17 February 2023

L. Alan Winters is Co-Director of the Centre for Inclusive Trade Policy, Professor of Economics at University of Sussex Business School and Fellow of the UK Trade Policy Observatory.

Given the recent discussions about future UK-EU relations and the review of the UK-EU Trade and Cooperation Agreement (TCA) after two years, I have started reading the Brexit literature again.  A recent paper – ‘What impact is Brexit having on the UK economy?’ by Graham Gudgin, Julian Jessop and Harry Western (GJW) from October 2022 argues there is no hard evidence of harm and that studies that claim to find harm are biased and/or incompetent! In this blog, I consider a few of their points in four areas.

First, however, I must note GJW’s grand conclusion

What does this say about the economics profession? We assume that much of the inaccurate work comes from not interrogating data sufficiently rigorously and settling too quickly on evidence appearing to support the anti-Brexit case. ….  If the profession cannot be trusted to put ideology aside in economic analyses, then policy debates will be distorted, and the public may lose even more faith in what Michael Gove called ‘experts from organisations with acronyms’.  (page 30)

---

Is this intentionally ironic? The paper is dripping with (pro-Brexit) ideology, and Gudgin is from the Centre for Business Research (which refers to itself as the CBR); Jessop, the Institute of Economic Affairs (the IEA) and Western won’t even tell us where she/he is from!  

Estimating effects on GDP

A pretty constant refrain from GJW is that

some of those who were critical of the vote to leave have retreated from catastrophe scenarios towards the cherry-picking of data, the tortuous use of ‘counterfactuals’, and the selective deployment of forecasts …. (Executive Summary)

When you are debating policy about the future, like, say, advocating Brexit, it is difficult to avoid statements about what will happen in certain circumstances – that is, forecasts (e.g. “there will be many opportunities” is a forecast). However, put that aside: I will focus here only on things that have happened and hence for which we have data.

“The hard evidence is that leaving the EU has had remarkably little impact on the UK economy.” (Exec Summ.)

But when you come to the analysis behind this ‘hard ‘evidence (pp. 3-10), we read

analysts (some of whom are clearly opponents of Brexit) have tried to use counterfactual scenarios …. The key weakness of all these counterfactual scenarios is that they have used as benchmarks past periods in which the UK economy was recovering from recession.

A ‘counterfactual’ is an estimate of what would have happened if Brexit had not occurred: we know what did happen, and to see if Brexit made a difference (is a cause) we need to know what would have happened without Brexit; but since we can’t observe the latter, we have to estimate it. Thus, the validity of a counterfactual cannot be proven, merely argued to different degrees of plausibility.

GJW argue that using 2009-16 (the post-financial crisis to pre-referendum (and pre-Covid period)) to construct counterfactuals for the UK biases them towards countries that regularly grow fast and hence renders them unsuitable as comparators for the UK that only occasionally does. Criticising the best-known counterfactual (from John Springford), GJW declare that he should, instead, have chosen France, Germany and Italy because they are large European economies like the UK. Springford describes the statistical method he uses to determine which countries best replicate the behaviour and structure of the British economy before the referendum and concludes that, for estimating GDP growth, the set is the United States, Germany, New Zealand, Norway and Australia. GJW’s approach to choosing the counterfactual is not described at all, and while their three candidates are similar to the UK in size and location, they also differ significantly in having the euro, larger public sectors, larger manufacturing sectors, and faster population ageing than the UK.

GJW also complain (p. 6) that the countries in Springford’s counterfactual may have been subject to different unrelated shocks; but apparently do not think this is a problem with France, Germany and Italy. Trying to sort out Brexit from the various other shocks that afflict countries, including Britain, has long taxed economists. GJW’s graphs contain useful information about the economies concerned, but eye-balling graphs is not, by itself, professionally accepted as an adequate way to separate out different causes behind economic performance.

Overall, GJW’s arguments for using a different counterfactual are worth considering, but they do not remotely reach the threshold to constitute ‘hard’ evidence. Indeed, GJW confuse an alleged absence of evidence (of harm) with evidence of absence – an elementary error.

One side-note in this section of GJW is that “losses related to the UK remaining aloof from future EU regulatory reforms [are] a matter of future UK policy, not Brexit per se,…” I thought that remaining aloof was the point of Brexit. This is the intellectual version of the political argument that ‘Brexit was a great idea, but has just been implemented badly’. But many argued that it could never be implemented into success, and nothing so far disproves this.

Investment

Springford and many others argue that Brexit curtailed investment after 2016. GJW respond that UK investment boomed over 2009 -16 and hence a slowdown was “to be expected”, and besides the decline applies only to business investment and not if you add in government investment and housing (substantially influenced by government). They say “apparently negative comparisons with other economies may just be picking up the effect of unsynchronised investment cycles.” (emphasis added). The UK fell behind others, especially over 2017-19. GJW observe that this predated the UK exit from the EU customs union and single market but, delicately, they don’t say whether that shows that Brexit wasn’t to blame or that, as a forward-looking activity, investment started to adjust to Brexit straight after the referendum.

On foreign direct investment (FDI) they hold that all is well because the UK is still the largest single recipient of greenfield FDI in Europe. They fail to observe that their figure (p.13) shows other countries closing the gap significantly after the referendum.

Trade

GJW state:

Far from collapsing as some claim, UK trade with the EU has fully recovered after some initial disruption, despite increased trade frictions.

This is just plain wrong. Aggregate exports have recovered – and economists who had not expected them to do so have admitted they were wrong (p. 16). On the other hand, UK goods imports from the EU relative to imports from non-EU are down by 28%. And on non-EU trade, GJW state

exports to non-EU countries were down by around 11% (Chart 15). Obviously, this cannot be due to Brexit as no significant new trade barriers with non-EU countries have been created.

This export weakness may be independent of Brexit, but not ‘obviously’ so. UK industry imports many parts and much equipment from the EU; the decline in such imports and the hit to investment could certainly curtail UK output available to export to other countries.

GJW contest the view that less trade would mean lower productivity, which, as they say, underpins some of the gloomiest prognostications about Brexit. And they cite a paper “suggest[ing] the weight of evidence” favours their view. Only the paper doesn’t quite say that. It says “whether high productivity businesses choose to trade … or whether trading causes business productivity to rise …  or whether causation runs in both directions, remains unclear.” And the section to which GJW refer does not even consider the impact of imports on productivity, a positive effect found by several persuasive studies.

Sterling

Of course, the precise timing and extent of the fall in sterling in 2016 was down to the shock referendum result. But this does not mean that sterling would not have weakened anyway, over a longer period.

Well, you never know, sterling may have weakened by itself! But such a counterfactual is much less persuasive than one stating that on 23^rd June 2016, there was no net expectation that sterling would depreciate and that an unexpected referendum result to dissociate from our main trading partner caused an overnight depreciation that has never been recovered. It is wilfully obtuse to claim that “there is no strong evidence that Brexit” affected prices via the exchange rate.

Analysing Brexit

It’s useful to debate technical details and GJW raise some legitimate questions. But their cherry-picking of evidence grates, their accusation that others fail to “interrogat[e] data sufficiently rigorously” is ironic given that they then proceed precisely in that fashion, and the tone of righteous victimhood is just risible. As I have noted elsewhere, Brexit reversed the usual burden of proof for public policy: previously, advocates of change had to show the benefits of a change, whereas with Brexit the policy change was presumed and opponents had to show it was a mistake. It is the same here: GJW raise doubts about analyses that show Brexit being harmful, but are unable, and apparently content to be unable, to offer any proof that it is helping.

A careful reading of the evidence shows that while there is little evidence yet that Brexit is doing much to help the UK economy, neither is there evidence of much harm. This is significant because it was generally agreed, even by Brexiteers, that there would be initial difficulties.

“Little evidence”? Is there any? “Much harm”? What constitutes much? “Generally agreed”? Some Brexiteers did admit sotto voce that there might be teething difficulties, but never what they were and how long they would last; and I can’t remember the big speech saying “ ‘I have nothing to offer but blood, toil, tears and sweat’, but Brexit will succeed”.

It would have been so much more useful if the technical issues had been left to speak for themselves without accusations of bias and incompetence, but that of course would have defeated the clearly political intent behind it.

I would be delighted to hear from Graham, Julian or ‘Harry’ if they would like to discuss any of this.

---

Disclaimer:
The opinions expressed in this blog are those of the author alone and do not necessarily represent the opinions of the University of Sussex or UK Trade Policy Observatory.

Republishing guidelines:
The UK Trade Policy Observatory believes in the free flow of information and encourages readers to cite our materials, providing due acknowledgement. For online use, this should be a link to the original resource on our website. We do not publish under a Creative Commons license. This means you CANNOT republish our articles online or in print for free.