Last week, voters across the UK (and indeed, across the European Union) took part in the European Parliament elections. Now that we know the outcome, Alan Renwick examines the impact on the results of both the rules that governed the election and the strategies of the parties.
The European elections raised important questions about how the voting system – and parties’ and voters’ reactions to it – might influence the results. Would the imperfect proportionality of the system harm the smaller parties? Should those parties – particularly the three Britain-wide anti-Brexit parties – have formed an alliance? Could voters maximise the impact of their ballots through tactical voting? Now that the results are in, it is time to take stock.
The impact of the rules
As I set out in an earlier post, European Parliament elections in Great Britain use a list-based system of proportional representation (while those in Northern Ireland use Single Transferable vote, or STV). This system is proportional, but not very. The D’Hondt formula for allocating seats favours larger parties. So does the fact that the number of seats available in each region (ranging from three in the North East of England to ten in the South East) is fairly low.
The results would certainly have been different had the elections been held using First Past the Post, as was the case for European elections in Great Britain before 1999. This system, still used for Westminster elections, awards a seat to the largest party in each constituency. Had voters cast the same votes as they did on Thursday, the Brexit Party would under First Past the Post have won almost every seat in England and Wales outside London and the Home Counties; the Liberal Democrats and Labour would have dominated in London and parts of its environs; the SNP would have captured every seat in Scotland; and the Conservatives would have been wiped out. In fact, many voters would not have cast the same votes as they did. For example, the anti-Brexit parties could probably have agreed joint candidates much more easily than under the actual system, helping them to secure some extra seats. But the Brexit Party would very likely still have scooped up most seats on less than a third of the vote.
What, then, was the impact of the disproportionalities in the system? Table 1 compares the parties’ vote shares first with their actual shares of the seats and then with their seat shares under two alternative systems. As expected, the D’Hondt system with relatively small electoral regions overrepresented the larger parties – particularly the Brexit Party and to a lesser extent the Liberal Democrats – while underrepresenting the small parties that lack a geographically concentrated base. Crucially in the current context, this meant that the combined seat haul of the five hard Remain parties (Liberal Democrats, Greens, Change UK, SNP, and Plaid Cymru) was two below the total for the hard Leave parties (the Brexit Party and UKIP), even though the former won more votes.
The remaining columns in Table 1 show that a different voting system could have removed this skew. The most proportional possible electoral system would use the Sainte-Laguë formula in place of D’Hondt and merge the eleven electoral regions into one. This would virtually eliminate the divergences between vote and seat shares. It would have given the hard Leave parties 25 seats and the hard Remainers 27. There is no chance such a system would be introduced: it would eliminate any remnant of a link between local areas and representation and would be particularly unacceptable in Scotland and Wales. But adopting Sainte-Laguë while keeping the existing regions – a very simple reform that barely anyone would notice – would have had much the same effect, giving the hard Leavers 24 seats and the hard Remainers 26. It would have left the smallest GB-wide parties unrepresented – but even most advocates of proportional systems think minimum thresholds are desirable.
Table 1. European Parliament election results 2019, Great Britain only, under alternative voting systems
The impact of party strategies
The disproportionality of the D’Hondt system thus punished the hard Remainers for their fragmentation. So could they have increased their representation by coalescing? Two possible strategies can be investigated: the parties themselves could have coordinated better so that they were not competing against each other; or the parties’ supporters could have pursued more effective tactical voting.
Table 2 shows what could have happened if, first, the three GB-wide hard Remain parties had formed an electoral alliance and, second, those parties and the two hard Leave parties had both formed alliances. I make two assumptions here. The first is that these alliances would have been such that each bloc had just one list of candidates in each region. As I explained in my previous post, the election rules made achieving that difficult: just one of the parties would have needed to run a list in each region. Second, in order for a pact to work, all of these parties’ supporters would have had to back a joint list. This would not in fact have happened: some Change UK supporters, for example, would have refused to support a Lib Dem. So the numbers represent the outer limits of what a pacting strategy could have achieved.
Table 2 indicates that, under these assumptions, the hard Remain parties could have increased their total seat tally from 23 to 29 had they entered a pact while the Brexit Party and UKIP remained separate. Had both the hard Remainers and the hard Leavers formed pacts, the Leavers would have come out ahead in seats, despite being slightly behind in votes. That reflects the fact that the Leavers did best in the smaller regions, where the disproportionality in favour of the largest grouping is greatest. As we would expect under D’Hondt, both pacts would have been over-represented relative to their vote shares, squeezing the seats won by the smaller GB-wide parties – namely, Labour and the Conservatives.
Table 2. European Parliament election results 2019, Great Britain only, with electoral pacts
Given that such electoral pacts were not formed, voters were left to decide whether to vote tactically. Tactical voting presumably explains some of the Brexit Party’s predominance over UKIP. Voters on the hard Remain side faced a less obvious tactical calculation, and several guides sprung up aiming to help them.
Table 3 shows the final recommendations of three of these guides, from Remain Voter, Remain United and Lewis Baston. The mere fact that they disagreed with each other in most regions illustrates how difficult giving sound advice is. This difficulty has two sources. First, accurate information is needed on how people plan to vote: the final seats in each region are often decided by margins of just a few thousand votes, and these margins can be crucial in determining which strategy is best. Second, it is necessary to know how many of these voters will switch in response to a tactical voting recommendation. Any recommendation would work if all the voters it targeted followed it: these voters would then all support one list in each region, maximising the seat return under D’Hondt. But all voters will never follow such a recommendation, and it is easy for a recommendation to do harm before it starts doing good.
This can be illustrated with a few regional examples. In the North East, the recommendation to vote Lib Dem was clearly correct. The Lib Dems were the largest hard Remain party in that region; had just one in five Green or Change UK voters switched to the Lib Dems, they would have picked up a seat. But the North West is more complicated. Tactical voting could have yielded an extra seat for the hard Remain parties there, most easily if all Change UK voters and a small fraction of Green voters had switched to the Lib Dems. On the other hand, had Green voters alone begun switching to the Lib Dems, the Greens would have lost the seat they won before the Lib Dems picked one up. Similarly, in the East of England, had Lib Dem voters started to switch to the Greens, the former party would have lost a seat long before the latter gained one.
There is a general lesson here: in the absence of very good information on voting intentions and responsiveness to tactical guidance, it may be safer under D’Hondt to concentrate your vote than to seek to spread it optimally.
Table 3. Pre-election tactical voting recommendations
There were fears in the early stages of the election campaign that the seat returns would be radically biased: that the combination of a system favouring larger parties with a pattern of fragmentation on one side and concentration on the other would mean that seat shares would deviate far from the votes cast. In the end, there was a degree of skew in the results, allowing the hard Leavers to win more seats than the hard Remainers on a slightly lower vote share. But this effect was fairly small – the resurgence of the Liberal Democrats meant they had no grounds for complaint.
None of that means there is a simple read-off from the election result to the state of public opinion on Brexit. Multiple reasons – low turnout, the salience of some issues beyond Brexit, franchise differences, and so on – make that difficult. But we can at least say that there was nothing egregiously unfair in the outcome.
About the author
Dr Alan Renwick is Deputy Director of the Constitution Unit. He is the co-author of Improving Discourse During Election and Referendum Campaigns andThe Mechanics of a Further Referendum on Brexit. He also served as Research Director for the Independent Commission on Referendums.