Since this piece was first published, an error in Lord Ashcroft’s Doncaster North poll has come to light. Ed Miliband is in no danger whatsoever in his own seat. I have struck out part of the text and updated the chart accordingly.
Following on from what I wrote on Wednesday, it turns out that Lord Ashcroft wasn’t just polling Ed Miliband’s Doncaster North seat, but also Nick Clegg’s Sheffield Hallam constituency and Nigel Farage’s target, Thanet South.
So to answer my own question, on this basis about half of Doncaster North Conservatives would need to go purple to unseat the Labour leader. As I wrote previously, Tories aren’t known for tactical voting – the chance to decapitate Labour will doubtless tempt a few, but how many? Miliband at least has a comfortable majority to begin with.
In Sheffield Hallam, Nick Clegg is behind on standard voting intention but ahead on constituency voting intention, consistent with Lib Dem incumbency, but seemingly with less of a boost than other MPs in his party. The Labour threat to Clegg in his backyard has been talked about for most of this parliament, after an Ashcroft poll in 2010 put Labour just two points behind in the South Yorkshire constituency, the only seat in that county that isn’t Labour-held. Clegg also starts with a substantial majority, but with such a large student vote, Labour feel it is “soft” (see here and here). Another complicating factor, perhaps unsurprisingly given the demographics, is the extent of the Green surge, nudging double digits with five times their 2010 vote share on both questions. Meanwhile the Conservatives are polling 23% on the first question but 19% on the second – if it really is this close, then this implied 4% tactical swing would be enough to save Clegg.
Nigel Farage was previously a strong favourite in Thanet South, but is now 4 or 5 points behind the Tories. There doesn’t seem to be much sign of tactical voting, and if there is, which way will it go? By-elections in the current parliament have seen suggestions of Labour supporters voting tactically both against UKIP and the Conservatives. The two differences in May will be (1) that it is a general election, with higher turnout than, and different consequences to, a by-election and (2) that it will be Farage himself standing there. Or will it? There were rumours that he might be thinking of standing elsewhere, which as a non-incumbent he could do easily enough, but what about the political consequences?
So it is not wholly inconceivable that the Conservatives could defeat the UKIP leader, UKIP could defeat the Labour leader and Labour could defeat the Lib Dem Leader. But it’s not particularly likely either.
Decapitations are rarely successful and leaders losing their own seats rarer still. No ‘big two’ party leader has lost his or her seat since 1931 and, excluding the exceptional circumstances on that occasion (the MacDonald “betrayal” leaving Labour in the hands of serial seat-loser Arthur Henderson) and Arthur Balfour in 1906 (change of PM and governing party the month before the election), no leader of HM opposition ever has.
What is more likely though, is that some party leaders have an unwelcome distraction close to home, with could divert their attention from national matters.
In the more traditional battleground polling, individual seats are interesting, but in aggregate, nothing hugely different from last time. The LIB-CON swing is little changed at 2%, a couple of points less than in Ashcroft National Polls. As usual we see plenty of variation, such that the Tories would ‘unexpectedly’ gain two southern seats – Portsmouth South and Devon North. If we adjust for the Ashcroft House effect (which typically has the big two lower than other pollsters) then Brecon and Radnorshire looks like being close.
There are only two LIB-LAB contests this time, including John Hemming holding Birmingham Yardley against the trend elsewhere in the battleground.
Survation also have a poll out for Camborne and Redruth, which was a Conservative gain from the Lib Dems by just 66 votes last time. The poll uses a split question as Lord Ashcroft does, and finds. UKIP in the lead on standard voting intention, but George Eustice holding for the Tories on the constituency question.
One caveat is that the sample size is only 500. It’s also worth noting that, Survation no longer use political weighting for constituency polling, a move which they admit trends to increase UKIP’s vote share compared to politically-weighted polls. Whether this makes it more-or-less accurate than Lord Ashcroft’s polling is open to debate, but either way it marks a clear methodological difference.