed in the EU.	Assumption first made in the OBR's November 2016 forecast; the OBR tracks it against outturn trade data in later forecasts Verification Checked 2026-11-06 against: official document. Wording confirmed in the OBR's Economic and fiscal outlook, October 2021 (Box 2.5), on gov.uk: 'total UK imports and exports will eventually both be 15 per cent lower than had we stayed in the EU', described as a reduction in trade intensity; the box says developments were still consistent with that assumption but that it was too early for firm conclusions. The assumption dates from November 2016 (date corrected from March 2021). Search-engine extracts of the OBR's Brexit analysis hub (obr.uk, HTTP 403 to automated fetches) say the OBR later judged it broadly on track; not confirmed from a primary copy: a check for the March 2023 Economic and fiscal outlook on gov.uk found no copy there, and obr.uk was not worked around. This is the Brexit trade assumption, not a trade-deal estimate (trade deals are on /trade-deals).	Critiques Graham Gudgin, Cambridge Centre for Business Research (interview, 31 March 2017). Earlier and about Treasury modelling: argues gravity-model trade estimates averaged across EU members overstate the effect for the UK, which trades less with the EU than most members, so the trade effect should be about half or less. Catherine McBride, Institute of Economic Affairs. 6 November 2023. Argues UK-EU goods trade from 2019 to 2022 shows no clear Brexit effect relative to trade with non-EU countries, and that the OBR's projected trade and productivity losses are not supported by the data. Acknowledges that Brexit adds compliance costs. Alternative estimates CER (Dec 2022), goods trade about 7% lower to Q2 2022. Outturn comparison against a synthetic UK; a different horizon from the OBR's long-run figure.
The cost of Brexit to June 2022 John Springford Centre for European Reform (CER) · December 2022 Think tank (not peer-reviewed)	Trade · by Q2 2022 central −7% % of goods trade vs the doppelganger UK. Services trade was about equal to the doppelganger in this update.	Synthetic control ('doppelganger') from 22 advanced economies Verification Checked 2026-11-06 against: original paper. Goods trade 7% lower, down from 13.6% in the previous update (per the CER page); services roughly equal.	Critiques Graham Gudgin, Policy Exchange. 17 June 2022. Critique of the doppelganger baseline (see the GDP entry).

What this site can't tell you

Studies of trends show correlation, not causation. Covid, energy prices and policy changes overlap the same years, and the methods above differ in how far they try to separate these effects.
Non-economic factors that people weigh, such as sovereignty, regulatory autonomy and migration policy, are not measured here.
Rejoining would not restore the terms of 2016. The euro, the UK rebate and the opt-outs would differ for a new member.