The UK is currently facing two wicked public policy issues – Covid-19 and Brexit. As the country U-turns into another national lockdown while also hurtling into a predictably high cost Brexit, I cannot be the only person to have been struck by the contrast between the ‘We’re following the science’ of Covid-19 and the ‘We’ve had enough of experts’ of Brexit. The intellectual effort expended on each crisis has been huge and an important question is whether that effort is helping/has helped to define better outcomes.
It is too early to know how Covid-19 will turn out, but for those of us who have spent the last five years using analysis to inform the Brexit process, we have to wonder whether our efforts have had any effect and whether – a different question – the effort was worth it. Short answer: it was.
Thus, this is an essay about the interface between research – ‘science’ or, nearly equivalently, analysis and expertise – and policy in a divided society where the stakes are high. It reviews the ‘evidence’ (as opposed to the commentary) on Brexit and then asks why so little use was actually made of it. It identifies a number of challenges in bringing evidence to the policy debate and asks whether eventually these will also marginalise science in the Covid-19 debate. It concludes, however, that even when decision makers do not act as their experts advise, investment in producing that advice is still good value. As an economist, much of my experience is in economics but the story applies to any discipline that uses current experience and knowledge to inform future policy – such as social policy or law.
Brexit attracted little analytical attention until the referendum kicked off in February 2016, and even immediately after that there was relatively little formal work on either side. Hence this essay is mostly about the post-referendum period: in principle, ‘the decision’ – to leave – had already been taken, but hardly a single detail was fixed, and between the 2017 general election and the premiership of Mr Johnson even ‘the decision’ was in doubt.
Some Brexiters opposed membership of the European Union out of principle – a respect for what they define as sovereignty and a belief that it would be unduly eroded if the UK remained within the Union. They obviously did not stress any economic costs of Brexit, but most did not really deny them either.
“Even when decision makers do not act as their experts advise, investment in producing that advice is still good value.”
Most advocates of Brexit, on the other hand, were late converts and many of them argued that there would be economic benefits. For some, including by his own admission the current Prime Minister, it was a very close decision. Some believed in a free trade nirvana once the UK threw off the shackles of EU protectionism, while others hoped for a retreat from globalism. The leave side of the referendum campaign picked and mixed these world views according to their audience and consciously decided not to resolve them into a coherent Brexit strategy. The Cameron Government decided not to dignify Brexit with any serious internal analysis and as a result we went into Brexit on 24 June 2016 with no road map at all.
The science – phase 1
When Theresa May became Prime Minister on 13 July 2016, the absence of a plan for Brexit was evident and Government effort mostly went into managing the split between Leavers and Remainers within the Conservative Party. Again vagueness – ‘Brexit means Brexit’ – ruled, except in a dramatic speech by May at the Conservative Party Conference in October 2016. In this, almost entirely unaided by formal analysis, she suddenly announced that Article 50 – the exit mechanism for effecting Brexit – would be invoked by ‘the first quarter of 2017’ and also that the UK had four red lines in negotiations with the EU on the future UK-EU relationship.
Economists had been busy, however, and were starting to set out the analytical tools for thinking about Brexit. One group – led by Patrick Minford – had argued for a while that there was a 4% premium to GDP from leaving the EU, but nearly all others were concerned by the potential cost. By August 2016 respected research organisations had identified most of the key concerns including, for example, work by the Centre for Economic Performance (LSE), the Centre for Economic Policy Research, the Institute for Fiscal Studies (via UK in a Changing Europe), the Overseas Development Institute and the newly established UK Trade Policy Observatory. They suggested, inter alia, that:
- customs formalities and rules of origin would hinder trade in any arrangement that left the UK outside the EU Customs Union;
- the ‘WTO option’ for UK-EU trading relations would be highly disruptive;
- the EU Single Market is a major facilitator of both goods and services trade, and the EU was unlikely to permit cherry-picking the parts of this that the UK happened to like;
- services trade matters for the UK economy – services account for well over half of production and consumption in the UK;
- equivalence in financial services regulation would not be sufficient to ensure the continued trading success of the City;
- developing countries had a strong interest in maintaining their access to the UK market; and
- breaking free of EU trade policy with third countries may offer opportunities, but it would need to be part of a coherent strategy, which would be difficult to define for a medium-sized economy like the UK.
By November 2016, researchers had noted that May’s four red lines implied that the closest feasible UK-EU relationship was a relatively simple free trade agreement – ‘Canada-plus’ – and that this would exclude the UK from the EU Customs Union and Single Market. And later we realised that this was not compatible with an open border in Ireland.
Policy is about trade-offs and so, despite being spot-on on the agenda, the early studies still left plenty to do. Thus 2017 and 2018 saw a rash of quantitative estimates of the effects of Brexit. These were based on simulation models, because, simply, there is no other plausible way of quantifying a future the like of which you have never seen before. At this stage there was no indication of how the future relationship with the EU – by far the largest consideration numerically – would work out, and so researchers had essentially to proceed with assumed scenarios. Figure 1 offers a neat summary of the main estimates of the long-run impacts of Brexit.
Leaving the details aside, two things stand out. First, all but one research group foresaw negative economic consequences. The exception was Minford’s Economists for Free Trade whose approach came in for intense criticism which, while apparently not causing them to change their minds, probably reduced their influence. Second, the range of estimates is huge. Analytically the variance between studies was a strength, because it suggested that the various different approaches and assumptions all produced a qualitatively similar answer (i.e. that Brexit was costly). But politically it was depicted as a weakness. I frequently heard the argument “if the economics profession can’t distinguish -10% from -2%, how do they even know that the effect will be negative”. But there was also considerable variance within studies across scenarios, which showed that the flavour of Brexit mattered – i.e. that even if Brexit were certain, there were still serious policy issues to decide.
“The distance between policy and analysis widened even further as the government sought to find a solution – any solution – within Westminster: politics took over entirely from economics.”
One other entry in Figure 1 deserves note: HMG (Her Majesty’s Government) projected losses under all circumstances and substantial losses from what we now term ‘No deal’. These projections had been available to government for most of 2018 and much of the government gradually came to realise the costs of Brexit and to try to devise plans to mitigate them. This realisation represented an unsung success for broad-based analysis. The HMG estimates were eventually published (with a little tweaking) in November 2018 to buttress the case for May’s proposed deal.
However, as we all know, May’s vision of a close accommodation with the EU came to naught; not only because of opposition within the ruling Conservative party, but also because the opposition parties could not (or would not) organise themselves into supporting it or reversing the decision. The distance between policy and analysis widened even further as the government sought to find a solution – any solution – within Westminster: politics took over entirely from economics.
As well as the dark arts of macro modelling, economists and lawyers provided huge volumes of analysis on specific elements of Brexit. Some of this fell on deaf ears but some did not. Among the former, economists showed that even before any policies had been changed, Brexit uncertainty had slowed GDP growth relative to expectations, (‘Ah’, one MP said to me ‘but GDP is still growing, so that’s not a problem’!) and that UK firms were withdrawing from EU markets and vice versa. The 2019 general election slogan of ‘Get Brexit Done’ was justified in terms of ending the uncertainty. In November 2020, the fact of Brexit has now been established, but its flavour has not and economic uncertainty was basically the same in September 2020 as in November 2018.
Similarly, the Protocol on Ireland/Northern Ireland in the Withdrawal Agreement negotiated with the EU in November 2019 was passed by Parliament in January 2020 and is now under threat of abrogation by the Johnson Government. Yet the features to which the Government now objects were pointed out to them in December 2019: lawyers showed that the Protocol left EU institutions firmly embedded in UK policy space and economists showed that it was likely to involve the EU determining the tariffs paid on 75% of imports into Northern Ireland (including those from Great Britain).
In other cases, analysis may have influenced policy – for example, the case for the UK adopting the EU ‘Everything but Arms’ scheme of trade preferences for least developed countries, and the dangers of Kent being grid-locked by lorries if border formalities increased at Dover. And while it seems unlikely to yield fruit, the UK draft for the UK-EU FTA did at least ask for a wrinkle in rules of origin called ‘diagonal cumulation,’ extensive mutual recognition of professional qualifications, and the inclusion of audio-visual services, which analysts had identified as critical issues for the UK.
Why didn’t they follow ‘the science’?
There are several potential explanations of why economists were side-lined so comprehensively in the Brexit process, but personal ambition and the sheer ruthlessness of Brexiter MPs has got much to do with it. The Brexiters never expected to win and never quite believed they had won. They fretted that the establishment (the ‘deep state’) was going to take it away from them. Given the lack of an articulated strategy, the best antidote was to avoid all analysis, as witnessed by the almost invariable response to any unwelcome argument: ‘Project Fear’.
“There are several potential explanations of why economists were side-lined so comprehensively in the Brexit process, but personal ambition and the sheer ruthlessness of Brexiter MPs has got much to do with it.”
Such absolutism was aided by the fact that the Brexit vote was not primarily driven by economic considerations, but rather those of identity and alienation (including, for some, xenophobia). However, the alienation was substantially economic in origin and identity policies have eventually to confront their economic costs, so I do not think that this explains the nearly permanent exclusion of economics from the debate.
Two other factors help to explain the exclusion of analysis. First, so-called “do-it-yourself economics”: in his Reith Lectures of 1985, David Henderson argued that “There is no doubt that the policies of governments … are influenced by economic ideas. But … these have not necessarily been the ideas of economists”. DIY economics resides in the intuitive ideas of ‘practical’ men and women, especially politicians. Henderson quotes the economist William Niskanen as saying “In contrast with their attitudes towards lawyers and engineers, …. every senior manager …. thinks of himself as an economist, as having a democratic right to make his own judgement about economic phenomena”, and he notes that this attitude is not limited just to managers. Thus, for example, it was held to be obvious that the fate of exports was the key consideration in assessing the economics of any Brexit outcome, in contrast to the economist’s view that imports and the terms of trade are just as important. Hence, it was argued that the devaluation of sterling that followed the referendum result would alleviate any economic problems, and that the UK would withstand Brexit better than the EU because the EU exported more to the UK than vice versa. Unlike the early days of Covid-19, there was felt to be little need to consult ‘experts’ on something so obvious as trade.
The second factor which hindered economists’ scope was the necessary imprecision of modelling – both the disagreements among scholars and the uncertainty that surrounds any modelling exercise. And both were compounded by the (bogus?) inability to understand that models aim to tell us the consequences of adding one policy measure to an underlying trend rather than to make unconditional predictions. Many times economists have been dismissed because the slump they were said to have predicted to follow Brexit had not materialised, when all they were saying was that eventually income would be x% lower than it otherwise would have been. The modern ‘aggressive uncertainty’ argument, which one associates with the sans-nuance world of social media, runs: ‘Unless you can prove to me that X will certainly happen, I will not recognise your contribution at all’. And the usual unspoken corollary is ‘and therefore I will believe whatever I damn well please’!
This phenomenon has now caught up with the medics, in the form of calls for evidence that this or that set of restrictions will halt the spread of Covid-19. But there is no evidence, because, as with Brexit, we’ve never done anything like this before. All we have is bits of analysis and experience of similar issues which we have to weave into chains of reasoning. There is room for disagreement about those chains and even more so about the factors outside the epidemiological models such as the economic harm of lock-down or mental health side-effects. The only solution is to get all the considerations out on the table and try to reach a consensus about sensible, not proven, ways to proceed.
With Brexit it was noticeable that, starting at the very top, there was little willingness to confront uncertainty or to tolerate alternative views. Because (the more principled) economists and other commentators would not pretend to certainty they were marginalised. Any reservation was treated as akin to treason and as a result the government largely refused to internalise, or even to hear, any of the many analyses designed to make Brexit less costly.
One victim of this was the business voice. The usual worry is that trade policy making is dominated by powerful producer interests, yet this time they were, until recently, more or less silent. Certainly Brexit-doubters were often excluded from government meetings but business also realised, I believe, that the Brexit vote had been partly a backlash against their supposed earlier influence. One group that had disproportionate influence before the financial crisis was the financial services sector – potentially one of the greatest losers from Brexit. It argued for close integration with the EU, was largely ignored by Theresa May and so, as Albert Hirschman would have predicted, it packed its bags and prepared a partial withdrawal to the continent.
The same logic lies behind the general quiescence of the traded services sectors. The UK services sector will undoubtedly suffer heavily after Brexit (witness, for example, foreign direct investment in it fell after the referendum), but the firms involved will suffer less because they and their key workers can work or set up shop elsewhere.
Back to Covid-19
Maybe surprisingly, I do not feel any schadenfreude as medics and epidemiologists struggle to explain their models and differing conclusions about dealing with Covid-19; indeed, I admire the way they have derived bits of information and tried to complete the picture by simulation modelling and also the way in which they have toiled to explain what they do and do not know, and how much the latter has been accepted by other players. But I am struck by the very different experiences of Covid-19 experts and Brexit experts. Put simply, ministers seem to have been more willing to listen to scientific experts over Covid-19 than to economics experts over Brexit. The contrast lies partly in the different natures of the two problems. At first Covid-19 appeared to be a rather narrow technical matter which called upon the skills and authority of technicians to solve. And the pandemic was an extraneous event – no-one had argued and voted for it, so there were no predetermined political positions as there had been with Brexit.
“Put simply, ministers seem to have been more willing to listen to scientific experts over Covid-19 than to economics experts over Brexit.”
However, as we are now seeing, even Covid-19 experts are struggling to be heard. As the pandemic has dragged on lockdowns have been seen to have major economic and social consequences. Debate has therefore become more overtly political, especially over the distribution of the burden, and the validity of the scientific advice has been challenged, notably within the Conservative Party. The Prime Minister has ultimately bowed to the majority scientific view and has taken an unpopular decision over lockdown, but the delay has been costly in health terms and the political cost in his party has been high. It will take both luck and skill to avoid Covid-19 slipping into the same cul-de-sac as Brexit, with poor policy ultimately emerging from a fog of willful ignorance and do-it-yourself expertise.
Will expertise come back into fashion?
It will be a while, if ever, before experts regain a respected place in the world of high-level policy making. Political discourse is deeply polarised at present and that reduces the space for reasoned and respectful policy debate, and most politicians see uncertainty as an avoidable weakness. These limitations may reverse, but achieving that will need remarkable leadership. Moreover, championing the centrality of expert advice as a matter of principle is not something that economists or medical researchers can easily pursue as professionals. Hence the immediate question is how professional researchers should operate in the current context.
Here are a few thoughts. In dealing with policy decisions, we must be scrupulous about separating what we know (to some acceptable degree of confidence) from what we think and what we hope. Listening to the many members of SAGE (the Government’s Strategic Advisory Group for Emergencies) on the radio every other morning, I think the medics have been better at this than were the economists.
“Policy issues are never settled by research alone – and this applies equally to Brexit and Covid-19.”
In particular we must not shy away from admitting uncertainty but be sure to emphasise that all courses of action ultimately involve some uncertainty. This is partly a matter of making sure that the burden of proof is correctly assigned. One of my biggest surprises in the Brexit debate was that, whereas previously it had been incumbent upon those advocating radical change to argue why doing so was a reasonable course to adopt, in Brexit it was reversed. The Brexiters would make claims (e.g. technology would solve the Irish border problem; a UK-EU free trade agreement would be the easiest negotiation in history) and it was somehow up to ‘the gloomsters and doomsters’ to disprove them.
Governments make decisions, but not in a vacuum. One of the things that has gone well over Brexit is that researchers have informed Parliament, civil servants and the general public to a much larger extent than on any previous trade policy issue. This has helped to improve the general debate and will, I hope, have planted sufficient markers that analysis can play a significant role as post-Brexit policy develops. Brexit – as a set of policies – is not over yet: we will be deciding the details for years!
One temptation when government is hostile is to trim your advice to what you believe the government will accept. Of course, one needs to address the government’s agenda, but it is very important for the long run health of policy making not to fall into declaring bad ideas to be good ideas on the grounds that at least now you are on the inside. Policy issues are never settled by research alone – and this applies equally to Brexit and Covid-19. Politicians have to balance myriad claims and so, infuriating as it is when they over-rule our advice, it is not necessarily illegitimate. Our response should just be “the ‘science’ says X, but I recognise that your job is broader than mine”. What is illegitimate in the 21st century, however, is simply to dismiss analysis with “We’ve had enough of experts”.
By L Alan Winters, Professor of Economics and Founding Director of the UK Trade Policy Observatory, University of Sussex.