27 April 2015

Short on detail on the huge question of Brexit

The first thing to note is the study is very short on detail. Open Europe’s comprehensive analysis and modelling of the potential impact of Brexit came in at 100+ pages. The Bertelsmann Stiftung (BS) contribution weighs in at 8 pages. Length is not indicative of much in of itself but Brexit is a complex issue and needless to say, this paper leaves a huge number of questions unanswered. Ironically, we are normally in favour of the academic German stiftungs being briefer and snappier in their analysis, sadly this is probably not the issue to do it on and the attempt seems to have largely backfired.

As BS itself notes:

The results of simulation calculations depend substantially on the underlying assumptions of how the UK would organize its relations with the remaining EU states and other trade partners after a Brexit.

This is undoubtedly true and applies equally to their own study. Unfortunately, they have left themselves little room to provide a full explanation of their numerous assumptions. Nevertheless let us take a closer look at their core scenarios.

Core scenarios show Brexit may not be the end of the world

The core scenarios are actually fairly reasonable and do not fall to far from what we found ourselves.

In their “soft exit” scenario (which is similar to our UK-EU FTA 1 scenario) the UK strikes a trade agreement with the EU similar to that currently held by Switzerland or Norway. It also maintains its free trade agreements with the rest of the world. This is not an unreasonable scenario and probably the most likely outcome. Our modelling put the cost at around 0.81% of GDP while BS puts it at 0.63%.

Their “deep cut” scenario is akin to our worst case Brexit scenario and sees the UK fall back on WTO rules and strike no agreement with the EU. We put the cost of this at 2.23%, while they strangely do not bother to say exactly what this comes in at. They do say that it is not as bad as their worst case scenario – “isolation of the UK” – which sees the UK also lose all of its Free Trade Agreements (FTAs) struck while under EU membership. This would cost the UK 2.98% of GDP according to BS.

BS does mention the potential EU budget savings of around 0.5% but it is not clear if these are actually included in any of the scenarios, the intimation is that it is not. If you apply this saving then the core range of the impact of Brexit is between -2.48% and -0.1% of GDP. Essentially, BS admits in an optimistic case (and one which many in the UK would see as the most likely outcome) Brexit could well be a wash from an economic perspective. Of course, this is all even before the UK has used any additional freedom to take any domestic policy decisions.

One sided assumptions and analysis undermine the study

The core scenarios seem superficially credible, though there is little to no detail on the exact modelling used or exactly how the results are reached. The study quickly loses credibility though when it pulls the potential 14% of GDP cost out of nowhere. This does not seem to be part of this study but of simply refers to other estimates:

“According to studies that estimate the influence of decreasing trade openness on the long-term real GDP (Freyer 2009 and Felbermayr/Gröschl 2013), a Brexit could lead to a long-term drop in the UK’s real GDP per capita ranging from 2% (“soft exit”) to 14% (“isolation of the UK”) compared to remaining in the EU.”

The 2013 study it cites seems to be about the GDP impact of natural disasters and makes no mention of Brexit or even the EU from what can see. It is also worth noting that the impact could be as low as 2%, though this receives no mention in the press release or anywhere else in the paper.

This is symptomatic of the one-sided approach taken by BS in this study. It is true that these types of studies stand on a number of assumptions. But the key is to test a variety of different scenarios to allow people to understand the policy choices they face but importantly to also test assumptions on both sides. Here BS only tries to examine one side – the down side – rather than giving even a cursory discussion to potential benefits or freedoms (beyond EU budget savings).

For example, BS assumes that the UK will not strike any new free trade agreements if it leaves the EU. While it is correct to note that such agreements “take time”, the model runs up to 2030. It seems very unlikely that the UK, a free-trading nation, would not make any efforts to strike FTAs over the course of more than a decade. Whether or not they deliver as much benefit as they would if they were struck as part of the EU is a marginal question, they would certainly deliver some benefit as our research shows and need to be considered.

Similarly, there is also no discussion of the potential savings from reduced EU regulation and greater flexibility in domestic legislation. It is all well and good to talk of the “dynamic” losses from Brexit as the paper does, but the scenarios used by BS are incredibly static. They do not provide any real variation on UK policy approach and essentially assume the UK would do nothing or change nothing outside the EU, a situation we think almost no one finds realistic.

In the end, it is somewhat ironic that this weak additional cost has put the BS at the less credible fringes of Brexit estimates along with huge positive gains (+10% of GDP) estimated by UKIP. Fundamentally, all studies such as these (based on little detail or analysis) do is to worsen the uninformed nature of the EU debate (which many in Europe including BS) complain about.