In the corridors of the Palace of Westminster and in those of the Union’s institutions, the air is rife with talk about extending the two-year period under Article 50 TEU. Both decision-makers and commentators regularly point out that delaying Brexit would be problematic if the delay ran into the upcoming European elections. It is worth taking a closer look to see just how problematic that would indeed be.

Firstly, the UK would have to hold European elections in May, or some other way to ensure representation of Union citizens living in the UK would have to be devised.

Secondly, Brexit delayed into the new parliamentary term, which starts on 2 July 2019, would render the composition of the European Parliament incompatible with the requirements of Article 14(2) TEU, an incompatibility that is very hard to justify and nearly impossible to address.


The duty of the UK to hold elections

The legal position until now has been that the UK will leave the European Union on 29 March 2019 and so the assumption has been that no elections will be held. It might seem a bit late now to organise elections after all, but the UK, with its tradition of snap elections and short preparatory periods, could still do it. The Secretary of State would have to set the day of the poll by order and then a notice of election would have to be issued to trigger the candidacy process 25 working days before the poll. With a poll taking place on Thursday 23 May – or perhaps pushed back to Saturday 25 May to gain a bit more time – and three Bank holidays in the weeks before, these formal steps would need to be completed around 15 April. More practically, political parties would have to draw up their lists of candidates and have them ready for submission by the day after the notice. The UK Electoral Commission reports (in private correspondence to us) that ‘we are making certain preparations’ and that ‘political parties, prospective candidates and other campaigners, along with returning officers and their staff will also be mindful of the actions they may want or need to take in the lead up to any scheduled or unscheduled election’.

Quite apart from the legal and practical feasibility of holding elections, the prospect of campaigners knocking on doors and explaining that they need votes to get candidates to Strasbourg because Brexit has been delayed, is a very unattractive one. So it is little wonder that ways to avoid the need for elections, even if Brexit is delayed beyond 2 July 2019, are being explored.

The House of Commons’ European Scrutiny Committee put questions on this to the UK government in this report (para 45). And the Brexit committee heard expert evidence from, amongst others, Professor Alison Young, and discussed the idea that representation of EU citizens living in the UK might be realised through co-opting Westminster MPs to sit in Strasbourg. This could only be done if the treaty requirement for ‘direct universal suffrage’ in Article 223 TFEU were set aside. We do not believe the EU’s own specific treaty-making power under Article 50 TEU could cover for that, but if the matter could be seen as merely one of ‘withdrawal’, then it might. A full-blown ad hoc treaty adopted under the ordinary revision procedure of Article 48 TEU would of course do the trick, but there is simply no time for that. Another idea mentioned is continuation of the term of office of current MEPs from the UK, something on which the end-of-term provision in Article 5(2) of the Direct Elections Act would have to be overruled. Leaving Union citizens from the UK entirely unrepresented for a very short stretch of time might also seem acceptable and might be justified with a reference to highly exceptional circumstances, but legally this would be an extremely narrow path to go down.

Overall, the different scenarios can perhaps be summarised in the following manner:

A. Brexit before 2 July 2019

In this scenario, no elections would be held. The Secretary of State would not to set a date for the poll and presumably the British legislation on European elections would also quickly be revoked.

B. Short delay of Brexit reaching into the new parliamentary term

Under this scenario, representation of Union citizens in Strasbourg would have to be ensured. Whichever exact form this might take, it would require legal underpinning both in UK and EU law and it is unclear what the latter would look like. A provision on the level of primary treaty law might be needed.

C. Longer delay of Brexit or Brexit cancelled

If Brexit were delayed for a longer period, full participation in the upcoming elections would seem to be the only viable choice. Should Brexit be cancelled altogether, holding elections of course would become inevitable.

Whilst any half-way house in scenario B looks most puzzling, holding full elections under scenario C is problematic as well, this time on a different, as it were ‘domestic’ EU level.


The problem with the old composition of the European Parliament being prolonged

The composition of the European Parliament during the upcoming 2019 to 2024 term, that is to say the distribution of Strasbourg seats over the Member States, was set in the in European Council Decision (EU) 2018/937 of 28 June 2018. The new composition of the Parliament post-Brexit is in perfect conformity with the treaty requirements in Article 14(2) TEU: the minimum of 6 seats per member state, maximum of 96, total number no more than 751 and degressive proportionality are all respected. The decision also provides for the eventuality of a delayed Brexit that reaches into the new parliamentary term. In that case, the old composition from the current 2014 to 2019 term is simply continued until Brexit does occur. Whilst that old composition is in order as far as the minimum and maximum numbers are concerned, it does not respect degressive proportionality.

Degressive proportionality means that, whilst more populous Member States can of course expect to send more representatives to Strasbourg and will certainly not send fewer than any smaller State, representation relative to population becomes weaker and weaker the larger a Member State is. The following two requirements, inherent in the notion of degressive proportionality, are spelled out correctly in the preparatory materials to the Decision:

(1) no less populous State shall receive more seats than a more populous State, and

(2) the ratio population/seats shall increase as population increases before rounding to whole numbers.

As population numbers change over time, the distribution needs to be adapted before each new election. The 2018 Decision chose not to do this for the scenario of delayed Brexit. By the time that Decision was taken, the following developments had occurred.

Through population increase in Sweden and a decline in Hungary, Sweden had overtaken Hungary. But Sweden still has 20 seats, one fewer than Hungary’s 21. This amounts to a breach of requirement 1) above.

There are several points at which the ratio population/seats does not increase but in fact decreases with increasing population. The Netherlands has a markedly weaker representation ratio of roughly 662,000 inhabitants per seat than Romania with roughly 614,000; as the Netherlands is smaller in population than Romania, this is the wrong way around. Similarly, Ireland is underrepresented compared to Slovakia and Finland; Denmark compared to Bulgaria; Austria compared to Hungary; Sweden compared to Portugal and the Czech Republic; Spain compared to Italy; and Spain, the UK and France compared to Germany. All of these discrepancies go beyond a rounding to whole numbers, so each of them renders the composition of the European Parliament in breach of requirement 2) above.

Article 3(2) of the 2018 Decision, which prolongs the old composition in case of a delayed Brexit, is therefore in breach of Article 14(2) TEU. It should be stressed that this breach is not a problem of representation of Member States as such, but of the Union citizens living in those States. It is a problem with regard to individual electoral and political rights, one that will cast a shadow on the democratic legitimacy of the European Parliament in case of delayed Brexit, unless it is addressed.


The lack of justification for this irregularity

Interestingly, the factual reason why the European Council took its 2018 Decision as it did seems quite clear: there was a strong wish on the side of Member States not to lose seats compared to the current term. The explanatory statement accompanying the Decision even stipulates this as one of the leading principles: ‘No loss of seats for any Member State’. For the post-Brexit composition of the European Parliament this could be ensured relatively easily, as the 73 UK seats falling open would provide for the necessary room for manoeuvre. Once this part of the Decision had crystallised, there was only one way to deal with the scenario of a delayed Brexit: if no Member State must lose seats, and the maximum of 751 is already exhausted, then logically no Member State can gain seats and the overall composition must remain exactly the same as before.

This motivation does not, of course, amount to a justification of a breach of the requirements in Article 14(2) TEU. The entire system of distribution of seats over Member States, as based in the Treaties, is predicated on regular revisions. A loss of seats is as much a part of such revisions as gaining seats. The wish not to lose seats, whatever one might think of it politically, cannot be a legitimate aim that could justify the breach.


Ways to resolve this problem under the different scenarios

A. Brexit before 2 July 2019

In this scenario, the breach of degressive proportionality does not occur at all. The new composition of the European Parliament post-Brexit conforms to all requirements in Article 14(2) TEU.

B. Short delay of Brexit reaching into the new parliamentary term

If a temporary alternative form of representation could be found, then the instrument establishing it could conceivably also be used as a vehicle for the readjustment of the European Parliament’s composition. Making provision on the level of primary treaty law, for instance, could also address the defect as far as EU law is concerned. However, any actual readjustment would have to be effectuated also within the Member States through their national laws on European elections. This is not straightforward. Similarly, if Union citizens from the UK remained unrepresented for a short period and if this would be successfully justified by exceptional circumstances, then a short period of irregular composition would seem justifiable as well. In that scenario, British seats in Strasbourg would remain empty. One might assume, then, that the new regular composition could kick in already, but national laws on European elections would have to facilitate this, which for example the relevant Dutch law does not.

C. Longer delay of Brexit or Brexit cancelled

A longer delay of Brexit, with regular European elections in the UK being held in May, would entrench the irregularity for a longer period. Should Brexit be cancelled altogether, possibly after a second referendum, this problem would persist for the entire parliamentary term from 2019 to 2024. Fully resolving the problem would require a fresh decision on the composition of the European Parliament under Article 14(2) TEU, this time with several Member States agreeing to lose seats. It is not probable that the result of the poll in May 2019 could then be used to re-distribute seats. In Ireland, for instance, elections are held in three separate districts, the number of seats elected in each of them is set by law, and the total number of seats has to be known before the poll. The Irish legislature recently passed a bill to amend its European elections law, which seems not to provide for a change of seat numbers after the poll. The same problem exists in the UK itself, where the only legal provision that is repealed already is the one on distributing seats over the several electoral districts. Similar difficulties will probably exist in other Member States. So a full-blown revision of the 2018 Decision on seat distribution would have to be implemented in the Member States and a fresh poll would probably have to be held. This seems extremely unattractive. A more realistic but legally less elegant solution would be to let the new post-Brexit distribution of seats kick in and give the UK a number of seats on top of that. This would take the overall size of the European Parliament beyond the maximum of 751 seats, so it would require some treaty-like instrument to set aside that maximum. It also would require revision of national laws, for instance of the Dutch law, which does establish the new seat number only after Brexit, not under delayed Brexit. There might be a bigger chance for this solution being viable without a fresh poll, but it is difficult to say whether problems would not arise in some Member States.


EU electoral law is, at the moment, well prepared for Brexit. It is ill prepared for delayed Brexit, and it is worst prepared for no Brexit. That is not a good position to be in.



Leonard Besselink holds the chair of Constitutional Law in the Faculty Law, is affilliated to the Amsterdam Centre for European Law and Governance and is presently Head of the Department of Public Law at the University of Amsterdam.

Bastian Michel is a PhD Researcher at the University of Amsterdam.


Author :

Leave a Reply