Featured Economy & trade Briefings

Goldman Sachs: Wrong on Brexit.

Written by Catherine McBride

A recent research paper by Goldman Sachs claims that the UK’s economy had significantly underperformed ‘other advanced economies’ because Brexit had reduced trade, lowered business investment and lowered EU immigration. The paper is unconvincing with evidence flimsy at best and some of their trade statistics are incorrect. If anything, the paper simply proves that the US would be a better economic model for the UK, not the EU.

Print Friendly, PDF & Email

Last week Goldman Sachs (GS) produced a paper on the structural and cyclical costs of Brexit which they distributed widely to the media. In their executive summary, they report that their analysis suggests that real UK GDP has fallen short of similar countries by about 5%. This was gleefully reported by media ranging from The Times in the UK to America’s CNBC. However, having been asked by a client of GS to review their analysis, I cannot conclude that the paper proves this at all. Even GS admit that not all of the UK’s economic problems can be attributed to Brexit, although most of the media have chosen to report their conclusions as both correct and solely due to the UK leaving the EU.

The uncertainties of ‘doppelgangers’

The GS paper is based on the use of a doppelganger model (in which a synthetic index is constructed composed of data from supposedly similar countries). A doppelganger is technically a twin, and twin studies are very useful in biology, but much less useful in economics primarily because there are no identical twin countries. And even if there were, unless their respective governments make the same decisions all the time, they are unlikely to get similar economic outcomes.

The nearest geographical examples of ‘twin’ countries are the result of political divisions of an island or peninsular – such as Haiti and the Dominican Republic, or North and South Korea. Both are great examples of why doppelganger economics doesn’t work. Haiti has a GDP per capita (PPP) of about $3,000 and a population of 11.7 million, while on the other side of the island, the Dominican Republic has a GDP per capita (PPP) of around $20,000 and a population of 10.5 million. Similarly, North Korea’s GDP per capita (PPP) is estimated to be under $2,000, while its near ‘twin’ South Korea’s GDP per capita is around $50,000. At the start of the Korean War, North Korea’s population was about 10.5 million and South Korea’s was about 20 million, both populations have grown at similar rates, but their GDP per capita (PPP) could not be more different. Both Haiti and the Dominican Republic as well as North and South Korea are good examples of how countries with similar locations, land masses and climates can have hugely different economic outcomes due to contrasting government policies or cultures.

GS claims to have used ‘similar’ countries to the UK in their doppelganger models. However, they have not revealed which ‘similar’ countries they have chosen to use in their comparative doppelgangers, nor have they explained in what way these countries are similar to the UK so we will never be able to check whether their analysis is fair. Nor do they explain why they change comparators between issues. Some are simple comparisons, others use a doppelganger index. When comparing real GDP per capita they use the US and the Eurozone; when comparing trade, they refer to a ‘G7 mean’ and a ‘range of doppelgangers’; and when comparing real GDP they use a G7 doppelganger but don’t say if this is made up of all of the G7 countries or just the G7 countries that they believe are similar to the UK. It does look as if they are cherry-picking their comparators.

Readers interested in further analysis of the inappropriateness of Synthetic Control Method (SCM or doppelganger) analysis might like to look at the 2023 article by Gudgin and Lu (The CER doppelganger analysis does not provide a credible measure of the impact of Brexit). Incredibly the Goldman Sachs paper erroneously claims that this article found a 5% Brexit hit on UK GDP, this is not true Gudgin and Lu view the impact of Brexit as essentially zero.

This doesn’t surprise me. Even the other EU countries in the G7 have very different economies from the UK, not least because 50 years of completely free trade within the EU has meant that most industries have expanded in the EU country with the greatest efficiencies and disappeared in the others. For example, since the UK joined the EEC/EU, UK steel production has dropped by about 60% while it has increased in Italy and Germany by a similar amount. France has a similar sized population to the UK but has 3 times the landmass, half of which is agrarian land, so France dominates the EU’s agricultural policy. France has almost no coal so its electricity comes from nuclear power, hydropower, and thermal power. An important difference between French and British economic performance when gas prices are high, as the UK still gets about 40% of its electricity from gas. Did the GS doppelganger take these differences into account? We will never know.

The US economy versus the EU

However, I am glad that GS’s analysis has clearly demonstrated the massive difference in real GDP per capita between the US and both the UK and the Eurozone. Obviously, being the world’s largest oil and gas exporter has greatly helped the US economy since 2021, as has having the cheapest industrial electricity in the developed world, as has being the home of the world’s largest tech firms, as well as the home of the producers of a Covid vaccination that became compulsory in many developed countries and was produced ‘for profit’ unlike the UK’s Covid vaccination. On top of a much larger fiscal expansion since 2016, all of these factors will have helped the US outperform both the UK and most other developed countries but none of these attributes have anything to do with the EU nor with ‘the structural nor the cyclical costs of Brexit’ which are, according to GS, the cause of the UK’s underperformance.

The reality of EU membership is that the US would have few of these advantages if it were a member of the EU. If the US had been in the EU, it would have been forced to follow the EU’s disastrous Net Zero plans preventing it from being the world’s largest oil and gas exporter. If the US were a member of the EU, it is also unlikely that it would be the home of the world’s largest tech companies –  the EU has failed to produce any global tech firms and is presently trying to undermine the US-based ones with its regulations and taxation. If the US were a member of the EU, it is also likely that Tesla would have been smothered at birth as at the time the EU was promoting its much larger ICE diesel car industry, with embarrassing results. If the US were a member of the EU, its farmers would be in the process of reducing their animal herds by 25% and its massive soybean and corn fields would have three crops growing in them instead. If the US were in the EU, GM crop development, which has reduced pesticide use globally, would have been banned or would have moved to Asia, followed closely by clinical trials for new medicines. Wall Street traders, like GS, would have already hit their Double Volume Caps even though we aren’t even at the end of the first quarter of 2023.  I could go on, but you get the idea. It is not a coincidence that the US has outperformed both the Eurozone and also a recent member of the EU which has failed to make use of its Brexit freedoms – the UK.

But despite the US’s stellar economic performance relative to both the UK and the Eurozone, GS is sure that the UK’s comparative under-performance was due to the costs of Brexit which it claims has caused lower trade, lower business investment and lower EU immigration. I shall examine their evidence in turn.


Rather incredibly, GS claims that UK goods trade has ‘declined significantly’ and has ‘underperformed other advanced economies by roughly 15%’ since the referendum. There are several things wrong with this statement:

  1. Total UK trade hasn’t declined. Measured annually, using Chained Volume Measures (CVM) total UK trade was 11% higher in 2023 than in 2016. Total goods trade was slightly lower, down 2% between 2016 and 2023 while total service trade was considerably higher up 33% between 2016 and 2023.
  2. GS seem to only be interested in goods trade and only in total goods trade, adding exports and imports together. Using total trade (i.e. exports plus imports) as a measure of economic health can be misleading as a country could have a large increase in imports compared to their exports but their total trade would still look healthy. It is more usual to subtract import values from export values to give a trade balance.
  3. Although GS is only interested in goods trade, the UK now exports more services than goods, helped by EU free trade rules moving industries to the most efficient EU producer. In 2023 UK service exports, in current prices (CP) were 54% of the UK’s total exports including precious metal exports, and 55% if we exclude them, as is customary. In 2023 UK service imports were 35% of total imports, with and without including precious metals. Even adding imports and exports, total service trade was 44% of total trade including precious metals and 45% if we exclude precious metals, not 40% as GS states on page 4 of its report.
  4. This brings us to another problem, if this paper is about the cost of Brexit, why is GS reviewing ‘total goods trade’ rather than simply ‘total goods trade with the EU’? The ONS publishes both figures. The UK’s two largest trading country partners are the US and China – neither are EU countries, but changes in their trade with the UK will be a large part of the UK’s total goods trade volume even though changes in this trade have nothing to do with Brexit. For example, UK goods imports from China fell by 21% between 2022 and 2023 while UK goods exports to China fell by 28%. This fall will be a big part of the drop in total trade in 2023 as China was our largest import supplier in 2022 but changes in UK trade with China has nothing to do with Brexit. Incidentally UK goods imports from Germany – the UK’s largest EU trading partner – were UP by 8% between 2022 and 2023 and our goods exports to Germany were unchanged, this was not enough to compensate for the fall in UK-China goods trade but still, it makes GS’s attempt to blame lower UK goods trade on Brexit look ridiculous as the graph below shows.


  1. Strangely, GS chose to index UK goods trade from the referendum in June 2016, but nothing changed in UK trade until February 2020 and UK trade with the EU didn’t change until January 2021, at the end of the Transition Period. If Brexit hurt UK trade, it wouldn’t have been apparent before 2021. Meanwhile, some of the other G7 countries seem to have had quite low trade in 2016 making the indexed comparison between the UK and the G7 Mean look more dramatic than it would have had they started their index on for example Q1 2021. The graph below shows UK goods trade indexed on Q2 2016 and on Q1 2021. They have the same shape but the one that actually measures the change in UK trade with the EU since Brexit, the line indexed on Q1 2021, is positive in Q4 2023, unlike the one indexed on Q2 2016. We can only guess why GS chose to use the Q2 2016 line for their analysis. The highs and lows on both lines are due to pre-no-deal stockpiling in 2019, covid lockdowns in 2020, pre-Brexit stockpiling in 2020 and the increase in oil and gas trade in 2022.


  1. GS Exhibit 4, seems to use some visual propaganda tricks, squashing 16 years of quarterly data into their three-inch wide trade graph, (their GDP graphs only use 9 years of quarterly data), as well as raising the x-axis to 70, both make UK total goods trade look much more volatile than it is.
  2. The GS Exhibit 4 Graph also compares the G7 mean with the UK, both indexed to Q2 2016. But four of the G7 countries: Japan, Germany, France and Italy import almost all of their oil and gas while Canada and the US are major oil and gas exporters. The US exported six times the volume of crude oil in 2022 than it did in 2016, and twice the volume of gas. Canada’s crude oil and gas exports by volume were 26% and 20% higher than in 2016, respectively. So, the GS graph showing UK total goods trade below the G7 mean is probably not something to get upset about, we still use our own oil and gas production as well as imports, so our trade in these commodities is lower than the four G7 oil and gas importer and the two G7 oil and gas exporters. It is just a pity that we don’t produce and export more oil and gas.
  3. Finally, I would also be very interested to know how their ‘range of doppelgangers’ outperformed the G7 mean for total goods trade. Did they include China, India or South Korea? Maybe they included the Netherlands, if so did they adjust for the Rotterdam effect? Perhaps they included Australia which exports large quantities of fossil fuels and agricultural products, unlike the UK. Who knows.

The GS report was badly timed, being published about a week before the ONS published the UK’s trade figures for 2023. I am sure this wasn’t done on purpose, however when actual trade figures are available, inventing doppelgangers to analyse UK trade is pointless. The graph below uses annual figures for UK goods trade for all destinations, in CVM, indexed on 2016, with the axis starting at zero. Not very exciting, is it?


If there is a story in post-Brexit UK trade, it is surely the massive increase in UK service trade, unlike the Project Fear predictions. (see graph below).  Considering Goldman’s primary business is financial services, I am really surprised that they chose not to mention services when discussing trade. It is only in GS Exhibit 5 that we notice that UK Service trade has outperformed both the G7 mean and their mysterious Doppelganger, but they don’t make a big deal about this in the text.  How strange.


Trade costs

GS does mention service trade in terms of cost as in: ‘service exporters are far less likely to report additional costs resulting from Brexit.’ This should hardly be a surprise to GS. They should know that the EU Single Market was never really set up for service trade, so whatever costs exporters were paying before Brexit should not have greatly changed post-Brexit. They should also know that the UK’s largest market for service exports is the US, not the EU.

GS Exhibit 5 has some clever nudge unit use of bright colours for businesses who claim their costs have increased but are now stable, or for businesses who claim their costs are still increasing, but the chart uses pale colours for businesses who have not experienced cost increases or are unsure about changes in their costs. This colour use emphasises the negative and makes things appear worse than they are, when in every category the majority of businesses have either not seen their costs increase, are unsure if their costs increased, or their costs increased but are now stable.

Increased costs for goods exporters should not be a surprise: trade with the EU was never free. The costs were borne by UK consumers rather than by the company benefiting from the imports or exports. Brexit has corrected this mismatch. So, while some businesses are complaining that they must now pay their own trading costs, (how unfair!), HMRC is now keeping the proportion of UK import duties, VAT payments and EU membership fees that it used to send to the EU. Technically this extra revenue should have reduced UK government borrowing or paid for better services or lower taxes.

Comparison to other advanced economies

As for the GS claim that UK trade is 15% lower than ‘other advanced economies’, unless we know who these mysterious advanced economies are, there is no way of rebutting this claim. But if we use the ITC COMTrade database for all G7 country goods exports measured in US$ and indexed on 2016=100: the UK was 129 in 2022 (ITC hasn’t updated their database for 2023 yet), Japan was only 116 in 2022, France was 124, Germany was also 124, the US was 142, Italy was 152, and Canada was 153. Goods imports were the opposite: Italy’s imports were 183 in 2022, the US was 150, Germany was 149, Japan was also 149, France was 145, Canada was 142 and the UK’s imports were only 128. (See the two graphs below).



In short, between 2016 and 2022, the UK had higher goods export growth than advanced economies: Japan, France and Germany, and the UK had lower goods import growth than advanced economies: Italy, the US, Germany, Japan, France and Canada. I am having trouble seeing the UK’s relatively balanced growth in both goods imports and goods exports as such a bad thing. Countries with massive growth in imports between 2016 and 2022 relative to their growth in exports over the same period, will have balance of payments problems eventually causing the value of their currencies to fall. The Euro adds some complexity to this process, obviously, but the Dutch can’t hold up the whole Eurozone by themselves forever.

Success of the UK’s new trade deals

Finally, GS dismissed the UK’s new trade deals, but they may be interested to know that although the UK-Australia trade deal only came into effect 6 months ago, UK exports of Cars to Australia increased by 116% from £521 million in 2022 to £1.14 billion in 2023, UK exports of Road vehicles other than cars also increased by 84%, Aircraft exports were up by 69%, Ships by 108%, even UK dairy exports were up by 13% and live animal exports were up by 59%.  Who saw that coming? Not the Treasury.  And before the Welsh Farmers start complaining about cheap food imports: meat imports from Australia only increased by 6% and most other food imports fell. Only sugar increased significantly but it was still only about 10% of total UK sugar imports.


While I understand why we might corporate nervousness about investment, especially between 2016 and 2020. A country where the Prime Minister is threatening to leave the EU without a trade deal on one day and then trying to tie the UK to EU regulations indefinitely through the Northern Ireland Protocol on the next might be thought off-putting  Even after that Prime Minister was removed and the UK agreed a tariff-free and quota-free trade deal with the EU, there are still reasons why many businesses might have been  cautious about investing in the UK: high energy costs that are only likely to get higher as more wind turbines are added to the grid; its high corporate taxation compared to its near neighbour Ireland; its windfall taxes applied if your business is in the right place at the right time; its threats of retrospective taxation even if you were obeying the rules at the time; its woke stock market listing requirements and its low tech listing valuations. All of these things should deter businesses from investing in the UK but none of them are due to Brexit.

Corporate tax is not an EU competency, it has always been in the UK government’s power to drive businesses away from the UK through high taxation, inconsistent taxation and expensive energy.  This makes a comparison between the UK and a doppelganger interesting when comparing the outcome of various government tax policies but not informative about the structural costs of Brexit.

And again, GS has used dark colours to display the line that they want you to notice. In Exhibit 7, they use dark blue for the 20% to 30% of respondents who believed that Brexit has had a negative impact on their capital expenditure between 2016 and 2020. But the line representing the 60% to 70% of respondents who said Brexit had no impact on their capital expenditure over this period is in pale blue and much harder to see. However, those who do notice it, will also notice that the proportion of respondents who claimed Brexit has had NO impact on their capital expenditure has not dropped below 60% between the referendum in 2016 and the end of the chart in March 2020. So why is GS claiming that lower investment in the UK was caused by Brexit?

Their second Exhibit 7 graph also shows that about 30% of their decision maker panel survey thought that Brexit was not an important source of uncertainty at the end of 2023, up from about 10% in 2019. While only about 3% thought Brexit was the main source of uncertainty at the end of 2023, down from about 20% in 2019. It would appear that businesses are getting used to Brexit.

In fact UNCTAD data shows that the UK has stayed well ahead of other EU countries as a destination for greenfield investment (much of it in financial and business services) so UK economic strengths have greater sway with investors than Brexit or even UK taxation.


GS is worried about the loss of EU immigration but they didn’t mention the 7.6 million EU nationals who have applied for Settled Status in the UK as of 30 September 2023, GS seems to believe that EU immigrants went home after Brexit, when in fact the majority settled in the UK.

Although GS does admit that immigration is now larger than when the UK was a member of the EU. They dismiss the present 745,000 net non-EU immigrants in 2022, claiming that EU immigrants had a higher labour market participation rate while non-EU immigrants are often students. But this difference is not necessarily bad for the UK economy.

According to the ONS, in June 2016, just over half of EU immigrants were from either the EU2 or EU8 countries. Both groups of former soviet bloc countries with considerably lower living standards and wages than the UK when they joined the EU. Many of these immigrants were happy to take menial jobs in the UK and live in very meagre circumstances so that they could send as much money as possible back to their families in their home countries. While most economic models count all immigrants as equal in terms of their additional consumption, when immigrants are sharing bunks with nightshift workers and sending their wages and UK child benefits back to their families in Eastern Europe, their contribution to the UK economy is much less than most GS economists would expect.

On the other hand, GS is unimpressed by the number of non-EU students immigrating to the UK, believing they add little to the economy. However non-EU students are bringing large amounts of money into the UK economy and supporting the UK’s tertiary education industry. Non-EU students have to pay full tuition fees, on average about £25,000, as well as paying for their accommodation and food. Each student is worth about £50,000 per annum to the UK economy. If they were employed instead, very few would be earning £50,000 a year. An income which would put them in the top 20% of UK wage earners. GS may value Bulgarian baristas more than rich Chinese and Indian students, but from an economic perspective, we are better off with the students who spend their money in the UK, than we were with the EU nationals who sent their wages back to Eastern Europe.

GS has also tried to prove that lower EU immigration has caused labour shortages in Accommodation and Food services but when an industry has an 85% increase in pre and post covid vacancies but only 17.5% of their workers were from the EU, it is a bit of a stretch to blame this on Brexit. An alternative explanation would be that the wages on offer in this industry are too low to attract the required staff.

According to the ONS, the wholesaling, retailing, hotels and restaurant sectors saw the largest annual regular pay growth rates in Sep-Nov 2023, up 7.2%. This wage increase is a Brexit benefit for workers in this sector, including the many EU workers who have remained in the UK.

The theory of market pricing is that shortages push up prices, which attracts greater supply, which in turn alleviates shortages. But unlimited EU immigration whilst the UK was a member of the EU, prevented worker shortages from occurring and so depressed wages in many low-skilled jobs. While this was great news for international fast food and coffee shop chains, it wasn’t great for the UK nationals who worked in these industries. Many of the people who voted for Brexit did so purely so they could get a pay rise because their employers could no longer replace them with an immigrant from Eastern Europe at a moment’s notice.

There were other industries on the GS Exhibit 10 diagram with an 80% increase in pre and post covid vacancies even though EU immigrants made up only 7.5% of their workforce. It would appear that not all staff shortages are due to Brexit, which makes the GS ‘line of best fit’ look somewhat tenuous.

As for their ‘correlation coefficient’ between the change in net immigrant inflows and the composite PMI employment index, the two groups with the highest correlation, All employees and Total EU employees, were only just over 0.5, that is, they only showed a moderate correlation. When things are highly correlated, the coefficient should be close to one. While the change in net inflows of EU15 immigration showed little correlation, and EU2 immigration, immigrants from Romania and Bulgaria, weren’t included in the analysis despite making up over a quarter of all EU immigrants in Q2 2016. I find this graph very unconvincing.

GS believes that cyclical inflows of EU workers have exacerbated labour market tightness and thus contributed to the UK’s higher inflation rates since 2016. While this is true, that isn’t necessarily bad. Many employed UK nationals are enjoying real wage rises for the first time since the expansion of the EU into the former Soviet bloc. Tony Blair’s Labour Government decided to allow unlimited eastern EU nationals to move to the UK from 2004, while France and Germany kept restrictions on EU8 workers for 7 years.


I am not convinced by the GS paper that the UK has significantly underperformed ‘other advanced economies’ – whoever they might be. Nor am I convinced by their analysis that this supposed underperformance is due to the economic costs of  Brexit. But Brexit has allowed the UK to make its own rules and to follow its own path if only our politicians had the courage to do so. If we want to have the same economic performance as the US, then we should be following the same or similar policies: develop our oil and gas, lower the cost of our electricity by using more gas generation, allow our farmers to plant GM crops, stop imposing woke rules on listed companies, encourage clinical trials back to the UK, etc. At least the UK have abandoned the EU’s crazy double-volume caps and its three-crop rule since Brexit.

If the UK wants to improve its trade, investment and immigration it should concentrate on improving its economy through better regulations, competitive tax rates and cheaper energy. If we are producing competitively priced goods and services, then people will import them, invest in the UK and immigrate here whether we are in the EU or not.

Print Friendly, PDF & Email

About the author

Catherine McBride