The economics consensus about Brexit Patrick Minford is Professor of Applied Economics at Cardiff Business School and Chairman of Economists for Free Trade (EFT), a group of leading economists Why did the consensus of economists oppose Brexit? This is an important question because after this election once the negotiations with the EU begin, that same consensus will be rooting for a ‘deal’ that is as close as possible to the status quo. The May government has said that ‘no deal is better than a bad deal’ and committed itself to leaving the single market and the customs union, taking back its laws, and resuming control of its borders. If the EU makes any trade agreement conditional on acceptance of EU Single Market regulation and the ‘four freedoms’ which include free migration, then there will be no deal and that economics consensus will be in full attack mode again. So why? I have spilt a lot of ink1 in detailed examination of several consensus cases: especially that of the Treasury and of the LSE group, which was heavily consulted by the Treasury. But there were many others. A partial list includes PWC for the CBI, Oxford Economics for various business clients, the IMF, the NIESR and the OECD. I also reviewed these cases in less detail. I should mention Open Europe which produced an assessment ranging from a small negative to a small positive according to a variety of policy assumptions: this was the only modelling assessment, apart from ours, which was not uniformly hostile to Brexit and as such was not in the consensus. The consensus assumptions about Brexit policies There is one main reason for this consensus hostility to Brexit. These modellers all, without exception, assumed that under a full Brexit where there was no deal with the EU, often called the ‘WTO option’, the UK government would continue with existing EU trade barriers against the rest of the world and would not alter its regulative approach from that of the EU. Some modellers went further and assumed that the immigration policy would greatly reduce skilled immigration. Yet they made these assumptions knowing that they represented bad policy: continued protection against the rest of the world plus protection against the EU, no regulative improvement from well-known EU failures, and in some cases a bad immigration policy not currently followed on non-EU migrants. During the referendum, when challenged about making them, the various groups somewhat embarrassedly could not justify them but simply argued that this was ‘inevitably’ what policy would be for ‘political reasons’. Plainly these groups were campaigning against Brexit in the referendum and so had the motivation to paint a grim picture; now however when they like the rest of us must be concerned to make a success of Brexit, this position will not do, especially as the government is setting out policy positions that are quite different - namely free trade with the rest of the world, regulative reform and an immigration policy that will target unskilled immigration with its large costs to the taxpayer and enable skilled immigration needed to supplement our native skills. Since none of these modelling groups have to date costed such optimal policies we do not know what difference it would make to their assessments. However, we can get an indication from one solitary exercise carried out by the LSE group, in which they assumed the UK unilaterally abolished the EU average tariff on manufactured goods, around 3% on their calculations. They found that this would add 0.3% to UK GDP, against losses from the WTO option of 1-2.5% of GDP. LSE dismissed this as trivial, barely changing their overall assessment. Yet I have pointed out at various times that this is a massive underestimate of the EU protection we would abolish by moving to free trade with the rest of the world. My research team’s estimates of EU protection of manufactures is around 20%, made up of significant non-tariff barriers in addition to those low tariffs; these estimates are based on very detailed estimates of price differentials against world suppliers. They also tally with direct estimates of non-tariff barriers cited by LSE. But this is only the half of it. LSE also ignored the massive tariffs on agriculture which have the effect together with the rest of Common Agricultural Policy intervention of raising farm prices by 20% also; this is the standard estimate from the OECD of ‘Producer Subsidy Equivalent’. Together with the manufacturing total barriers these two pieces of protection raise consumer prices by around 8% according to our modelling. But even in the LSE model they multiply the gain from unilateral free trade by nearly 7 (20/3) which raises it to about 2% and more or less offsets the worst end of their range of losses from Brexit. What one gets out of this calculation is rather the conclusion of Open Europe using a model from the same general stable: the gains or losses under the WTO option could be quite small in either direction, depending on the choice of Brexit policy. It is interesting that as we move towards actual policy choices in this Brexit era the rhetoric from the consensus is becoming more muted. In a recent interview with Ambrose Evans-Pritchard of the Telegraph the leader of the LSE group, Prof. John van Reenen, candidly admitted that for a largescale regime change like Brexit he could not be too certain of the outcome. We can also note a variety of detailed ways in which the various groups calculated the losses from the Brexit departure have been criticised, taking their modelling approach as given. For example Ken Coutts and colleagues at the Cambridge Centre for Business Research have noted that the way the EU effect was estimated by the Treasury relies heavily on other countries’ experience and is unlikely to be accurate for the UK. In my original critique of the Treasury study I noted a number of similar points about estimation, as did David Blake. Much of this critique carries over to the other studies to a more limited extent. The general modelling approach of the consensus and the ‘gravity model’ However, this is all to take the general modelling approach of the consensus as correct in principle. We come then to the other major reason for the hostility of the consensus towards Brexit: the way that they think about the origins of trade. The classical view of trade, going back to Ricardo, John Stuart Mill and later Heckscher and Ohlin, is that it originates with differential supplies due to comparative advantage. Ricardo had wine and cloth in his famous example: Spain produced wine relative cheaply, Britain cloth. The later economists went behind these relative cost facts to argue that they in turn came from the supplies of ‘factors’ or resources native to each country, whether it was the ‘land’ (including the weather) or the skilled labour or the capital. These ‘resources’ would include such things as the climate for investment and property rights since these would enhance the effectiveness of the factors in production. We can think of this analysis as being a ‘supply-side’ theory of trade. Of course demand has a role in just how much is traded; but the broad pattern of what a country supplies and so exports after satisfying domestic demands compared with what it does not supply and so imports to meet home demand is set by supply forces. In these classical models if a country faces a trade barrier in another market it simply diverts any supply that is not sold there to another market, all markets being governed by competition across the world, so producing ‘world prices’. Now of course if you join the EU which is a vast customs union, you find it advantageous to switch your output entirely into that market wherever you can, since its trade barriers against the rest of the world will raise the prices within it above world prices. In the classical literature on customs unions, it was shown that while in general everyone would be better off with completely free trade, a particular country could gain from the customs union if it produced much larger amounts of the protected goods than it bought: this was because the gains to its producers from higher prices exceeded the losses for its consumers. Unfortunately the UK is in precisely the opposite situation: it buys much more of the protected food and manufactures than it sells! Hence in the last referendum my LSE mentor, Prof. Harry Johnson, vocally opposed the UK’s joining of the EU. Now contrast this with the view that dominates the current consensus - often known for short as the ‘gravity’ principle. What dominates trade is demand from countries that are large and also close by, because transport costs rise with distance (‘gravity’ equals size/distance). So the picture is of consumer markets in which a lot of different goods, with differing ‘brands’, are sold by producers from around the world supplying those brands. The prices of those brands are raised by transport costs the more distant the supplier. Consumers choose between them all and buy most from the cheapest sources. Clearly big countries have a lot of demand and as producer your share of those markets will depend on your price and so your transport costs. Tariffs and other trade barriers are simply one element in these ‘transport costs’. Now take a country like the UK: it sells exports to its trade partners according to how expensive its goods are in those markets. So if the EU puts up barriers to UK exports it will sell less. Carrying out FTAs with ‘distant’ markets is unlikely to replace the loss of sales to the close-by market of the EU. Now consider import trade: the UK buys from other countries according to its GDP and if it leaves the EU, its imports from the EU will go up in price while under FTAs its imports from the rest of the world will go down in price but overall its imports will not be changed much. So because exports will go down overall trade will fall. This demand-side effect is then compounded in these models by the assumption that foreign investment and also R&D depends on trade and so demand. The resulting fall in FDI and R&D then lowers productivity. So what we have is a model of trade and production that is dominated by demand. There is an interesting parallel between these current demand models of trade and the Keynesian models that once dominated macroeconomics after the second world war. In those too there was no role for supply. In both cases the suppression of supply came about because of the assumption of rigid prices and generalised lack of competition, ‘imperfect competition’, which underpinned this price rigidity. In current trade models competition plays very little role as the responses to relative prices are generally low. It is a curious fact of economic thought that just as Keynesian models were rightly displaced by models based on classical principles where supply is dominant except in the short run, their sister-models in trade came into fashion, even though their assumptions plainly contradict those of modern macro thought and appear totally unsuitable for the long run. In trade theory the focus is on the long run because trade regimes work themselves out over this time frame and the regimes last for a very long time: so it is particularly odd to find that demand is dominant over what amounts to a decade or more. If we revert to the obvious facts of world trade and competition, we can see that there is free entry into globalised world markets and that brands are frequently brutally displaced by competition: note the experience of Nokia and Blackberry, or the way in which Amazon has come to dominate retailing, or Google to dominate internet search. In whole areas of consumption such as the car market there is upheaval (‘disruption’) as the entry of electric and driverless, computerised, cars threatens to destroy the main car producer model. Then there is the steady grinding out of weak performers in worldwide supply chains. It has never seemed more appropriate to assume world competition, as in the classical model. Now think about the UK’s trade trends. The UK is now one of the world’s major export supplier of services; for example, in foreign currency trading it has the largest share of the world market. Back in the 1970s the City was a small part of the UK economy and of little consequence worldwide. Manufactures were 35% of UK employment, now only 8%. Have these changes had anything to do with demand from the EU or the world? Plainly not. They are the result of the UK’s own policies during the 1980s to transform the economy in the direction of free markets and competition. Workers left inefficient manufacturing industries, and new workers acquired skills whether in education or services and supplied competitive service products to the world economy. Capital flowed in from abroad after the abolition of exchange controls. Then consider how UK industries would respond to some EU protection. 60% of our car industry’s exports by value already go to world markets, outside the EU’s protected market. With the UK only 3% of the world economy it is really not difficult to imagine those exports being somewhat expanded either in existing or in new markets around the world, at world prices just like now. On this view the demand side is essentially irrelevant: supply will seek out demand where existing demand falls back. Which models are consistent with the facts? One rather amusing claim made by these gravity modellers is that their models must be right because they ‘fit the facts’ as found in the ‘gravity equations’. Yet what these modellers have done is fix (‘calibrate’) their models precisely so that they mirror these equations! These equations do indeed summarise certain facts about trade, as has been well-known since the 1950s when they were estimated by Jan Tinbergen. As we have seen demand will affect trade; and where supply has an effect it will be captured in these equations by ad hoc additions such as varying constant terms. The ‘gravity models’ are little more than a demand structure that produces these gravity relationships at a small remove. The gravity equations in no way ‘test’ these models by their consistency - since plainly they have been set up precisely to be consistent. Similar claims were made back in Keynesian times about the models fitting the facts. So indeed they did: modellers simply looked at the macro facts of consumption being associated with income for example and estimated models that gave back these facts. But what we learnt from brutal later macro experience is that these were not causal explanations of the facts; and so when policies or the environment changed so did these ‘fact’ relationships. We concluded that these Keynesian models were simply not proper causal models. In much the same way these gravity models are not at this stage full causal models. Full causal models need a full treatment of supply, competition and pricing for the long run. To test them we also need a much wider range of data than just the trade shares covered in gravity equations: for example the behaviour of a country’s production sector shares also has to be accounted for. Whereas we have now acquired the tools to test and reject proper causal macro models, these tools have yet to be applied in trade theory, partly because they have such a long-time perspective that long tracts of data are ideally required. However, such testing is overdue and we academics need to get on with it. While this academic process is going on, policymakers need a robust and plausible theoretical framework to make policy with. The classical model of trade and comparative advantage provides an approach that does explain the main trade developments of a medium-sized country like the UK and its assumptions of high competition and free entry in global markets are plainly in line with recent experience. What does such a model say about the best policy for the UK to pursue? The best trade deal is no deal - echoes of the Corn Laws The key element is the high rate of EU protectionism on food and manufactures, which erects a peripheral wall around the EU keeping up the prices of imports from the rest of the world and so raising prices to EU consumers for not just imports but all EU-made products competing with them. In both sectors the protective rate (from tariffs and non-tariff barriers) is as we have seen around 20%, raising UK consumer prices by around 8%. This in turn artificially boosts farming, the price of land and the inefficient parts of the manufacturing sector. By removing it with Brexit and going to free trade we reverse this and in the process raise consumer welfare and productivity, with a 4% boost to GDP. There are two routes to free trade: a negotiated route via free trade agreements, with the EU and then with significant others - and the route of unilateral elimination of our own protection, such as happened in 1846 when Peel abolished the Corn Laws. He got fed up with foreign recalcitrance over reducing trade barriers and simply struck out with unilateral free trade. We too could well get fed up with similar recalcitrance today as the mercantilist EU insists on special demands for its industries or its migrants and even other countries hold out for demands we cannot meet. The FTA route to free trade depends on others cooperating in genuine free trade. It might just work and go well. But realism suggests it could get bogged down and derailed. So suppose it falls at the first fence with no EU deal. What is the UK’s best option? It is to go unilaterally for free trade, with the gains described above. We simply say to the EU: look we abolish these barriers against you anyway and by implication under WTO rules we will do so against all others too. We so reduce consumer prices, increase competition and productivity and boost GDP. The EU would no doubt levy their tariffs on our exports; and other countries too would maintain their existing tariffs against us. But in a competitive world market where we are now selling at world prices, this has no effect on our national welfare. The reason is straightforward, as explained above: these world prices reflect world demand and supply and the EU tariffs we are speaking about do not affect the EU’s total demands and so do not affect world prices at all. All they do is cause EU demands to move towards home products away from us, but as they do so their home output is now not available in third markets where we will make up the deficit. The EU tariffs are noted earlier rather low - around 3.5% on manufacturing industry. Edgar Miller and I have estimated2 that they can easily absorb this cost in the short run when sterling is low and boosting their profits; and in the long run they can raise productivity to offset it. As for our farmers, they will after Brexit face world prices: protection of the CAP and high EU tariffs will be removed. They will sell on world markets for food instead of on EU markets where prices are artificially raised. So EU tariffs on our farming are simply irrelevant. We will revert to helping struggling farmers whose activities are necessary for the rural environment directly from the public purse. We have many large and efficient UK farmers who will change their practices, and adapt by raising productivity. So no deal is better than a bad deal. Indeed what the above shows is that no deal is better than any deal! But of course we will try to get a sensible EU deal in good faith, simply to maintain good relations even if it is not so good in pure economic terms. Endnotes 1. See the relevant publications on 2. ’What shall we do if the EU will not play ball?’ on There are two routes to free trade: a negotiated route via free trade agreements, with the EU and then with significant others - and the route of unilateral elimination of our own protection, such as happened in 1846 when Peel abolished the Corn Laws