Several institutions have published model-based forecasts about the economic consequences of Brexit. The forecasts range from a positive impact of 1.4%, to a substantial negative cost of 9.3% of GDP. The range may seem to shed more darkness than light. But by looking at the different assumptions we can pinpoint exactly why institutions differ in their forecasts.
Model-based forecasts have advantages. First, they are simplifications and internally consistent. Some may be better than others for answering specific questions; but, they are coded and logically consistent. Second, they force modellers to be explicit about their assumptions; and this often reveals the prior beliefs of the modeller. Models are, of course, simplifications and leave out many of the interesting non-linear relationships in the real world.
The table below summarises the headline forecasts. Most institutions present at least four different forecasts; two post-Brexit trade scenarios and over the short and long term. The trade scenarios are that the UK strikes a free trade agreement or relies on its World Trade Organisation (WTO) ‘Most Favoured Nation’ status to continue trading with the EU.
A free trade agreement is the less restrictive regime with no barriers for trading goods but access to limited service sector markets. It would also have to be agreed with each of the 27 remaining EU members. Under WTO rules, the usual goods tariff is around 3% (higher on some goods) but again with limited coverage in the services sector (see my blog on services trade here). It is certainly worth noting that services trade has more value added (i.e. generates more net income for the UK) than exports of goods.
Table. Comparison of forecasts of the economic impact of Brexit on the UK
Download Comparing Brexit forecasts table
All of the institutions in the table make two assumptions about the UK’s post-Brexit relationship with the EU. First, the UK would make zero or a lower contribution to the EU budget (depending on market access). Second, there is a change in trade relationship between the UK and EU.
Here the key assumption is what happens to the trading arrangements. In particular, to the services trade arrangements, although this is not always clearly laid out by the forecasters. While trade in goods is mostly regulated by tariffs, services are regulated by non-tariff barriers such as equivalence to a single set of standards, regulations and rights of establishment that are not included in most trade agreements.
Explaining the difference between forecasts
Two obvious results need to be explained. First, why do Economists for Brexit (EfB) find a positive impact while all other institutions find a negative impact? This comes down to the assumed trade relationship between the UK and EU after Brexit and how EfB choose to model this.
EfB assumes that the UK unilaterally cuts tariffs on all imported goods. They then crucially argue that if the UK faces tariffs from the EU on exports they could simply divert trade elsewhere. Moreover, they argue that the markets for services in the EU is heavily regulated and that after Brexit these restrictions and so the UK’s share of services trade will be broadly unchanged despite leaving the EU.
They also model the import tariff cut by a reduction in expenditure tax (e.g., VAT). The former would be bad for the protected domestic industry sector, whereas the latter is good for both the protected and unprotected domestic sectors.
By contrast, all of the other institutions in the table assume that the UK would have less trade with the EU as a result of Brexit. They mostly justify this assumption by looking at how countries that joined the EU subsequently had greater bi-lateral trade. They then assume a degree of symmetry: if joining increases trade, then leaving reduces trade. This may give the right sense of direction but the estimates are imprecise because no country the size of the UK has ever left the EU.
The second result that needs to be explained is why some institutions forecast a much bigger impact of Brexit than others. These are summarised in panel 2 of the table. In short, as the number of transmission mechanisms increases then this tends to lead to a greater overall impact. However, the greater the number of mechanisms, the more difficult the incremental mechanisms are to justify. For example, explaining productivity differences is difficult at the best of times.
It is also notable that for all of the political heat around migration, only three of the forecasters include any change in migration policy after Brexit in their forecasts. The reason is partly technical. In ‘representative agent’ models more or less migration simply scales up or down the model; there is no real economic impact on households unless a more sophisticated mechanism is included.
By Dr Angus Armstrong, Director of Macroeconomics at the National Institute of Economics and Social Research.