Easter 🔊 Sunday also 2772th Anniversary of the Foundation of Rome, or so the story goes, and quite definitely James McAvoy's birthday
The Road So Far 🔊
Elections for the European Parliament are to be held in the UK on 23 May 2019 for the last time 🔊. Or not. Holding them is part of the UK's commitments in the latest Brextension deal but they could be cancelled at the last moment, if Commons miraculously pass Theresa May's Withdrawal Agreement they have already rejected thrice. It would have to happen before 22 May anyway, which not coincidentally was the backstop deadline in the Brexit 2.0 arrangement. Otherwise the vote goes on come what May wants or doesn't.
Let's just see first who the current UK MEPs are. Which, believe it or not, is not the same as they were after the last election in 2014. And it changed again a few days ago when two former Tory MEPs switched to Change UK. Guess they wanted to play it safe and waited to be sure the TigChukers were actually registered as a real party by the Electoral Commission. The 73 MEPs are grouped by national party and by European Parliament group (which are somewhat different from, though mostly overlapping with, European parties but never mind as we're bordering on nerdy TMI here). Unlabelled denotes just one MEP to avoid cluttering the graph. One UK seat is currently vacant (formerly a Labour seat in Scotland held by Catherine Stihler).
I have tracked the changes in a 'where they were and where they are now' chart. Hope it's not too confusing. Same colour code as above and again unlabelled means one MEP in category. Note that the left-wing parties and the Northern Ireland representation have not changed either their national party or their parliamentary group affiliations. Only UKIP and the Conservatives have gone through some kind of more or less intense musical chairs.
And that's all the changes we'll see as the 2018-2019 season has come to an end. Right now MEPs no longer have the opportunity to carry on with the noise and confusion 🔊 as the European Parliament is in recess and will not reconvene until 2 July with the new intake of MEPs.
Can EU election polls be trusted?
British pollsters have become quite of a laughing stock since they failed so dismally to predict the results of the 2016 EU referendum and the 2017 snap GE. Whatever the changes in their methodologies, there is still an ominous cloud of distrust hanging over their heads. But their performance back in 2014 was not that bad. I managed to retrieve six EU election polls from the pollsters' sites, all ten days or less away from the election. There were other polls fielded in that timeframe but they seem to have been woodchiporized into some cyber-bin since. Anyway here's what the six polls I have and their weighted average said, and it's not really far from the actual GB-wide results (Northern Ireland was not polled back then just as it is not right now).
Scottish polls on the other hand are a different kind of a strange beast. Contrary to popular belief they did not miss the Independence Referendum by much, with the trend over the last week being a steady 47-53 No victory despite a handful of outliers predicting a Yes victory. They weren't that bad either at the 2015 GE but that was the easy one when the only doubt was about SNP bagging 'only' 54 seats or going all the way to 59. Then they totally fucked up the 2016 Holyrood election, the 2017 Council elections and the 2017 snap GE, all the time wildly overestimating the SNP vote and the 2014 EU election was no exception. Below are the results of seven Scotland-only polls fielded back then and the weighted average of the last two fielded just before the election.
If these polls had been right the result would have been three SNP MEPs, two Labour and one Conservative with SNP snatching the sixth seat from UKIP by quite a wide margin. Marco Biagi (formerly the SNP MSP for Edinburgh Central) recently published a witty and informative article about the forthcoming EU election, warning everyone to 'be more wary of polls than normal'. And of course he is totally right and we should bear this in mind when assessing current polling.
So then what do current polls say?
Before you ask (again) all the polls I use are listed here with links to the source data on the pollsters' sites that include regional crosstabs. Nine polls have been fielded so far, starting in January when holding EU elections was nothing more than a fantasy scenario and until this week with three polls in close succession. There is little to no consistency in these polls and nothing so far looking like a reliable trend. The press will of course be looking for a winner in a context that could eventually deliver only losers. Only sure thing is that the Conservatives will not be the winners as their choice of options has narrowed down to walloping and clobbering, until voters go one step further and hand them a quartering. So here is what all these polls say:
Saying that there is some fluidity in EU election polling would be quite an understatement as volatility might indeed be the word we're looking for. So I have restricted my EU Poll'O'Polls to just the three most recent ones, all fielded less than a week ago. But even polls fielded literally on the same day show quite interesting discrepancies so the usual caveats strongly apply, and then to the power of 73. So here is the current weighted average of voting intentions, supposed to be as faithful a snapshot of public opinion as pollsterly possible. Followed by its breakdown by electoral regions, based on what regional crosstabs pollsters see fit to provide. Northern Ireland is excluded here as they have not been polled so far.
These numbers clearly demonstrate that nobody has any reason to feel safe in the current political climate. UKIP would be irrevocably replaced by the Brexit Party as the Poster Boys for the New Model Blackshirts. Though voters choosing borderline-batshit-crazy over terminally-batshit-crazy is not a sign of sanity from any perspective. Conservatives are predicted to take a massive bashing especially in the traditional Southern Little England heartlands. And it can only get worse as more and more Tory grassrooters refuse to campaign in the name of 'Brexit betrayal'. Labour have no reason to be feasting either as EU polls show the same symptoms as GE polls: Labour looking goodish only because Conservatives are nosediving into a crashlanding. Even their results in the Northern Powerhouse heartlands and in London are alarming, notwithstanding an unexpected resilience in the North East.
What kind of representation would this deliver?
The current 73 MEPs are elected via proportional representation (the famous/infamous d'Hondt method) in twelve electoral regions that overlap the European Union's NUTS-1 regions. These have been designed by the EU for statistical purposes only, not electoral ones, which explains why the number of MEPs per region varies wildly from three to ten. There is in fact no strong and convincing rationale for the UK using NUTS-1s as electoral regions. France also used regional constituencies three times for EU elections before reverting to a single national list, but their constituencies did not overlap their NUTS-1s in any way.
But enough useless trivia for now. Here is the projected representation of the UK on the evening of 23 May, or more probably the wee hours of 24 May, based on the aggregation of seat projections by region. There is an example of how d'Hondt works just below in my Scotland section and it works the same way for all regions, you just have to adjust the number of seats. I have included Northern Ireland here hypothesizing their representation would be the same as in 2014. Best that can be done without any available polls, and the only reasonably expectable change would be one seat switching from UUP to DUP, and this would have no impact on the overall picture.
The seat projections show that Labour are indeed in a dangerous position even if they're on their way to gain a couple of seats. They certainly walk on thin ice in London where some of the most dedicated Remain voters would switch to Change UK on top of those who have already switched to the LibDems. And their smallish gains in the North definitely have more to do with Conservatives crashlanding than Labour conclusively winning back C2DE Leave voters. Ambiguity never wins.
And what about Scotland then?
Scottish polling so far is just as inconsistent and puzzling as the GB-wide results. Furthermore we don't have real Scotland-only polls now as we had in 2014 but only subsamples from the GB-wide polls, which adds another level of uncertainty and calls for extra caution. With these caveats duly registered, the SNP is doing significantly better than their 2014 result and this is supported by other non-EU polls so we can accept the trend as valid, even if the precise numbers have to be subjected to reasonable questioning. Here is what we have now from all nine available EU polls and the weighted average of the last three.
To make it clearer here is the current weighted average of voting intentions we have for Scotland. Not outlandishly good for the SNP as some earlier polls were, but still some 10% above the 2014 vote. Unionist parties are not doing that well and most probably below their most pessimistic expectations. Greens and Brexit Party breathing down the neck of the Ruth Davidson Party should ring an alarm bell if only they ever listened to the people. Which they don't. And won't. And then will come whining about the walloping they took.
The volatility of polls is reflected in the seat projections (more on how the sausage is made right after this). Let's just say it's safe to conclude that the SNP can (and will) bag three seats (Alyn Smith, Margaret Ferrier and Christian Allard if the conference votes the way I would) while the other three would be kind of marginals if you can picture such a thing under PR. Labour probably have a head start and should bag one unless some kind of upset happens, while the remaining two could be decided by just a few votes between Greens, Conservatives and the Brexit Party. So SNP brace yourselves for gaining only one seat 😅😇🐶. But don't give up on the fourth just yet.
With only six seats at stake Scotland really pushes the highest averages method to its limits. Remember first that 'average' here does not mean 'average' as in [ Votes / Seats ] but 'what if average' as in [Votes / Seats+1 ]. What we're trying to do here is not statistics about the actual distribution of seats but allocating the next seat in a way that represents voters' choices best. This can be done only by simulating how many voters would be represented if all parties were allocated an extra seat and then giving it to the party with the highest calculated average of votes per seat. Just watch what we have on current weighted average and hypothetical turnout same as 2014. Of course it works just the same whatever the number of votes cast, as only the vote shares actually matter. On Round 1 all parties are allocated one virtual seat, then on Round 2 the winner of Round 1 gets two virtuals and all others one again, and so on…. Simples.
Also bear in mind that d'Hondt (and for that matter any PR method whatever the way they allocate seats) sets a de facto threshold for representation that is closely linked to the number of seats but also to the distribution of the popular vote. Highest averages method is known for (and indeed designed to) favouring larger parties while the concurrent largest remainders method favours smaller parties. Marco Biagi rightly points out in his article that the SNP missed a third seat (the controversial one that went to UKIP only because of anti-SNP tactical voting) in 2014 only by some 30k votes or 2% of all votes cast. If the SNP had had these extra votes, hypothetically from actual non-voters, they would have bagged 50% of the seats on 31% of the vote, clearly a counter-intuitive result and closer to FPTP than to what you would expect from PR.
We have a pretty similar situation on current polling. Only the SNP and the three next contenders (Labour, Conservatives, Greens right now with Brexit Party also a contender in an earlier poll) would clear the de facto threshold which is about 10-11% for Scotland on six seats. So if you recalculate vote shares for the four finalists only on a 100% basis, what you get right now is 51% for the SNP and about 15-18% each for the other three, simply because non-SNP vote is highly fragmented on current polling. Put it another way and what you get is roughly a 3-1-1-1 distribution, precisely what the seat projection says, and that makes d'Hondt truly proportional eventually no matter how it looks like at first glance. QED.
But what if Scottish polls are totally off again?
Just rewind to the part about 'can polls be trusted?' and look again at the numbers for Scotland. So in 2014 pollsters overestimated the SNP by almost 8% and underestimated Tories by about 4%. All other parties except Labour were underestimated too but in all cases the error was within the usual margin of such polls. Which does not mean it would not have mattered in a close election. So here you have a reminder of what current Scottish polling says and what it would be if the polls were off the same way they were in 2014:
This might look like not much of a big deal at first glance. After all the SNP would still be comfortably ahead of all other parties and up from 2014. But just remember d'Hondt's de facto threshold is 10.5% in Scotland and look at the Brexit Party's corrected vote share. The recalculated allocations of seats says it all: SNP 2, Conservatives 1, Labour 1, Greens 1, Brexit 1. And the SNP would lose the sixth seat by barely 2k votes or 0.1% of votes cast. Quite frightening, isn't it?
What this tells us is not just that we should take all polls with a muckle smitch of salt, much more importantly this tells us strong campaigning and a massive GOTV drive are of the essence for this election even more than for previous ones. Bear in mind that this time Scotland is ONE constituency, not 59 or 73, so all votes have the exact same weight no matter where they're cast. The SNP definitely needs all these votes to be cast. And above all the SNP should not take anything for granted, especially not the people's votes. Not now. Not ever.
And now what comes next?
As usual what happens next is anybody's guess and yours is as good as mine. The only thing I'm certain of is that EU elections are not predictors of what might happen at the next GE as results over the last twenty years show. Past elections show that EU elections are the perfect outlet for venting frustrations and casting a protest vote, whatever the protest is actually about. And don't think it's just a UK thing, it works pretty much the same way all over the EU. Quite simply because most people don't have the slightest clue about what the European Parliament does and the EU election results don't change the national balance of power in any way.
Which does not mean we should all abide by common popular wisdom especially when it is plainly and demonstrably wrong. Even the strongest believer in democracy must admit it: yes the people can be wrong especially when the people's vote turns into mob's vote. Which has been a constant factor in past EU elections and goes a long way explaining why far-right and narrow-brained nationalism are so heavily represented in the European Parliament.
But closer to home our main concern should probably be that the likely result of next month's EU election will project the worst possible image of the UK in the current context. Just look at who the two most likely front-runners are. Labour who haven't yet sorted out what species of Brexit, not-totally-Brexit or almost-Brexit-but-not-quite they actually support and are ready to send Remainer MEPs to Strasbourg/Brussels while still fantasizing about some incarnation of 'progressive Brexit' in London. And Brexit Party who's only clear goal is to get elected to destroy the institution from the inside while still cashing in lofty expenses.
Once again Scotland will have to set the tone. I obviously wish for us returning four SNP MEPs but three SNP and one Green would be the second best option and I could live with that. As long as the message is clear: a majority of pro-Independence MEPs and a strong commitment to having Scotland's voice heard loud and clear both in the UK and abroad. We can do it so let's do this. A month left to win. Fingers crossed.
We have helped to write European history and Europe has helped write ours (David Cameron)
© Jean-Jacques Burnel, 1979
No comments:
Post a Comment