<

Blog

23 may 2016, nate of the north

Monday, May 23rd, 2016

I'm going to indulge in some amateur demographics. At last year's uk general election, almost 37m people voted, a 66.1% turnout. Roughly 11.3m voted conservative, 9.3m labour, 3.9m ukip, 2.4m libdem, 1.4m snp and 1.2m green (full results here). Using the economist's tracker as starting point, some 54% of conservative voters can be expected to vote leave in the referendum, alongside 28% of labour's, 92% ukip and 45% libdem. For scotland we have an all-voter proxy, which imported to the nats, is 36%. For the greens, let's go with the telegraph's lowly 20%. Plugging that blizzard of stats into the referendumresult© gives us a forecast, based on the same turnout, with everyone voting as expected. It's a win for remain, but with a wafer-thin 52%. Let's play. Start with the (typically older) tory vote, where factoring in party membership being far more strongly out (around 70%) and that generally outs, being typically more passionate, are more likely to actually vote on the day than ins, that together translates to a modest 5% adjustment. For labour, I think the polling is wrong. Extrapolating from last month's council elections suggests a far bigger chunk of the labour vote, especially in the northern cities, has ukip-leanings generally and will vote leave in the referendum. With a combination of the labour leadership's distinct lack of engagement, the temptation to give cameron a bloody nose and general disengagement caused by the debate being between white tory males significantly depressing labour-voting turnout (I'd say around 30% down on the general election, still higher than the locals) - this gives remain a much closer 55% for this group. Together, these two adjustments decisively swing the vote - to a 53% win for leave. Widen that lack of passion from ins to depress the rest of their votes by just 3%, and mirror the same for outist-enthusiasm, and leave wins with 55%. On turnout, let's try a scottish referendum-like voting surge, with a 76% turnout on the same trends: it helps remain, but they still lose with 49%. More likely, there's a 10% lower turnout (and even 56% would be a pleasant surprise): leave nudges ahead, on 53%. All this helps explain why although I am unambiguously in favour (see 24 february 2016, in or out ?), I fear the vote, by a clear-enough margin, will be to leave (see 1 january 2016, the year of leaving).