Some of the most widely cited predictions of the economic effects of Brexit rely on flawed analysis, particularly of the performance of the UK after it joined the EEC, and on the link between trade and productivity, write Ken Coutts (left), Graham Gudgin (University of Cambridge) and Jordan Buchanan (right) (Ulster University Economic Policy Centre). In order to restore public confidence in economic forecasting for major policy issues, economists need to use more relevant analyses, based on a wider range of evidence.
The Brexit debate has been distorted by several myths. One of the most persistent and widely repeated is that the economic performance of the UK improved after joining the EEC in 1973. This claim was made by the OECD and was regularly stated in the media during the Brexit referendum campaign. The link between trade and productivity also plays an important role in economists’ assessments of the economic performance of the UK within the EU, and the short-term and long-term economic effects of the referendum decision to leave the EU.
Many of these assessments have been by government departments and international agencies. In estimating the economic effects of Brexit on living standards, these rely on a range of analytic approaches, including the use of gravity models, computable general equilibrium models and macroeconomic forecasting models.
In our working paper, we conclude that much of this work contains flaws of analysis, and a treatment of evidence that leads to exaggerated costs of Brexit. Gravity models are well established as a technique for estimating the impact of trade associations or currency unions but require more care than has been in shown, when being applied to a specific issue like Brexit. The Treasury has been particularly cavalier in its approach, both in its application of gravity analysis and in applying a ‘knock-on’ impact from trade to productivity. Other organisations have been a little more circumspect about the productivity link, which we doubt exists to any significant degree for advanced economies, but several have used it without much questioning. The short-term forecasts which have turned out to be wrong have further damaged confidence in economists’ contributions to public debate.
Partly as a result, very little attention is currently being given by politicians or the public on either side of the debate to the impact assessments published at the time of the referendum. The potential damage to the UK’s negotiating position on Brexit may have been limited by the indifference of policy-makers to economic impact assessments. Although the UK government has steered away from further work on economic assessments of Brexit, devolved governments have felt less constrained.
The Mayor of London, reacting against the UK Government’s reluctance to publish assessments, commissioned Cambridge Econometrics, who showed that a modelling approach without gravity models or general equilibrium, will generate moderate and plausible results. Even so, only the most pessimistic of their conclusions on Brexit received any publicity. CE’s gave little attention to their prediction that per capita GVA was little changed by Brexit and hence the media ignored it.
The Scottish Government was much less inhibited and ploughed ahead with an analysis incorporating all of the flaws in the Treasury and CEP analyses with no acknowledgement of published criticisms. The consequences of these shortcomings go well beyond Brexit itself.
We believe that the credibility of the economic forecasting profession and some of the major parts of the economic press, have been damaged again. It will take more than a decade to be sure of this, but the failure of the short-term forecasts indicates what could happen. The fact that the flaws we identify all point in the direction of pessimism on Brexit, and hence in the direction that most academics and economists tend to lean ideologically, will increase the scepticism of many. The refusal of the Treasury to discuss their approach, at least until the issue was aired in Parliament, is in our view unacceptable in an open democracy.
Our conclusion is that in order to restore public confidence in economic forecasting for major policy issues like Brexit, economists need to use more relevant analyses, based on a wider range of evidence. We expect that econometric models used by commercial forecasters like Cambridge Econometrics, will prove to be most accurate in the long-run. If so, the academic profession needs to reconsider both the relevance of its current attachment to theory based on unrealistic assumptions, and to the general quality of policy-relevant applied work.
Whatever techniques are used need to be applied with more balance and scepticism. The CEP in discussing the Treasury reports could only think of changes which would have made the HMT predictions even more pessimistic on Brexit. In the words of Oliver Cromwell to the General Assembly of the Church of Scotland, the economic forecasting profession needs to ‘think it possible ye may be mistaken’.
Our conclusion is that most estimates of the impact of Brexit in the UK, both short-term and long-term, have exaggerated the degree of potential damage to the UK economy. We stress at this point that this is not a politically-driven exercise. Most of the four-person team behind the research for this and our other papers voted ‘Remain’ in the 2016 referendum and would do so again if given the chance. Our purpose is rather to establish a sound basis for the ongoing debate on the likely potential economic impact of Brexit, and more generally to question the quality of economic analysis in dealing with major, macroeconomic policy issue like Brexit.
This is an edited extract from ‘How the economics profession got it wrong on Brexit’, Working Paper 493, Centre for Business Research, University of Cambridge, January 2018. It represents the views of the authors and not those of the Brexit blog, nor the LSE.
Ken Coutts is Assistant Director of Research at the Faculty of Economics, University of Cambridge.
Graham Gudgin is Honorary Research Associate at the Centre For Business Research (CBR) at Cambridge Judge Business School, University of Cambridge. He is also visiting Professor at the Ulster University and Chairman of the Advisory Board of the Ulster University Economic Policy Centre.
Jordan Buchanan is an Economist at the Economic Policy Centre at the Ulster University.
“Although the UK government has steered away from further work on economic assessments of Brexit, devolved governments have felt less constrained.
The Mayor of London, reacting against the UK Government’s reluctance to publish assessments”
Um! The Mayor London (London Mayor) is not a devolved government he’s a public service employee
One problem with the arguments over Brexit is how technical they can get, so that the man in the street is left bewildered, and follows emotional arguments from political tub-thumpers.
Even though I have a degree in economics from LSE (40 years ago, mind you!), I struggle to grasp all your propositions. I leave that to better and more up-to-date “experts” (are they allowed to take part, pace Michael Gove?).
One small point. You refer to “flawed Treasury and CEP economic models”. I presume you mean here the Centre for Economic Performance at LSE?
CEP – yes they do. There is a link to their full paper in the article. It is well worth a read it gives a critique of all of the main Brexit papers released or rather a critique of the assumptions used as inputs to the models.
The individual if so interested can make their own mind up as to whether they believe the inputs to be correct or not.
Our conclusion is that in order to restore public confidence in economic forecasting for major policy issues like Brexit, economists need to use more relevant analyses, based on a wider range of evidence. We expect that econometric models used by commercial forecasters like Cambridge Econometrics, will prove to be most accurate in the long-run. If so, the academic profession needs to reconsider both the relevance of its current attachment to theory based on unrealistic assumptions, and to the general quality of policy-relevant applied work.” A worthy but anodyne paragraph that few would dispute but not very illuminating.
Much worse is: “Our conclusion is that most estimates of the impact of Brexit in the UK, both short-term and long-term, have exaggerated the degree of potential damage to the UK economy. ” Do you (the authors) not fall into a trap similar to the one you accuse others of – minimal evidence in order to proclaim a wished-for outcome?
Surely you can do better than this?
On the other hand – https://infacts.org/brexit-impact-analysis-wildly-optimistic-non-eu-trade/
You pays your money and takes your choice! To me all logic says we will suffer pretty badly from leaving the single market and the customs union.
Your linked article written today states “The US is more likely to prioritise the EU for a trade pact” than the UK.
What has the author missed about Trump and his sanctions?
I am not really sure that this is primarily about Brexit. In so far as I can read what is going on in this is that this is just “crossfire” in a dispute between some neo-keynseian economists and ideas they do not agree with in mainstream economics. They may have an interesting perspective but I do not think they have been very successful in explaining its relevance and why people should favour their model and conclusions.
The title seems sensationalist, which does not seem constructive in the present political climate. With a title like this you would expect it to be supported by a very clearly explained and powerful logical argument backed up by a massive body of evidence. This may exist, but if it does, it was not very effectively communicated.
It would be useful for readers if the authors just stated their own predictions in ways that are directly comparable with those of others. i.e answer in simple terms . “What will be the economic impact of Brexit?” Are they predicting a Patrick Minford type of outcome or something more in line with the vast majority of other economists?
I am not particularly impressed or surprised that the immediate impact of the vote was not the economic collapse that some predicted. It was rational to fear that market panic was a real possibility but it is not a surprise that we avoided it. The announcement that we were leaving the EU had only one possible direct effect. An effect on confidence in the UK economy. This impacted the value of the pound and businesses have slowly started to respond in making changes to their trading arrangements one way or the other. What was really feared was that the announcement would cause a market panic. This did not materialise but that should be no surprise. Since market speculation is now a major part of how markets function it becomes much less predictable whether markets will panic or not. Speculators may well indulge in brinkmanship and hang fire gambling that that the market will not collapse immediately when all the evidence shows a need for an adjustment. Governments will often support markets if market panic starts and this can be a source of profit for speculators. I do not think that economic theories explain the unpredictability of markets, it is often just about speculation. On the other hand I believe that when the reality of the economic changes that Brexit will bring about become more immediately imminent the speculators will take their profits and re-value UK assets at a more realistic level in line with the change in the UK’s long term prospects.
It is easy enough to download per capita gdp data for the OECD countries from the World Bank website.
Then plot UK vs various countries in the periods before and after 1973. Excel will even add the trendlines.
I have done this exercise here: https://twitter.com/mac_puck/status/1123537360606629889
It shows that Gudgin, Coutts and Buchanon’s analysis is simply wrong. Badly, blatantly, suspiciously wrong.
It is funny how Gudgin lists his roles in Cambridge and Ulster, but doesn’t mention his role at the Policy Exchange (a right-wing, opaquely-funded “think tank” with heavy pro-Brexit bias)