Evidence-based policy or fake news? Eleven Economic Propositions made by Remain

fake news

The article by a Senior Economist working in the UK private Sector is a longer version of the article ‘Evidence-based policy or fake news?’ which appears on our BLOG page. The report examines ten widely believed facts about the economic impact of Brexit and finds all of them wide of the mark. In some cases they are hugely wrong.

The UK government is still apparently undecided about what kind of Brexit ‘end-state’ to pursue, a state of paralysis that in our view has much to do with the nature of ‘official’ opinion on some of the key economic questions posed by Brexit.

We believe much of the analysis behind official opinion on the economics of Brexit is flawed, based on incorrect methodology or cherry-picked evidence.

In this article we confront eleven key propositions that the Remain ‘camp’ and UK ‘official opinion’ have advanced to try to push the UK towards a so-called ‘soft’ Brexit that would leave the UK as part of a customs union with the EU and/or part of the EU single market (de facto or de jure).

We find that ‘official’ opinion has been shaped by estimates of key ‘costs’ that are greatly exaggerated, in many cases by a factor of 5-10 compared to plausible numbers.

Proposition 1: the UK economy is hugely dependent on trade with the EU

While the UK is a relatively open economy, it is not hugely dependent on exports to the EU. In fact, of all EU members the UK is the least dependent on trade with EU partners.

Around 47% of UK goods exports are dispatched to EU destinations and this number is probably overstated by 2-3 percentage points by the so-called ‘Rotterdam effect’ – UK goods being shipped via the Low Countries to destinations beyond the EU. For services exports, a large and increasing part of UK exports, the EU share is around 40%. This gives an overall UK export share to the EU of around 43%[1].

Moreover, the value of UK exports of goods and services to the EU amounts to only around 12% of UK GDP. Even this is an overestimate. GDP is a value-added measure (i.e. it excludes the value of intermediate inputs to production) and so we should compare the UK value-added generated by exports to the EU with UK GDP. Adjusting for this we estimate the share of UK value-added resulting from exports to the EU is less than 10%, so 90% plus of UK GDP is not generated by exports to the EU[2].

In addition, the UK economy grew faster before we joined the EEC/EU. Even if one uses the USA as a benchmark, to control for changing global conditions, there is no indication that economic growth in the UK has improved through membership of the EU.


Proposition 2: EU membership has massively increased UK trade with the EU

The Treasury’s 2016 study claimed that EU membership had increased UK exports of goods to the EU by 115%, and services by 25%, making a 78% total rise above what they would ‘otherwise’ have been.

These estimates look greatly exaggerated, by a factor of 4-5.

  • They are not consistent with historic UK trade data. In the years 1969-72, just before the UK joined the EU, UK goods exports to what are now the EU countries were around 40% of total UK goods exports, and 6% of UK GDP[3]. These numbers are below current levels of 47% and 8% respectively – but certainly are not less than half of them as the Treasury study would imply! Part of the rise will also be a general globalisation effect.
  • The estimates are based on a flawed application of ‘gravity model’ methodology. As Gudgin, Coutts & Gibson (2016) have shown, the Treasury’s numbers for extra UK exports generated by EU membership are based on estimates for the ‘EU effect’ averaged across all EU members. This is a major source of upward bias in the Treasury estimates because the UK’s trade with other EU countries is much lower than the EU average. Gudgin et al. estimate that for the UK alone the increase in goods exports to the EU generated by EU membership is more like 23% – about a quarter of the Treasury estimates.
  • An earlier unpublished Treasury paper from 2003 showed even weaker effects of EU membership on trade. Again using a gravity equation, UK-EU trade was found to have increased by just 7% as a result of EU membership, partly offset by negative trade diversion effects of 4%. The paper also posited a further 9% rise in intra-EU trade from the single market programme, but this was not a UK-specific estimate and even if it were, the total net positive impact on UK trade would still be lower than that estimated by Gudgin and Coutts and only around a fifth of the Treasury’s 2016 estimate!
  • The evidence for a large ‘EU effect’ on UK exports of services is shaky. Services exports would not have benefitted from the UK joining the EU customs union in 1973 as this reduced barriers only to goods trade. Conceivably there might have been benefits from moves towards a single market in services in the 1990s and 2000s and (more speculatively) from free movement of people within the EU. But big positive effects of EU membership on services trade are hard to discern. The share of UK services exports sent to the EU has been steady at about 40% since the early 1990s. Gravity model-based studies by Walsh (2008) and Dettmer (2014) struggle to find any statistically significant evidence of a positive effect of EU membership on services trade.


Proposition 3:  Outside the EU, UK exporters will face big barriers to trade with the EU

It is an article of faith in the Remain camp that outside the EU, UK firms will face massive barriers to trade with the EU, which might even prevent trade entirely in some areas.

There will certainly be an increase in trade barriers, but the extent of the potential increase has been hugely overstated.

Tariffs: Based on current trade patterns, UK exporters would face a weighted average tariff rate of 3-4%. Around 30% of products face zero tariffs under the EU’s WTO schedule, including important UK export sectors such as pharmaceuticals, oil, and most iron & steel products. Under the WTO agreement on civil aircraft, UK exports of civil aircraft and parts would also be duty free. Overall 35-40% of UK exports to the EU would be duty-free even in a ‘WTO’ scenario, with the bulk of the remainder facing tariffs below 5%. Only about 15% of UK exports would face a tariff of 10% or more and only around 5% (mostly agricultural products) would face a tariff of 15% or more[4].

It should also be noted that on the import side, UK exporters can and would use inward processing relief to avoid tariffs on components. This system allows import duty to be remitted on parts imported that are later used to produce a good for export. A large share of UK imports would potentially be covered by this in a ‘WTO scenario’.

Non-tariff barriers (NTBs): ‘Official’ opinion emphasises the importance of NTBs rather than tariffs as a future barrier to UK exporters, but greatly exaggerates them, perhaps by a factor of 5-10.

NTBs cover a wide range of barriers including customs/certification/administrative costs, technical barriers to trade (e.g. relating to product standards, labelling etc.) and, in free trade agreements, so-called rules of origin costs. The latter relate to costs incurred proving that a good originates from a given destination, or does so to a sufficient extent to obtain preferential tariff treatment.

There is no question that NTBs are a real phenomenon and that they are frequently more of a problem than tariffs, especially for trade among advanced economies. But estimates of the likely future NTBs faced by UK exporters from some ‘official’ sources and industry lobby groups look much too high. Some such estimates, including those by the Treasury, suggest UK exporters will face NTBs equivalent to a tariff of 20-30% (often with reference to estimates of NTBs between the EU and US).

But detailed work by Kee & Nicita (2017) finds that the average tariff equivalent of NTBs facing UK exporters to the EU, weighted by imports, would be 3.4% in a ‘WTO scenario’.  These econometric estimates can be supported by real-world examples. A detailed study of post-Brexit NTBs for several sectors in the Netherlands, using actual cost estimates for administrative and other processes, found NTBs might total around 1% of trade values (including customs costs).

Moreover, some NTBs relating to product standards are unlikely to be a problem for existing UK exporters who already produce to these standards and have done for many years. In this respect they are in a different position from, for example, a US exporter attempting to enter the EU market and finding the necessity to produce goods to a set of standards different from US (or in some cases, international) ones. This is why using estimates of NTBs between the US and EU as a proxy for NTBs that might be faced post-Brexit by UK exporters is a flawed approach.


Proposition 4: UK exports will collapse in a ‘WTO Brexit’

Having claimed that EU membership increases UK exports to the EU by some 80%, the Treasury in 2016 went on to claim that all this ‘excess’ trade would disappear after EU exit, so that UK exports to the EU would fall by 43% and total exports by 24%.

These estimates look much too high, being based on unrealistic assumptions:

  • The estimated ‘excess’ trade is an exaggeration (see above)
  • Assuming that all ‘excess’ trade will unwind is an extreme assumption. Patterns of trade formed over several decades will tend to linger given relationships between firms and established facilities and capital equipment.
  • The implied trade rise in trade barriers needed to produce such a colossal fall in UK exports to the EU is implausible given realistic estimates for these (see above) and the size of trade elasticities. Indeed, Kee & Nicita note that EU imports from the UK are biased towards products with a low elasticity of demand, which will reduce the negative impact of trade barriers on exports.
  • Official sector estimates assume no change in exchange rates.

Estimates of the decline in UK exports to the EU based on more realistic assumptions are far lower, sometimes in low single figures.

Kee & Nicita estimated that the imposition of WTO tariffs on UK exports to the EU would cut UK exports to the EU by only around 2%. Even doubling or tripling this for the impact of NTBs would leave the estimated fall in UK exports to the EU at between one-seventh and one-tenth of the Treasury estimate.

If we also assume some exchange rate depreciation, the impact may be even lower. Lawless and Morganroth (2017) found that a 10% decline in sterling would entirely offset negative impact of tariffs and NTBs on UK exports to Ireland.


Proposition 5: long-term economic losses from leaving the EU single market and customs union will be very large

The Treasury’s 2016 document and the more recent ‘cross-Whitehall briefing’ claimed UK GDP would fall by 8% in the long-run in the event of the UK leaving the EU and reverting to ‘WTO rules’. This estimate is probably 3-8 times too high.

These numbers fail to pass the most elementary ‘sniff test’. As we have already seen above, less than 10% of UK GDP comes from exports to the EU. The Treasury is asking us to believe that, essentially, all this GDP would vanish – permanently.

But this is not realistic. What will actually happen in the case of a decline in trade with the EU is that some exports to the EU will be diverted elsewhere, including to the home market, and some imports from the EU will be substituted by imports from elsewhere and by domestic production.

There may indeed be costs from this, but as Paul Krugman has recently shown , the economic welfare losses from declining trade can be surprisingly modest. As Krugman notes, if increased trade barriers drive up import prices for consumers, the welfare loss can be approximated by the formula: loss=fall in imports*1/2 tariff rate*share of imports in GDP.

So what would this loss look like in the case of the UK moving to ‘WTO-based’ trade with the EU? UK imports from the EU are about 15% of GDP. If we assume that trade barriers rise by 4% (for tariffs) plus 10% (for non-tariff barriers)=14%, and that this leads to UK-EU trade falling by 25%, then using the formula above the welfare loss would be ½*14%*0.25*15=0.3% of GDP.

Even if we plug in much more lurid assumptions, the welfare loss is much lower than the Treasury claims. With a rise in trade barriers of 25% and a drop in UK-EU trade of 50%, the welfare loss is still less than 1% of GDP. So even using very pessimistic assumptions, estimated welfare losses from declining trade with the EU are perhaps one-eighth of the Treasury’s claims!

A more sophisticated approach uses ‘computable general equilibrium’ modelling to estimate the impact of rising trade costs. But studies based on this approach also fail to generate anything like the Treasury numbers.

Ciuriak et al. (2017) used the same modelling approach as the ‘cross-Whitehall briefing’, and found that even in the most severe ‘WTO rules’ scenario, UK GDP only fell 2.5% in the long run – less than a third of the Treasury results.

The LSE’s Centre for Economic Policy agree that tariff and non-tariff barriers on a WTO basis would have a small impact – just 0.1% of GDP for tariffs and 1.3% of GDP for non-tariff barriers[5].

How do we get from these sorts of numbers to GDP impacts of -8% as posited by the Treasury? The answer is twofold: first, by assuming trade barriers that are implausibly high (see above) and second, crucially, by making a series of further extreme negative assumptions:

  • No trade reorientation – the loss of UK exports to the EU is not offset by any increases to other locations.
  • Sharp declines in UK productivity – the decline in trade, or in FDI, or both, is posited to have big negative effects on UK productivity.
  • Subtracting hypothetical future gains from further EU integration

Importantly, the evidential bases for these large negative add-ons are weak. There will of course be some trade re-orientation and there is no clear correlation between trade and GDP levels for advanced economies[6]. The dubious basis of including hypothetical future EU integration as a Brexit loss is self-evident.


Proposition 6: The EU single market has given UK industry a big boost

The EU single market for goods came into existence in the early 1990s and much was made at the time of the way it would boost industry by improving the prospects for specialisation and realising economies of scale.

Unfortunately, the evidence does not support the notion that UK industry has received such a boost. Since 1992, UK manufacturing output has risen by just 12%, or a paltry 0.4% per year. This is somewhat slower than the rate of growth of the previous quarter century – even though the latter included the massive industrial recession of the early 1980s.

If we look at a more recent period, the record is even worse; UK manufacturing output in 2017 was essentially at the same level as in 1999[7]. The manufacturing sector, so loud in lobbying, has been a massive underperformer – contributing essentially nothing in accounting terms to the UK’s relatively strong overall growth performance of the last two decades.

Nor is there much evidence of a productivity surge related to the single market; since 1992, UK manufacturing output per job filled has risen 2.3% per year but this compares to a pace of 4% per year in 1979-1992. Experimental multi-factor productivity series show the same picture – a marked slowdown in manufacturing productivity growth since 1992 compared to the 1970s and 1980s[8].

The geographical pattern of UK export growth, meanwhile, contradicts the notion that the EU single market has been a dynamic, extended ‘home market’ for UK businesses. Since 1998 (when the data begin), UK exports to the EU have grown by just 1.5% per year[9], less than half the pace of UK export growth to the rest of the world.

Given the biases towards EU trade that exist via the customs union and single market, this underperformance is remarkable – the EU has been a very sluggish market for UK exporters who have increasingly looked for opportunities elsewhere.


Proposition 7:  Outside the customs union, customs processes will be expensive and time consuming

There are several strands to this argument, and again high levels of exaggeration are present.

  • Customs costs. In June, HMRC claimed the cost of running standard customs processes (they referred to these as Maximum Facilitation customs processes but it is clear that is not what was being considered) for UK-EU trade could total £20 billion. This would be equivalent to around 1% of UK GDP or about 6% of the value of UK exports to the EU.
  • Customs delays. Industry lobby groups and others have claimed that outside the customs union there could be long delays at borders for UK exports due to border checks.

HMRC’s estimates for the costs of customs processes are deeply flawed and likely to be over-estimated by a factor of at least six and possibly as much as 20 times.

As previously noted by Briefings for Brexit, the HMRC estimate adds together potential costs on the UK and EU side and implies this is a UK cost; it utilises inflated estimates of rules of origin costs (see below); it appears to confuse estimates of costs of customs declarations per container with cost per consignment (where consignments will often be much smaller); and it assumes no behavioural changes by business such as bundling up small consignments into larger ones.

Looking ahead, the estimate also ignores the opportunity frequently trading businesses will have under new customs rules to self-assess their customs liabilities with quarterly reporting – which would massively reduce the number of individual customs declarations.

The HMRC estimate also flies in the face of international evidence which shows much smaller customs costs. Swiss customs estimate the cost of running their systems (which are outside the EU customs union) at 0.1% of GDP – a tenth of the HMRC estimate. KPMG estimate customs costs for Dutch customs at less than 1% of the value of consignments, less than a sixth of the HMRC estimate. Businesses such as Tate & Lyle have meanwhile produced estimates of their customs compliance costs (including staffing) at, at most 0.04% of the value of imported products[10].

Claims about border delays are also exaggerated, with much of the commentary on this issue not relating to the way modern borders work. Pre-clearance of most consignments is normal and the overwhelming bulk of ‘paperwork’ is conducted electronically.

Data from customs services across Europe shows that border checks are infrequent on goods arriving from outside the EU. UK customs check around 4% of consignments arriving from outside the EU (overwhelmingly agricultural products) and Irish customs just 1%. French customs reports 95% of customs declarations are released in less than five minutes.

This evidence renders highly dubious the popular argument relating to the supposed burden of checks or tests at the border relating to whether goods conform to EU product standards. In reality, such checks and tests are sparse and concentrated on produce from high-risk countries and foodstuffs. For the great bulk of products, under the EU’s New Approach, manufacturers can self-certify their goods as EU-compliant, affixing the CE mark – including manufacturers outside the EU.

Proposition 8:  Outside the customs union and single market, complex supply chain industries will decline

Much attention has recently been given to the way just-in-time manufacturing systems might be disrupted by Brexit. In particular, it has been emphasised that components criss-cross boundaries in modern UK-EU supply chains. This is of course correct. However, this phenomenon is not unique to the EU, or indeed to customs unions.

A good example of this is the NAFTA trade agreement that covers the US, Canada and Mexico. This is not a customs union and has border controls. Yet, as the FT noted recently, the NAFTA pact has spawned ‘integrated supply chains’ in which ‘components cross the world’s busiest border multiple times as they are transformed into finished goods’, especially in the auto industry.

This is exactly the kind of supply chain structure that we are told cannot exist outside of a customs union. OECD value-added data confirm the FT is correct; for the motor vehicle sector, 35% of the value of US exports is imported content, and 50-60% of the value of Mexican and Canadian exports[11].

This, plus the evidence above on customs checks makes recent suggestions that the UK should continue to align fully with EU product standards to reduce border checks appear redundant. Any ‘reduction’ will be minimal as average wait times are so short and such a small part of EU imports from the rest of the world are tested or checked now.


Proposition 9:  ‘Rules of origin’ costs are high, making free trade agreements inferior to a customs union

This notion has been popularised recently by a study by the IoD, which purports to show that rules of origin (ROO) are so costly as to greatly reduce the value of a UK-EU free trade deal (and indeed other free trade deals).

Rules of origin (ROO) costs relate to the administrative costs of proving products originate in the UK rather than in other countries and feature in free trade agreements. ‘Official’ opinion plays up the importance of ROO in two ways:

  • It is claimed the costs of paperwork and other compliance with ROO is very high, so much so that a lot of exporters don’t bother to go through the process at all if they face small tariffs as the cost saving isn’t worth it. HMRC recently cited estimates that the compliance costs could be as high as 7-8% of the value of trade
  • It is claimed for some industries, e.g. cars, a low share of domestic value added in the final product will disqualify them from getting preferential treatment under a future UK-EU FTA so they would face high tariffs.

Concerning (i) above, if compliance costs for ROO were really as high as HMRC claims, then it is hard to see why many FTAs would be signed at all given how low average world tariffs now are – especially among the advanced economies.

We should indeed expect to find that the utilisation of existing FTA tariff preferences by exporters is low. But this is not the case: recent studies suggest utilisation of tariff preferences in FTAs is high, often over 80% and as high as 90% in a WTO study for firms in North America and the EU.

In our view evidence suggests that the HMRC estimate that ROO for the UK in a free trade deal with the EU would be up to £7bn is likely to be at least four times too high.

The WTO find that 80% of imports eligible for tariff preferences in the EU and US, and for which the preference margin is less than 1%, nevertheless enter under preferential tariff regimes. Moreover, the WTO also suggest that rather than being a recurring cost, ROO costs are likely to be largely fixed in nature. The WTO suggests fixed costs in the range of US$14-1500 – negligible for large firms. The WTO findings strongly suggest that the real costs of ROO compliance are less than 1% of trade values, and may often be negligible.

Concerning (ii) above, details of ROO vary but on average if a product can be shown to have ‘local’ content of around 45-50% it passes the test. Moreover, it is often the case in FTAs that the parties to the FTA allow ‘cumulation’ so that content originating in one of the FTA partner countries is also counted as local.

On this basis, the great bulk of UK exports to the EU are not going to have a problem meeting the thresholds for ‘local content’ and so qualifying for zero tariffs in a future FTA. About 64% of value added in UK manufactured goods exports in 2011 was local content, and about 55% even in the car industry[12]. With EU content likely to be around half the foreign content, the UK+EU content totals would be around 80% in both cases.

It is still possible that some firms will have problems and need to readjust their production patterns, but given these industry-wide figures, the issue would not be a large one. A potentially bigger issue is for firms exporting not to the EU but to countries currently covered by FTAs the EU has negotiated such as Switzerland or Korea. If EU content is no longer considered ‘local’ some UK exporters to these third markets might lose tariff preferences. However, this issue relates to a minority of firms in a handful of sectors exporting to countries that account for about 10% of UK exports

Proposition 10:  Free trade agreements won’t generate a significant rise in UK exports

Official opinion has also been keen to downplay the possible positive impact of future free trade agreements the UK might sign with partners outside the EU. The leaked ‘cross-Whitehall briefing’ claimed an FTA with the US would add only 0.2% to UK GDP, which is at the bottom of academic estimates. European Commission estimates for the impact on the UK of the proposed EU-US TTIP deal suggested UK GDP could rise by 0.4% and UK exports to the US could rise by about 20%. The Bertelsman-IFO study suggested much higher gains.

Some recent evidence following free trade deals also points to potentially significant gains for UK exporters. Of particular note is the EU-Korea FTA, which came into force in 2011. In 2017, UK exports to Korea reached £6.5 billion – almost triple the level in 2010. Looking at the US-Australia free trade deal which came into force in 2005, we also see large effects on bilateral trade. In 2015-17, Australian exports to the US were 63% higher than in 2002-04, and US exports to Australia 78% higher[13].

Proposition 11:  The UK financial services sector is heavily reliant on ‘passporting’ for its revenues

Financial services is one of the UK’s most successful export industries and official opinion and some industry lobbyists have been keen to play up the notion that this industry is heavily reliant on the UK’s EU membership for its success. In particular, much has been made of the supposed importance of the EU financial ‘passport’ for selling financial services across the EU.

In fact, the evidence suggests the passport is of rather limited importance for UK-based financial institutions. An earlier paper by Briefings for Brexit showed that measured trade barriers in services between the UK and EU had changed little since the early 2000s, despite the various EU directives that created passporting.

This should not be too surprising because the great bulk of UK financial services output is wholesale in nature (including trading, cross-border lending to large firms, M&A etc.) and not much affected by EU membership. The City’s exposure to EU client bases (retail, small businesses, private banking) that will need EU servicing when the UK is outside the financial passport is quite small. For many financial activities where EU regulations do impinge, there are also low-cost work-arounds as shown in Reynolds (2017).

This explains why early predictions of mass job losses in the UK financial sector after Brexit have not materialised. The announced job relocations to date of a few thousand employees are a mere 5% of some of the more lurid early claims.

[1] Author’s calculations based on UK Pink Book

[2] Author’s calculations based on OECD

[3] Author’s calculation based on IMF Direction of Trade Statistics and UK national accounts

[4] Author’s calculations based on WTO

[5] Dhingra, S., Huang, H., Ottaviano, G., Pessoa, J. P., Sampson, T., and van Reenan, J., (2017), The Costs and Benefits of Leaving the EU: the Trade Effects. CEP Discussion Paper No 1478 April.

[6] This link is derived in the ‘official’ estimates largely with reference to one outdated study mostly looking at emerging countries.

[7] Author’s calculations based on ONS data

[8] Author’s calculations based on ONS data

[9] Author’s calculations based on ONS data

[10] See Gudgin & Mills ‘Customs costs post-Brexit’ Briefings for Brexit

[11] Author’s calculations based on OECD value-added trade database

[12] Author’s calculations based on OECD value-added trade database

[13] Author’s calculation based on IMF Direction of Trade Statistics

About the author

Briefings For Brexit