Swings and… swings

We’re not still going on about the European elections and what happened to the Labour vote, are we. That’s a statement, not a question, and actually I’m quite disappointed that we aren’t; as soon as minor-party voting intentions dropped below 20%, and the shouting about ‘four-party politics’ subsided, people seem to have lost interest in what happened. But, while we are clearly back in the world of two ‘main’ parties, the Brexit Party and the Lib Dems do seem to have put quite a large dent in both the Tory and the Labour vote; it would be worth knowing whether this is likely to fade between now and, oh, say for example the end of October.

Fortunately, the Euro elections have been run before (who knew?) and – as I said in an earlier post – voters have shown a tendency to use the Euros to “send a message” before now. But what does this mean in practice? If we compared the Euro election vote with the previous General Election, we could establish that the Labour vote had dropped from 40% of a 69% turnout in 2017 to 14% of a 37% turnout in 2019, but what did that actually mean – particularly when Labour’s vote at the previous European election had been 24% of a 36% turnout, which was down from 35% of a 65% turnout in the previous General Election, which in turn was up from 15% of a 35% turnout at the Euro election before that? (Labour on 15% of the vote, eh? Dreadful! To be fair, Wikipedia says that Gordon Brown “faced calls for him to resign” after this result – but the linked news story shows that what he faced was calls to resign as Prime Minister, from the Leader of the Opposition. There doesn’t seem to have been any internal opposition to Brown – or if there was they kept their traps shut.)

Anyway, I tried for some time to work out the significance of 24% of 36% vs 40% of 69% vs 14% of 37% – or, failing that, to work out a way of representing the relevant figures in a readable chart so that I could see the significant bits – before it hit me that the only way to do it was to ditch the percentages and go back to the raw numbers. Which gives us these two little beauties. (Complete with titles. I’m spoiling you, I really am.)

Top Tip #1: look at the X axis – and in particular look at the origin. The Y axis is not centred at zero – for reasons which will be obvious when you look at the Y axis. Everything above zero is an increase in votes – or rather in millions of votes – as compared to the previous relevant election; everything below the line is a decrease, in millions of votes. The first big thing to take away from these charts is just how asymmetrical they both are. At all but one General Election from 1997 to 2017, around 15 million more people turned out to vote than had done at the previous European election; the exception is 2005, and even then the rise in turnout was over 10 million as compared to the previous year’s Euros. The negative difference between General Election turnout and turnout in the next European election varies more widely, but again mostly ranges between 10 and 15 million; the exception is the 1999 European election, where turnout was down 20 million on the General Election of 1997. (There’s a story there – or a sub-plot – about voters getting swept up in high-enthusiasm, high-turnout elections, and coming down to earth when they’re asked to vote again a couple of years later. (“What, another?”)) The main point here is that the story of the difference between a Euro election – any Euro election – and the previous General Election is not a story of swings and voter movements; it’s primarily a story of voters staying at home, or rather of who stays at home. Who stays at home, and who goes out muttering “voting? damn right I’m voting, this‘ll show ’em…”.

Top Tip #2: trend first, anomaly second. Is there a trend? We can’t understand what people are doing now without having some idea of what they were doing previously. Were voters behaving in a particular way for the run of Euro elections before 2019, and/or the run of General Elections before 2017? Fortunately in this case the trend is pretty clear; look at the columns for 2004, 2009 and 2014 in the first chart, and those for the General Elections in the following year – 2005, 2010 and 2015 – in the second chart. What do you see? In 2004, 2009 and 2014, between thirteen and seventeen million people who had voted for one of the three major parties in the previous General Election – four to seven million ex-Tory and ex-Labour voters and two to six million ex-Liberal Democrat voters – didn’t; while about four million people who hadn’t voted for the Greens or UKIP at the previous General Election, did (in a ratio of a million Greens to three million Kippers). Some people stayed loyal; a lot of people stayed at home; a minority of people cast a protest vote – and that minority was made significant by the low turnout. The chances are that most of the Euro Kippers had voted Tory rather than Labour or Lib Dem at the previous General Election – and that the opposite is true of the Euro Greens – but this is less important than the scale of these numbers: the main thing that happened at all those elections was abstention. Relative to the previous General Elections, the Tory vote fell by between half and two-thirds, Labour’s by between half and three-quarters and the Lib Dems’ by between half and five-sixths. For the most part this wasn’t a swing to anyone; the total combined Green and British nationalist vote at each of those European elections was, at most, half of the Tories’ vote at the previous General Election.

Now look at the second chart. Relative to the previous years’ Euro elections, in 2005, 2010 and 2015 the major parties are up thirteen to seventeen million votes. (Labour: up five to six million; Tories: up four to six million, and seven million in 2015; Lib Dems: up three to five million, and one million in 2015. That coalition was powerful stuff.) The Greens and British nationalists, on the other hand, are down a total of three and a half million in 2005 and 2010, and one million in 2015. Again, we can assume that these voters went back to their ‘home’ parties – and we can assume that the British nationalists probably went back to the Tories and the Greens probably didn’t – but, again, this is much less important than the change in turnout, which in each case was up by 10-15 million as compared with the previous European election. The swing away from UKIP and the Greens was far less important in determining those results than the swing away from the sofa.

So those are the trends. What about the last couple of elections? 2017, as you may remember, saw an unusual election campaign and an unusually high degree of polarisation between the two main parties. Relative to the 2014 European election, the Labour vote was up by nearly nine million and the Tories’ by nearly ten million, three or four million more than the increase in 2015. The Lib Dems, by contrast, only put on a million relative to 2014 – and, since I’ve measured both elections relative to 2014, this was effectively the same million that they’d put on in 2015 (in other words, the party’s vote was almost completely unchanged from the previous General Election; in fact it was down a bit). The Green and British nationalist votes fell by a total of five million relative to 2014 – but, again, the main swing was the swing away from not voting at all: overall turnout was up by nearly sixteen million. These were familiar changes, in other words, but on a larger scale than usual: compared to the 2014-15 vote changes, the rise in turnout, the rise in Tory and Labour votes and the decline in British nationalist votes were, respectively, 1.5 million greater (+11%), 2.3 million greater (+30%), 3.6 million greater (+67%) and 3.3 million greater (+330%). Presumably some Euro-election Kippers swung to Labour in 2017, but the numbers won’t have been huge. The main effects were turnout effects, as usual, but on a larger scale: the Tories were better than usual at getting out the vote, while Labour were a lot better than usual. Also, thanks to the EU Referendum seeming (temporarily) like old news, both parties did better than they had done in 2015 at calling roving Kippers home.

What happened in 2019? Those bars look pretty big, but I wonder if there’s less there than meets the eye. Over and over again, we’ve seen what are at first blush fairly huge movements of voters, between General Election and the following European election, followed at the subsequent General Election by an equally huge movement in the opposite direction; the burden of proof is surely on anyone maintaining that this time is different. So, this time, Labour and Tory vote shares – having gone up by 8.9 million and 9.8 million between 2014 and 2017 – are right back down again, dropping by 10.6 million and 12.1 million respectively; so too the British nationalist vote share, having gone down by 4.3 million between 2014 and 2017 – is up again, by 5.2 million. There’s a story, perhaps, in the ‘extra’ four million votes that the big parties lost, and the extra 0.9 million British nationalist votes; polarisation is increasing, even if it’s only at the margins. But it is at the margins – once again, there are some relatively small voter movements which have been made to look much bigger by the one big movement, the (usual) slump in turnout. (The Brexit Party topped the polls with 5.2 million votes; a party gaining that many votes would have been in a narrow third place at the General Elections of 1997 and 2001, and a firm fourth place in every other General  Election from 1983 to 2010.) There’s also a story in the results for the Lib Dems, who – for the first time ever – appear to have been seen as one of the ‘alternative’, ‘insurgent’ parties, and actually increased their vote as against the General Election; they put on a million votes as compared to 2017. But, just as the crash in votes for Labour and the Tories needs to be set against the unusually high votes for those two parties in 2017, the Lib Dems’ result needs to be set against their own crash in 2015 and their non-recovery in 2017: their total of 3.4 million votes, although higher than the party’s vote in those two General Elections, is lower than any other General Election that the party has ever contested. To find a General Election vote lower than 2017’s 2.4 million you need to go back to 1970, and even that represented a higher proportion of the (then) electorate than the 2017 result (5.4% vs 5.1%); in those terms Farron plumbed depths that the Liberals hadn’t seen since the 1950s and Jo Grimond’s leadership. All credit to the Lib Dems for their outstandingly clear – if opportunistic and misleading – positioning in the Euros; arguably they’ve reaped a deserved reward. But it’s also arguable that there’s only so low that the Lib Dem vote can go – Farron’s 2.4 million was lower than the party’s vote in four of the previous eight European elections. Really, after 2017 the only way was up – just as, for both the Tories and Labour, the only way was down.

What of the narratives? What of Theresa May’s Brexit strategy hitting the rocks and Farage moving in to pick up the survivors? What of Labour’s Brexit fence-sitting and the Lib Dems’ positioning as the party of Remain – what of the potential Remain Alliance, the Lib Dems and Greens piling up the votes while Labour’s vote plummeted? I think you’ll find it’s a bit less exciting than that. The 2019 results showed both Labour and the Tories doing a bit worse than might have been expected, the Brexit Party doing a bit better (at the expense of the Tories) and the Lib Dems doing substantially better (at the expense of both Labour and the Tories). But they’re not wildly out of line with earlier trends. Perhaps polarisation is increasing, but only at the margins: the main trend at this European election was abstention, just like it always is. Vote flows are a pain to model, but arithmetic is a limiting factor. The Labour and Tory votes were down (relative to 2017) by ten and twelve million respectively; the total votes for the Lib Dems plus the Greens, on one hand, and BXP plus UKIP and all the minor British nationalist parties, on the other, were 5.4 million and 5.8 million respectively.

What that means is that, in and of themselves, these figures don’t give any reason to believe that voters won’t be returning en masse to Labour and the Tories at the next high-turnout election – just as they did in 2005, 2010 and 2015, as well as 2017. In particular, if the next election follows the pattern of 2017, with a highly polarised campaign and a focus on getting out the vote – and why wouldn’t it? – we could easily see a similar bulge in the Labour vote. And if that’s followed by yet another slump – complete with the obligatory prophecies of doom and calls for Corbyn’s resignation – at the European elections in 2024, that’s a price I’d be prepared to pay.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: