Post Segments
What have the Economist magazine, the Financial Times, the BBC, Bank of England and UK in a Changing Europe got in common? They all agree that Brexit has done substantial damage to the UK economy. Moreover, they go further and assert that this view represents a settled consensus. There is, of course, no consensus. Plenty of economists disagree, although in our intemperate times not all can identify themselves in public.
The widespread belief that that the UK economy is already 4% or even 8% smaller than it would have been in the absence of Brexit rests on much more flimsy foundations than most commentators realise. There are two strands of research supporting the ‘Brexit damage’ thesis. The first strand involves forecasts made before the UK left the EU customs union and single market in 2021, constructed using economists’ forecasting models. The second strand involves analysis of what has actually happened to the UK’s economy since Brexit, now that sufficient data is available. We will examine the problems with both of these strands in turn.
The OBR’s phantom 4% loss of GDP
The most famous prediction of a 4% hit to GDP comes from the Office of Budgetary Responsibility (OBR). Not so well known is the fact that the OBR has made no calculations of its own. Instead, it averaged the results from 13 academic studies conducted using disparate methods, and all completed before the final shape of the UK’s exit agreement was even known. These studies made a wide range of predictions for the damage done by Brexit, from under 2% of GDP to 10%. The fact that the predictions varied so wildly tells us that most of them must be inaccurate. Treating ‘4%’ as a settled fact hides how much the underlying studies actually disagreed. This is an example of what Stanford academic Sam L. Savage called ‘the flaw of averages’ – the idea that averaging a diverse set of numbers somehow gives a reliable estimate, when in fact the underlying variance actually makes the average unrepresentative or misleading.
Moreover, the higher estimates depended on a false assumption about the link between trade and productivity which doubled the impact of Brexit on GDP. Excluding the studies which use this important, but false, assumption leaves a conclusion that any hit to UK GDP would be small.
Ultimately, the famous 4% figure was only a prediction, not fact. It should also be remembered that it is easier for economists to identify what might be lost than it is for them to factor in potential but unidentified future gains from being outside the EU’s regulatory orbit, for example a larger AI industry.
What’s actually happened to the economy since 2016
The second strand of research involves attempts to measure what has actually happened in the UK since leaving the EU’s trade system in 2021. With five years of data, there is now sufficient evidence to observe how the UK is actually doing, and indeed academics have made a raft of attempts to measure Brexit’s impact on the economy.
The data shows that the UK economy has grown as fast as France or Germany since 2016. Anti-Brexit enthusiasts find it “irritating” that so many people point this out. They assert that, since the UK grew faster than these countries before 2016, it would have continued to do so if we had remained inside the EU. But this assertion ignores the fact that faster UK growth was concentrated in the 1990s – when among other things Germany was dealing with the high costs of re-unification, and the UK was benefitting from a variety of factors such as a long boom in financial services and house prices which could not (and did not) go on forever.
Why the doppelganger method does not make sense
All other studies are a variant on this same logic: comparing UK growth to an extrapolated trend and treating the gap as ‘Brexit damage’. The most influential are the ‘doppelganger’ studies which use a computer program to select a group of countries which, in aggregate, mimicked UK economic growth in the years leading up to the 2016 referendum. The selected group can then be used as a supposedly reliable indicator of how well the UK economy would have performed without the shock of Brexit. The most recent of these was a widely publicised working paper published in late 2025 by British authors in the (non-peer-reviewed) series of the American National Bureau of Economic Research (NBER).
The NBER study constructed a counter-factual (doppelganger) data series based on quarterly changes in GDP and other economic variables over the period 2006-2016. The countries included in this composite index reveal the fatuousness of the exercise. They are the USA, Estonia, Finland, Greece Italy, Ireland, Lithuania, and Iceland. Why several of these tiny nations, with very different economic structures to the UK, should be selected as likely indicators of UK performance is a mystery. The inclusion of Greece with its catastrophic recent economic history is frankly absurd. Those who use this technique are following an academic fashion and give little thought to the hapless results which tumble out of the computer.
Moreover, the whole doppelganger approach is rendered even more absurd if we apply it to other countries only lightly impacted by Brexit. If we do this, we find that apparently France’s GDP is 7% lower than it would have been without Brexit, and Germany’s a massive 12% lower – as bad or even worse than in the UK. No one can take these studies seriously in the light of these results (which we will present in more detail in a forthcoming article). What the doppelganger is picking up is a range of post-2016 negative influences on growth that have had an impact across economies – and which have nothing to do with Brexit.
It’s not Brexit – it’s America
In fact, the only country whose historic growth path matches that of the UK to a degree that is statistically significant is the USA. Indeed, the doppelganger index for the UK in the NBER study closely matches the USA alone. This is not surprising, because the USA constitutes 60% of the composite index. In effect, the NBER study is comparing the UK with the USA. The same distortion applies to every other country’s doppelganger figure, since the USA dominates each composite index in the same way.
But the recent economic history of the US is very different to that of the UK. The US economy took off after 2016 under both Trump and Biden, as successive US governments ran huge deficits and allowed fracking, bringing energy costs down to a fraction of UK and EU levels. The AI boom has added another big stimulus to the US economy. As a result, the US has grown much faster since 2016 than any other G7 economy. It is therefore wrong, and ridiculously so, to use the gap between UK and US performance to indicate the impact of Brexit, as the NBER and similar studies do.
Whether you apply the doppelganger method to the UK, France or Germany, what is being measured is the failure of the UK and major EU economies to keep up with the surging growth of the USA. This has little or nothing to do with Brexit, and everything to do with the US-specific factors listed above. The fact that so many commentators have given credence to these studies is an embarrassment. It also damagingly distracts attention from the real causes of our economic difficulties, which are centred on the productivity growth slowdown since the banking crisis in 2007.
The illusion of resemblance
The NBER article has attracted support from many commentators partly because it includes other comparisons alongside its flawed doppelganger analysis. These include weighted and unweighted comparisons of the UK with 33 other European and G7 economies, again on a before-Brexit and after-Brexit basis. Other than the USA, the GDP growth trends of these countries are not significantly similar to the UK prior to 2016. Indeed, some of them are wildly dissimilar. Importantly, some have faster growth than the UK, and some slower. As a result, if they are bundled together, then their mean is not far from the UK trend. However, this resemblance is just a statistical artefact: the pluses and minuses cancel each other out and the combined effect is a growth pattern that resembles the UK’s, even though none of the individual countries do. So the resemblance tells us nothing about comparability — it’s just an arithmetic side-effect of the mix, not evidence the group is a valid stand-in for the UK. It is therefore not something from which any conclusions about the effect of Brexit can be drawn.
The dubious microeconomic evidence
The NBER study also adds a microeconomic analysis alongside the macroeconomic analysis discussed above. This uses a large-scale, regularly conducted survey of UK companies with more than ten employees. The authors divide the sample up into firms more exposed to the EU and those less exposed. The more exposed firms tended to be larger and in manufacturing or retail (for imports). The performance of these two groups is then recorded over a pre-Brexit referendum period (2012-2016) and a post-referendum period up to 2025.
The results showed that the more-exposed firms had significantly larger slowdowns in growth after 2016 for investment and employment, but strangely, not for sales. This matters: if EU exposure were really damaging these firms’ trade, one would expect it to show up in sales first. The absence of a hit to sales suggests the investment and employment slowdowns may reflect caution and uncertainty rather than any actual Brexit-driven loss of business.
Perhaps more seriously, the NBER survey results do not seem to match the performance of the UK economy as a whole. ONS and OECD data for the UK private sector show that employment has grown faster than major EU economies since 2016, adding 2.3 million extra jobs. Business investment is also higher now relative to GDP than it was before 2016, and has grown as fast as France or Germany. If the NBER’s sample were genuinely capturing a Brexit effect, the national picture should look similarly weak. The fact that it doesn’t suggests that the NBER’s sample is unrepresentative, rather than reflecting a real Brexit effect.
A notable weakness in the NBER methodology is that the pre-referendum period was so short that it is likely to have picked up cyclical upturns following the banking crisis – and hence potential cyclical downturns after 2016. These may have affected industrial firms more than service firms, and may potentially have had little to do with Brexit. Interpretation is not helped by statements such as “Since the vote for Brexit in 2016 Q2, UK GDP has grown by less than other comparable countries” – which is simply untrue, as we have shown earlier in this article.
Conclusion: the evidence for a large Brexit ‘hit’ remains flimsy
Overall, the widespread view that Brexit has had a large negative impact on UK growth remains remarkably flimsy. The UK has lagged behind the US since 2016, but then so have most other advanced economies, and there are very good, non-Brexit-related, explanations for this. The popular doppelganger approach, which purports to show that the UK would have grown far faster than it actually did in the absence of Brexit – rivalling the stellar performance of the US – is hopelessly flawed. Its methodology makes little sense and its conclusions are implausible. The only surprising thing is that anyone takes it remotely seriously.
Dr Graham Gudgin is Honorary Research Associate at the Centre for Business Research at Cambridge University. Wilf Glasby is the pen-name of a senior economist working in the private sector in the UK.
