Brexit: Who Voted How? Evidence from Ashcroft
By 
Good morning, newly sovereign Britain and welcome to your new leadership team! We had Blair-Brown, then Cameron-Osborne, and now we’ll enjoy Prime Minister Johnson and Chancellor of the Exchequer Rubble.
Unlike yesterday’s immediate reaction post, this one is written on 24 hours of sleep uh, written 24 hours later on seven hours of sleep.
First, Scott L has it bang-on. While there are distinct similarities in the motivations of the Leavers and Trump supporters (as well as parts of Sanders support), the the contexts of the two votes are different enough such that the lessons that can be drawn are background, and not worthy of basing a forecast upon. Of course, given my very recent track record of superiority in forecasting, this should probably make people worry. I did get the Cameron resignation right, but then low hanging fruit . . .
We now have some (patchy) data to assess the two variables I said that we should be attentive to in the run in to the referendum. Turnout we’re not going to be able to assess yet, and in terms of raw turnout, we have nothing to compare it with really so that might never be properly assessed. However, the received wisdom of political science and psephology, that undecideds significantly break for the status quo the closer we get to polling day (I estimated 3:1), is not receiving support, and this might be one key to understanding how my 52:48 prediction went ass-backwards (as well as why NCP’s forecast of 53:47 was just a bit off).
Those data are from an Ashcroft poll (details here, a lot of interesting stuff to pore over) conducted immediately after the referendum, with an N of over 12,000. Those who decided a week out split 7%-6% for remain, a few days out 8%-7%, on the day 10%-9%. While there’s a marginal advantage for remain in the late deciders, these data suggest it’s only 53%. Not the 2:1 or 3:1 breaks for the status quo we typically expect.
Demographics did work out as expected. The age gap (which we’ve known about for well over a month) is receiving a lot of attention in the media at present, as though everybody is surprised. Social class worked out as well; the higher up the socio-economic ladder one is, the odds of voting remain increase. However, where pre-referendum models suggested social class C1 (lower middle class) would just support remain, they ultimately just supported leave (and the professional classes, A & B, did not support remain at the rate initially thought). The age gap is striking.
The next two figures illustrate the support for each side in party political terms, and how the parties own supporters voted. I have two observations here. First, it was Conservative voters who drove Brexit. 40% of leave support was Tory, 25% UKIP, and 21% Labour. Yes, a nice cross-party distribution, but there were over three Tory/UKIP (in the parlance of Plymouth Labour as we’re now enjoying a Con-Kip coalition in this fair city, “Blukip”) supporters for every one Labour supporter voting leave. Additionally, this also tests my off-the-cuff suggestion yesterday morning that greater than 35% of Labour supporters voted leave. Ashcroft estimates the figure at 37%. While the traditional Labour heartlands of the northwest and northeast went huge for leave, Labour by and large did not. 58% of Conservatives did.
As the Labour Party itself is going through an uncertain period (charitably stated), one thing I’ve been hearing and reading that does need to stop now is that we didn’t “lose” more of our supporters than originally expected. Indeed, the 37% estimate is in line with expectations. Additionally, it doesn’t mean that all of those supporters we “lost” to leave were the traditional working class base. That figure must include a large degree of Lexiters, as evidenced by the 25% of Green Party supporters who inexplicably voted leave.
And just who are those 4% of UKIP supporters who voted remain? Statistically, there had to be some, of course, but it’s still hilarious fun to point it out.
And now, the money shot. Why did the lies of the leave campaign resonate? Why was Gove sage to suggest that Britain has had enough of experts experting us to death with facts in their expert ways? Self-reported political attentiveness breaks as we would think.
It’s not as stark as I’d have thought, but this is self-reported. There’s a lot more there in the link above worth looking at, of course. Later today, but more likely tomorrow, I’ll have some further thoughts on the result of the result, Article 50, and speculate as to just what the hell Boris Johnson is up to.
