Brexit at Ten: How Much Has Britain Actually Lost?

Seven Prime Ministers in Ten Years

The referendum took place on June 23, 2016, and this past June marked exactly ten years since then. During that tenth-anniversary week, Prime Minister Keir Starmer announced his resignation, and on July 20, Andy Burnham was sworn in as the seventh prime minister since the referendum. Looking at just that scene, you’d be forgiven for saying “Britain has collapsed.”

But to actually make that claim, you need to split it into two questions. One is how much worse things have gotten; the other is whether that’s because of Brexit. With a decade’s worth of data now in hand, the first question can be answered fairly concretely. The second is the kind of question that no amount of data ever answers cleanly. I want to get into why that is, too.

A benchmark map of the US economy | MercorThe US Benchmark Map places AI benchmarks across the American economy, measuring which parts of the labor force are covered by unsaturated, public, agentic benchmarks.mercor.com

How Much: Four Sets of Numbers

Growth. The most widely cited figure comes from a November 2025 NBER working paper. Written by Stanford’s Nick Bloom and researchers at the Bank of England, the paper estimates that UK GDP per capita in 2025 is 6-8% lower than it would have been without Brexit. UK Chancellor Rachel Reeves cited that 8% figure in a March 2026 speech, and ever since, “8%” has circulated as though it were settled fact. Goldman Sachs ran its own model in June 2026 and arrived at about 6%, noting that this was up from its estimate of 5% two years earlier. John Springford of the CER (Centre for European Reform), who first developed this methodology, put the figure at around 5% in 2025. The UK’s Office for Budget Responsibility (OBR)¹ has built a “4% long-run productivity hit” into its official forecasts continuously since 2020.

Tally it all up and the range is 4-8%. What matters is that these numbers don’t represent how much the UK economy actually shrank — they represent the gap between the UK as it is and the UK that would have existed without Brexit. In reality, the UK economy has grown by about 12% since 2016, outpacing Japan, Germany, Italy, and France. But because the population has grown quickly, per-capita growth over the decade comes to just over 4%. “Stagnant” fits the per-capita numbers better than “collapsed.”

Trade. This is closer to measured reality than estimation. The Trade and Cooperation Agreement (TCA)², which took effect in 2021, eliminated most tariffs, but leaving the single market created new non-tariff barriers³ — customs procedures, certification requirements, regulatory divergence. A gravity-model analysis the CER published in June 2026 finds goods exports to the EU are 16% lower than they would be had the UK remained, and goods imports 14% lower. Services exports are down 7%, and services imports are down 19%. ONS actuals tell the same story: 2025 goods exports to the EU were £34 billion (15%) lower than in 2018. Even so, the EU still accounts for over 40% of UK trade, and the UK remains the world’s second-largest services exporter after the United States. Services holding up is one reason the UK hasn’t fully buckled.

Investment. This is where the clearest wound is. By ONS figures, business investment in Q1 2026 was only 11% higher than in Q1 2016, while over the same period US business investment rose 52%. The UK’s overall investment rate sits at 18.9% of GDP, the lowest in the G7. The Bloom paper estimates business investment is 12-18% below its counterfactual path. Businesses freezing investment right after the referendum was the first and largest effect to show up.

Productivity. OBR says 4%; the Bloom paper says 3-4%. When investment stalls, the equipment and systems each worker has to work with stop expanding, so productivity stalls too. The Bloom team adds two more factors on top of this. Management spent time dealing with Brexit instead of other work, and the most productive firms — the ones that traded heavily with the EU — took the biggest hit, which worsened resource allocation across the economy as a whole.

Is It Really Brexit? Where the Counterarguments Push Back

Most of these estimates come from a doppelgänger model⁴. You blend together countries that moved similarly to the UK before 2016 to construct a “counterfactual Britain without Brexit,” then read the gap with the real UK as the Brexit effect. The method assumes that Brexit is the only thing that set the UK apart from these other countries after 2016. The counterarguments attack exactly this assumption.

Julian Jessop, former chief economist at the IEA (Institute of Economic Affairs), took issue with the composition of the comparison group as early as May 2026. The United States, which carries the heaviest weight in that group, is an outlier case: cheap energy thanks to shale, plus massive fiscal stimulus and an AI boom on top. Italy, Spain, and Greece were still clawing their way back from the eurozone crisis of the early 2010s, so their numbers are muddied by catch-up growth, while Ireland’s GDP figures are unusable because of multinational corporate accounting quirks. Once you set the UK side by side with individual countries instead of an average, its per-capita GDP looks comparable to France’s and slightly better than Germany’s. The actual G7 laggard turns out to be Canada, not Britain. Even Bloom’s own research team wrote in their paper that the assumption that “the UK would have grown at the same rate as the average of other countries” could be wrong, and that Covid and the energy shock hit different countries differently.

And indeed, plenty happened to the UK after 2016 besides Brexit. GDP shrank 10.3% in 2020. In October 2022, inflation hit 11.1%, the highest in 41 years. The OECD judged that the UK took the hardest hit from that year’s energy crisis among G7 countries. That same September, Liz Truss’s government’s tax-cut plan rattled the bond market so badly that the prime minister was out after just 49 days. The fact that seven prime ministers cycled through in ten years is itself evidence that policy kept shifting under everyone’s feet.

Go back further and a bigger picture emerges. UK productivity growth averaged 2.3% a year from 1974 to 2008, then fell to around 0.5% from the 2008 financial crisis through 2020. That’s eight years before the referendum. Investment rates have sat near the bottom of the G7 every year since 2000, and Oxford Economics calculated that if the UK had invested at the G7 average, it would have had an extra £1,900,000,000,000 accumulated over 25 years. Research consistently confirms that austerity, which began in 2010, cut into public investment as well. In other words, Britain’s “chronic under-investment” wasn’t created by Brexit — Brexit was layered on top of it.

Briefings for Britain, a pro-Brexit outlet, goes a step further. It argues that UK productivity has risen a cumulative 11% since 2015 — twice the eurozone’s rate — so there’s no evidence of a “Brexit productivity hit.” Given that the site has backed Leave from the start, its claims can’t be taken at face value. But the underlying point — that evidence is weak for productivity growth bending below its 2007–2015 trend after 2016 — hasn’t been cleanly refuted even by those defending the OBR’s 4% figure.

So how does the other side respond?

The answer to the doppelgänger critique comes from micro data. Alongside the macro comparison, Bloom’s research team used the Bank of England’s firm-level panel survey. They compared firms that had high EU exposure before 2016 — in terms of EU export share, EU import dependence, share of EU-born workers, and EU regulatory exposure — against firms with low exposure. Before the referendum, firms with higher EU exposure grew faster; after the vote, that relationship flipped. Even controlling for COVID effects, the gaps show up as 12% in investment, 3–4% in employment, and 3–4% in within-firm productivity. Because this approach isn’t “comparing one country to another” but “comparing more-exposed firms to less-exposed firms within the same country,” it screens out shared shocks like COVID, the energy crisis, or changes in prime minister.

Jessop pushes back on this too. He argues that the underperformance of EU-exposed firms might be capturing weakness in the European economy itself — Germany’s, for instance — and Britain’s high energy costs, rather than Brexit per se. At this point, both sides are attacking each other’s identification strategy, and all a reader can really do is judge whose assumptions are less of a stretch.

As I see it, the common ground between the two sides is wider than it looks. Jessop, too, concedes that uncertainty and friction caused early damage to the labor market, investment, and trade. The disagreement is over what happened next: he thinks that damage was “temporary and fading,” while Bloom’s team thinks it “accumulated slowly and grew larger over ten years.” In the Bloom team’s own framing, economists’ 2016 average forecast of a “4% long-run loss” turned out to be accurate at the 5-year mark (4–6%) — but overly optimistic at the 10-year mark (6–8%). In other words, the economics was largely right; it was politics’ timetable for how long the effects would drag on that was wrong.

Where Did Brexit’s Promises Go?

Setting aside the costs, look at what leaving the EU was supposed to deliver, and you see another side to the phrase “because of Brexit.”

Immigration hasn’t fallen. Net migration in the 12 months before the referendum was 321,000; by the 12 months to June 2023 it had hit an all-time high of 906,000, and it was still 728,000 in 2024. EU arrivals plunged, but the new work-visa system filled the gap with migrants from India, Nigeria, and Pakistan — by 2024, 86% of immigrants came from outside the EU, up from 42% in 2019.

The OBR classifies the effect of new trade deals as having “no material impact on the forecast.” The deal with Australia adds just 0.1% to GDP over 15 years. The agreement with Japan is hard to count as a net gain at all, since if the UK had stayed in the EU it would have been covered by the EU-Japan deal anyway.

So what the current government is doing now is reversal. At the UK-EU summit in May 2025, the two sides agreed to negotiate a sanitary and phytosanitary (SPS)⁵ agreement, a link between emissions trading systems, and a youth mobility scheme, and in December they agreed to rejoin Erasmus+. Reports from late September 2026 say all three have taken shape, with a finalizing summit set for Brussels on November 20. The SPS agreement is targeted to take effect by mid-2027. Still, the Starmer government has drawn a line against rejoining the single market or customs union, and there’s no confirmation yet on whether a Burnham government would move that line.

How much this reversal actually restores is itself contested. In 2024, the Centre for European Reform (CER) estimated that even combining the SPS agreement with eased movement for performers would add only about 0.1% to GDP a decade from now — because even a 22% jump in agri-food exports barely moves the needle on an economy where that sector is a small slice. Goldman Sachs argues that alignment at single-market or customs-union level could reverse a substantial share of the costs — but that option isn’t on the table right now.

Oswarld’s Lens

I think the question “is it because of Brexit?” is actually a question about measurement. That 4-8% figure gets captured first in things that are easy to measure—goods trade and corporate investment, where the statistics are clean. What’s hard to measure—where ten years of executive meeting time actually went, which investments got “put on hold for now” and never came back—gets treated by the model as a residual. So both sides end up refuting each other with whatever they can measure, and filling in the unmeasurable parts with their own assumptions.

This week’s Mercor map of the US labor benchmark landscape struck me as the same pattern. They mapped 29 public AI benchmarks against 923 US occupations and 18,797 tasks, and found that only 208 occupations have 5 or more benchmarks covering them, while 90% of US labor hours go into tasks with almost no benchmark coverage at all. Benchmarks cluster around occupations with large aggregate wages and tasks with high AI exposure. Lawyers’ work gets measured densely, while legal secretaries and court clerks barely get measured at all. So our answer to “how good is AI?” is really an answer to “how good is AI at the things we decided to measure?” The Brexit debate has the same structure. In goods trade, which we measure well, the damage is clear; in the areas we don’t measure, both sides just state their assumptions.

That said, I don’t want to land on “so nobody really knows.” Investment statistics are actual measurements, not assumptions, and 11% below Q1 2016 levels is a low number under any counterfactual you construct. My read is that Brexit didn’t so much wreck Britain as add one more layer of decade-long uncertainty onto an underlying condition that had already stopped investing since 2008. Whether that one layer amounts to 4% or 8% depends on the model’s assumptions—but the numbers themselves show that underlying condition would have been bad with or without Brexit.

Closing

The “how much” has an answer. GDP per capita essentially flat, goods exports down roughly 15%, business investment up 11% over a decade. The “why” has a range. Somewhere between 4% and 8% — and the rest gets split among post-financial-crisis structural shifts, COVID, energy costs, and political instability.

Once the results of the November 20 summit are in, we’ll know how far the UK has decided to roll things back. The number worth revisiting then isn’t a GDP estimate — it’s business investment from 2027 onward.

---

💬 Reader, how do you separate “caused by this” from “layered on top of it” when you’re evaluating the outcome of a decision? Whether it’s at an organization or in your personal life, share your examples in the comments.

📨 If you have a colleague who covers the UK or European markets, or who makes decisions based on macro indicators, pass this one along.

Fixing the “Oswarld's Lens” reference — glossary specifies “Oswarld’s Lens” for this section name.

References & Further Reading

Primary sources

• Nicholas Bloom et al., “Brexit’s slow-burn hit to the UK economy”, VoxEU, 2025.12.5. ··· This is the original source for the 6–8% estimate. It walks through both the macro-comparison and firm-panel methods, and is candid about its own limitations. If you only read one thing, read this. The full paper is NBER WP 34459.
• Julian Jessop, “Three reasons why you should never trust a “doppelgänger"", ThinkScotland, 2026.5.7. ··· The most detailed pushback from the skeptic side. It covers the problems with constructing comparison groups and includes country-by-country comparison charts.
• John Springford & Anton Spisak, “The cost of Brexit, ten years on”, CER, 2026.6.18. ··· The source for the decline rates in goods and services trade. The same institution’s 10th-anniversary essay collection contains the £40 billion tax calculation.
• Goldman Sachs Research, “UK: The Economic Cost of Brexit, Ten Years On”, 2026.6.24. ··· The source for the 6% estimate and the 4% floor once the financial sector and the US are excluded.
• Office for Budget Responsibility, “Brexit analysis”, updated 2026.7.20. ··· The official source for the 4% productivity hit, 15% trade hit, and the assessment of new trade-deal effects.
• ONS, “Business investment in the UK: January to March 2026”, 2026. ··· The raw data behind the 18.9% investment rate, the lowest in the G7.

Background

• Oxford Economics, “The UK’s Investment Shortfall”, 2026.9.8. ··· Calculates a cumulative £1.9 trillion investment shortfall since 2000. The place to look if you want to see the UK’s underlying condition predating Brexit.
• NIESR, “Why is UK Productivity Low and How Can It Improve?” ··· The source for the productivity-trend comparison: 2.3% from 1974–2008 versus 0.5% since.
• Briefings for Britain, “New UK data further undermines the ‘Brexit economic damage’ narrative”, 2026.9. ··· A productivity-based counterargument from the pro-Leave side. Read with that stance in mind.
• House of Commons Library, “Debate on the UK applying to join the European Union”, 2026.10. ··· Lays out the progress of the “reset” negotiations from the May 2025 summit through the November 2026 summit.
• Mercor, “A benchmark map of the US economy”, 2026. ··· The source used as an analogy in the Oswarld’s Lens section. This is where the figures of 923 occupations, 29 benchmarks, and 90% of labor hours going unmeasured come from.

---

The author is Oswarld (Kwangseob Ahn). Current roles: Adjunct Professor at Sejong University, Strategy Consultant at INLEVEL9. Career, research, books, and recent work are kept current on the About page. Latest · July 2026: HEMA-2: A Consolidation-Aware Tri-Memory Architecture with Multi-Channel Scheduling for Lifelong Conversational AI.

📝 Glossary

Footnotes

1. OBR (Office for Budget Responsibility): The UK’s independent body for producing the economic forecasts that underpin government budget assumptions. Since 2020, it has built a long-run 4% productivity hit and 15% trade hit from Brexit into its official forecasts. ↩

2. TCA (Trade and Cooperation Agreement): The UK-EU agreement that took effect in January 2021. It eliminated most tariffs and quotas on goods, but because the UK sits outside the single market and customs union, customs procedures and regulatory certification became necessary again. ↩

3. Non-Tariff Barriers: Things that make trade harder without being tariffs — customs paperwork, certificates of origin, product certification, inspections, and regulatory divergence. Costs arise even when the tariff rate is zero. ↩

4. Doppelgänger model (synthetic control): A method that builds a weighted average of countries whose trajectories moved similarly to the subject country before some event, in order to construct a “what if this event never happened” counterfactual. Results vary enormously depending on how the comparison group is built, which is exactly why that choice is the crux of the debate. ↩

5. SPS (Sanitary and Phytosanitary): The inspection and certification rules governing trade in food, animals, and plants. If the UK agrees to align with EU rules, most border inspections and paperwork for agri-food products would disappear. ↩