Why did so many people vote in 2017 who had never voted in a general election before? Not only did the headline figure for turnout across the United Kingdom go up by 2.5 per cent, but that masked a much bigger increase in the number of first-time voters, for two reasons.
The first was that turnout dipped little among the Conservative-voting over-55s in England and Wales, which Rob Ford has dubbed the “grumpy grannies” effect (Theresa May irritated elderly voters, but, rather than vote for the Opposition parties, they opted not to bother). The second, and perhaps more interesting factor is that in Scotland, turnout went down by 4.7 per cent. (This will be important later on.)
We know that there were two types of first-time voters in 2017: people who had never voted in a general election before, but had cast their first ever vote in the European referendum (mostly but not wholly for Remain), and that these voters backed Labour more often than not.
We know, too, that a second group of non-voters had never voted in any contest before and they also largely backed the Labour Party. Although we can’t say for absolutely certain these people voted solely down to Jeremy Corbyn and the breach he offered with Labour’s recent past, I see no good argument not to declare that these non-voters were the responsibility of the Labour leadership and leave it at that.
The more interesting group of non-voters are people who first started voting in the referendum and kept up the habit in 2017, and that’s where Scotland comes into play.
What happened in Scotland is the fascinating known unknown of the last election and potentially of the one to come as well. All three of the Unionist parties gained seats there – the Conservatives picked up 12, Labour six and the Liberal Democrats three.
But of the three parties, only the Conservatives gained votes in a particularly significant way. The Liberal Democrats actually got fewer votes in 2017 in than in 2015, while the Labour vote went up by just 9,860 votes over the whole of Scotland.
What mattered was that half a million people who voted for the SNP in 2015 did not vote in 2017, which allowed Labour and the Liberal Democrats to gain seats while treading water in terms of their actual performance as far as votes went, and in some cases actually losing votes. (Labour gained Kirkcaldy and Cowdenbeath with 638 fewer votes than they got in defeat in 2015, while the Liberal Democrats were able to pull off a similar feat in Caithness.)
And the thing is, it is not wholly clear why that happened. You can make a good case about the SNP’s record in office or the change in the Labour party but unlike with those 2017-era non-voters, it isn’t open-and-shut. (Not least because you also have to account for the fact that turnout was way up in Northern Ireland – about 20 points higher in fact.) Another, drier explanation may also work.
Turnout also went up in the election following the European referendum of 1975, in Scotland after the referendum of 2014, and across the United Kingdom after the 2016 referendum. But in the election after that (1983 and 2017 in Scotland) turnout went back down again.
One of the difficult things about analyzing elections is that they are “statistically noisy” – which is a polite way of going “there is a lot of stuff going on and it’s no clear how this stuff cancels each other out”. There are lots of very plausible reasons why turnout would be higher in 1979 than 1983 that have little to do with a high-stakes referendum occurring beforehand – though it is less easy to come up with a similar case for Scotland in 2015. There are good reasons for turnout to have been higher in 2017 in England and Wales (though it is worth pointing out that you have to explain the turnout drop in Scotland and the increase in Northern Ireland too) that aren’t referendum-dependent. There are even plenty of plausible reasons that turnout could be lower at the next election that have nothing to do with this “referendum effect”, even if there is one.
But it speaks to one of the intriguing things that we know we don’t know about the 2017 election: we know that the referendum of 2016 played a major role, but we don’t know how major and how enduring that effect will be.