Blog Featured

The IMF’s latest no-deal warnings are sloppy and overblown

Written by Harry Western

The International Monetary Fund has produced another analysis suggesting that a no-deal Brexit would lead to steep short-term declines in UK output. As before, its claims look exaggerated with some elements of their work not standing up to serious scrutiny. The IMF estimates are based on a host of highly pessimistic assumptions, some of which are out of date while others are not supported by academic evidence or real-world examples of how trade works. Worse still, some of the numbers the IMF uses – for example those related to ‘border disruptions’ – appear to be plucked from thin air. Claims about the effects on the financial sector are based on sloppy analysis and look utterly implausible.

In its April 2019 World Economic Outlook, the IMF has produced an analysis of the economic effects of a ‘no deal’ Brexit. Predictably, it is very negative and predictably it is based on a host of highly pessimistic assumptions. Parts of the analysis are downright sloppy and in some cases the IMF can justifiably be accused of simply plucking figures out of thin air.

The IMF produces two scenarios, a scenario A with no border disruptions and only limited financial market effects and a scenario B with large-scale disruptions and more severe financial effects. Both end up with UK GDP being around 3.5% below the IMF’s baseline forecast by 2021, with a temporary and deeper initial decline in output in scenario B. Their estimates imply full-year recessions in the UK in both 2019 and 2020.

What kind of assumptions are needed to generate such massive effects on UK GDP?

Trade barriers: the IMF assumes that the EU applies its standard tariffs (which average 3-4%) to UK exports straight away, with the UK mirroring these tariffs after a year – for the first year the UK is assumed to impose only its relatively liberal no-deal tariff schedule recently published.

Non-tariff barriers to trade (NTBs) equivalent to 14% of trade values also appear. This is on top of 10% NTBs the IMF already assumes will materialise in its baseline forecast – so the IMF is claiming total NTBs facing UK exporters to the EU after Brexit will average a colossal 24% of trade values.

  • Our take: the IMF’s assumption that the UK will ditch its relatively liberal no-deal tariff schedule and opt for increased protectionism after a year is arbitrary and unjustified.
  • IMF claims about the likely scale of NTBs are extreme. Detailed work by Kee & Nicita (2017) shows that the average tariff equivalent of NTBs facing UK exporters to the EU, weighted by imports, would be just 3.4%. Other recent model-based studies also have NTBs in a 3-8% range. A detailed study of post-Brexit NTBs for several sectors in the Netherlands found they might total only around 1% of trade values.
  • With UK and EU regulatory systems fully aligned at the moment of Brexit, claims of massive NTBs are even harder to justify. Many UK firms have also completed work-arounds to avoid such NTBs including switching testing and certification to EU-based bodies. The Bank of England’s latest survey showed 80% of firms were as ready as they could be for a no-deal Brexit.

Trade elasticities: the latest IMF study does not say explicitly what effect the rise in trade barriers it assumes will have on UK exports. But their previous work does, as it presents estimates of the price ‘elasticity’ of demand for UK exports to the EU. These estimates are around 4 for goods and 5 for services, meaning that a 1% rise in the price of UK exports would cut demand by 4-5%. On that basis, the rise in trade barriers the IMF assume would lead to a huge decline in UK exports to the EU.

  • Our take: the elasticity estimates are extremely high and contrast sharply with many academic estimates that point to UK exports being relatively price inelastic. These include Kee and Nicita, Aiello et al., and Imbs and Mejean. Typically, such studies suggest price elasticities for UK goods exports of 0.5-1.5 i.e. between an eighth and two-fifths of the IMF numbers. For services, the Bank of England finds a very low degree of price sensitivity of UK exports.

Trade deals: the IMF assumes the UK loses ‘most’ third-country free trade agreements in place via EU membership, with these being replicated only after two years.

  • Our take: the IMF’s assumption on EU free trade deals is inaccurate and out of date. EU free trade deals currently cover about 9% of UK trade (not 15% as the IMF claim), and 11% if the Japan FTA is included – which will only abolish tariffs over several years. The UK has already agreed rollover deals covering 60% of the 9% of UK trade covered by these EU deals, which will rise to 70% when the Canada deal – believed to be imminent – is agreed.

Financial services: the IMF appears to be assuming incredibly negative impacts on the UK financial services sector. In their 2018 study, they claimed that UK financial services exporters would face NTBs of 50% of trade values, with their output down 25% in the long run compared to a baseline where the UK stayed in the EU.

  • Our take: the IMF’s numbers here are totally implausible. The IMF relates them to the loss of ‘passporting’ rights for UK financial institutions, but the great bulk of UK financial services are of a wholesale nature and so are little affected by passporting which relates more to retail financial services. For retail activities there are also low-cost workarounds.
  • The IMF’s estimates are based heavily on a consultancy report by Oliver Wyman which is now widely discredited. That report claimed Brexit would lead to 75,000 job losses in the City of London, but the latest estimates suggest a total of just 2000 roles have moved or been created overseas – a fortieth of the Oliver Wyman claims. Relying on this poor-quality report to calibrate such an important element of their work is sloppy practice by the IMF.

Financial market effects: in its more adverse ‘scenario B’ the IMF claims UK sovereign bond spreads would rise by 100 basis points and corporate bond spreads would rise by 150 basis points.

  • Our take: again, these numbers do not look plausible. In the aftermath of the 2016 referendum these spreads did initially show some ‘knee-jerk’ widening – but only for a very short period; the effect had unwound within a few weeks. The Bank of England helped this along with a modest programme of quantitative easing (including buying £10 billion of corporate bonds) and would have ample scope to repeat such an exercise if needed.

Border disruptions: the IMF’s ‘scenario B’ features ‘border disruptions’ that cut 1.4% of GDP in the first year and 0.8% off in the second.

  • Our take: here, the IMF appears to be essentially plucking figures out of the air. The numbers appear to be related to estimates the Bank of England presented in late 2018, which were treated with great scepticism at the time – including by Nobel Laureate Paul Krugman – as they had no obvious evidential base. To make matters worse, the Bank then abruptly halved its estimates a few months later – again with little or no quantitative evidence to back this up.
  • Claims of massive border disruptions ignore the very significant progress made in preparations by governments and businesses over recent months. These include new systems developed by Eurotunnel and the port of Calais to ensure smooth movement of goods, substantial expansion of facilities by the port of Rotterdam, the UK’s accession to the Common Transit Convention, new streamlined systems introduced for VAT and customs by the UK’s HMRC, the planned waiving of import duties on 87% of goods by the UK, UK-EU agreements on air and train travel and the UK’s recognition as a third country by the EU for animal and plant trade.
  • It is also not credible to assume disruptions would last for two years ­– behavioural changes and new agreements would mostly limit the timescale to a few months.

Policy responses: the IMF assumes UK interest rates are lowered in response to the negative impact of a no-deal Brexit, and that fiscal stabilisers are allowed to operate (so that the budget deficit rises ‘passively’). But no other policy responses, such as quantitative easing, are included in their analysis and the analysis also assumes the UK does not strike any new trade deals in two years.

  • Our take: ignoring the potential for the Bank of England to respond to any no-deal downturn with quantitative easing has no justification and looks like merely a convenient way of keeping the results a negative as possible. In practice, we know the Bank would do this, exactly as it did after the 2016 referendum result. Discretionary fiscal stimulus would be extremely likely too, especially given that the UK budget deficit has narrowed to less than 2% of GDP – there is plenty of room for it.

In conclusion, the IMF’s results are based on extremely pessimistic assumptions including trade barriers and trade elasticities that are as much as five times too high. The IMF has also added on massive ‘disruption’ effects and financial sector effects that have no proper basis in quantitative work at all.

With most policy mitigations and potential Brexit upsides also assumed away, we are left with a skewed and in places somewhat shoddy analysis. The IMF unfortunately has form in this area, having often ‘fitted’ its analysis to a particular political imperative. Its systematic exaggeration of growth prospects in countries receiving IMF funds – most dramatically, and embarrassingly, in the case of Greece – is a good example. Their latest Brexit analysis comes from the same stable of politically convenient analyses and is best filed under ‘ignore’.

About the author

Harry Western