In principle, EU law still applies in the UK until the day the country formally leaves. However, as Arthur Dyevre writes, the UK’s impending exit may have already altered the application of EU law in British courts. Drawing on new research, he explains that UK courts have submitted substantially fewer questions to the Court of Justice of the European Union since the Brexit referendum – a situation that could be made further complicated if the UK opts to delay or withdraw its plans to leave.
While negotiations go on in Brussels in pursuit of a hitherto elusive breakthrough, the UK is still formally in the EU. Well over two years after the Brexit referendum, EU laws, regulations and doctrines continue, in principle, to apply with full force to UK residents and state officials. Could it be, however, that, similar to businesses and multinational companies, UK judges and litigants have already started to move away from EU law in anticipation of a post-Brexit world? Have they made post-Brexit plans that they are already implementing?
Obviously, the referendum has created huge uncertainty over the terms of the UK’s future relationship with the EU. At this stage, a wide array of scenarios are still possible. A hard, cliff-edge Brexit, whereby the UK would return to third country status on 29 March 2019, remains a serious possibility. In case the negotiating parties manage to agree on a post-Brexit deal, there will be a transition period during which things will basically remain the same, but there are lots of question marks as to what the agreement, if ever adopted, might contain on matters like the Court of Justice and the direct effect of EU law in the UK.
Nor can the possibility of a second referendum, in which voters could decide to keep the UK in the EU, be entirely ruled out. Under some of these scenarios, the European Court of Justice would lose its authority over British subjects early next year or, in the event of a deal, at the end of the transition period. Under other scenarios, the jurisdiction of the CJEU could extend beyond the transition period or, if voters change their minds and elect to keep the UK in the supranational club after all, it could even emerge formally intact from the Brexit chaos.
Uncertainty puts UK judges and litigants in a delicate situation. Why submit a reference to the Court of Justice if, by the time it is announced, the preliminary ruling has no legal authority? Of course, litigation and the argumentation strategy of litigants are endogenous to judicial decision making. Litigants adapt their choice of arguments to what they believe the judges are likely to accept. So, knowing that it will be harder to convince a UK judge to request a preliminary ruling, litigants may choose to prioritise domestic law arguments over EU law arguments. Such a flight-to-domestic law would precipitate the unravelling of legal integration even if Brexit, in the end, does not materialise.
Comparing a world with to a world without Brexit
Whether this intuition captures what is actually going on in the UK legal system is an empirical question. We may be tempted to answer it by simply comparing the referral rates of UK courts before the referendum to their referral rates post-referendum. If we do this and look at quarterly referral counts (how many Article 267 references are submitted by British courts every quarter), we do, indeed, find post-referendum referral rates that appear low. In the second quarter of 2016 (April-June), UK courts submitted nine references, but only three in the third quarter immediately following the referendum.
However, it is difficult to attribute this to a Brexit-effect for several reasons. First, relative to the size of the UK’s population, referral rates for British courts have traditionally been below the EU average. Second, referral rates exhibit strong quarter-to-quarter fluctuations. Typically, judicial activity tends to be lower in the third quarter of any year, as it coincides with the judicial holiday. Finally, there are all the confounding factors that can affect referral propensity, such as lower intra-EU trade (EU law is to a large extent about trade in goods), lower public support for the EU (judges who don’t like the EU also don’t like the ECJ), variations in migration patterns (if fewer people cross borders there are fewer instances where EU law can be invoked) and so on.
UK courts are submitting 23 to 30 per cent fewer questions to the Court of Justice
To measure the effect of Brexit on British courts, what we really want to do is to compare the real-world UK to what the UK would look like had the referendum not taken place. How can we do this? We took up that very challenge in a recent paper. What we did, basically, is construct a synthetic UK from the other bits of the EU that look like the UK and then compared the real-world UK to this counterfactual. We should mention, for the technically minded, that our methodology combines machine learning with an advanced statistical technique called “difference-in-difference”.
In short, using all the data on preliminary references from 1973 to June 2018, we first identify the country attributes associated with referral activity and then construct a weighted average of these attributes to make it look as close as possible to the UK. Our synthetic UK predicts referral rates that are (minus random fluctuations) very close to the real-world UK until the referendum. After that they diverge. UK judges are referring 22 to 23 per cent fewer references after the vote than our judges in our synthetic UK. The effect is even larger and reaches 30 percent when we count the first two quarters of 2018 (this requires dropping some covariates for which data is missing, though). Overall, we find strong evidence that the referendum has adversely affected the use of EU law.
The future relationship between the EU and the UK is still in limbo. But our study suggests that the uncertainty created by the political process has already begun to unravel the fabric of legal integration. The broader lesson for the future of supranational law is that effective disintegration may precede formal withdrawal or may occur even if formal withdrawal is delayed or, for some reason, does not come to be.
This article is based on research conducted as part of the EUTHORITY Project. For more information, see the author’s accompanying paper (co-authored with Monika Glavina, Nicolas Lampach, Michal Orvadek and Wessel Wijtvliet).
Arthur Dyevre – KU Leuven
Arthur Dyevre is a Professor in the Centre for Legal Theory and Empirical Jurisprudence at KU Leuven and Principal Investigator of the ERC-funded EUTHORITY Project.