Labour's rising poll share

Huge excitement over YouGov's most recent poll, which shows a Conservative lead of 5%. This is somewhat totemic, as it's the smallest lead *any* pollster has given the Conservatives since Theresa May became prime-minister -- and, for the first time, allows an outcome where her small majority could even be reduced (though not eliminated). So I thought I'd take a closer look.

Here's how the polls have changed since the Brexit vote (individual polls are circles, bold ones are YouGov ones): it's always best to look at the average, not at individual polls. I've already written about how the growing Tory lead was driven largely by the collapse of UKIP (with Jeremy Corbyn in charge, most genuine swing voters were already on the blue team or had gone elsewhere), and how Labour are consolidating the left in response (a good chance to remind everyone: voters aren't stupid, and they can count!). The combined Con+Lab share is now around 80%, something which hasn't been de rigeur since the end of the 1970s.

Clearly, public opinion is moving very quickly, and although my hunch is that this particular poll is an outlier in Labour's favour (every time you measure a party's *lead*, instead of their share, there's a 1 in 20 chance of your being wrong by more than 6%), it's hard to know for certain when the trend is changing so sharply.

So if it's real, what's driving it? A look at the detailed information* from the last two YouGov polls reveals something very interesting, namely that only three things have substantially changed. First, the fraction of C2DE (less well-off) voters intending to vote Conservative is now HIGHER than the fraction of ABC1 (more well-off) voters. Secondly, the Conservatives have not lost an inch among the 65+ vote, but they have gone backwards amongst 50-65 year-olds. Finally, the vote shares among men haven't really changed, but this poll shows a lot of women moving away from the Conservatives and towards Labour.

We need to be cautious about drawing conclusions from sub-samples (the statistical uncertainties get a lot bigger), but the fact that the swing to Labour is coming from these groups is interesting. If (big if) the next poll brings us comparable numbers, then I'd be reasonably confident in saying that the Conservative proposals for social care changes have hit them amongst well-off people with retirement on the horizon. Whether that would actually hurt them under first-past-the-post remains to be seen.

---

* http://d25d2506sfb94s.cloudfront.net/cumulus_uploads/document/wvyc3lofp5/SundayTimesResults_170519_VI_W.pdf

and

https://d25d2506sfb94s.cloudfront.net/cumulus_uploads/document/dcfgflapq2