Monday, 21 September 2015

Corbynmania: early evidence

One good thing about having very strong opinions is that you can then test them against the evidence. Now, there weren't many opinions stronger than those we came out with last week - namely that Labour's new leader, Jeremy Corbyn (above), was doomed to fail on psephological, domestic and foreign policy grounds. Now, the fascinating thing is that - having shown our workings - we can get down to actually testing those views. The fact that we've already put those concepts and computations down in hard, cold print has one enormous advantage: it shows not only where we were right, but (still more interesting) where we may have gone wrong.

Mr Corbyn's foreign and domestic policies will probably take a while to assemble - especially as his Shadow Cabinet and MPs seem to have an open license to just ignore them. But we already have some polling, and actual electoral, data that we can use to test and explore our assumptions about Mr Corbyn's likely popularity - or, to be blunt, lack of it. So let's get to it.

Last week we said that the new look Labour Party would probably be generally unpopular, given its leader's past, as well as his views on key touchstone issues. But we thought there was a likelihood that his novel experiment might get quite a lift or 'bounce' in the polls, due to its freshness, apparent lack of spin, and clear opinions. However, we furthermore argued that he would have a hard time attracting voters back from the Scottish National Party (due to that party's patriotic and centrist, as well as 'left wing' appeal), but that Mr Corbyn might be able to get Green and United Kingdom Independence Party adherents to join his new brand of anti-austerity politics. Since Labour seems no longer even to want Conservative voters to take a second look at them, those three pools are where they're going to have to get new voters from. As well as from last time's non-voters, of course, but that's difficult to test as yet.

How are we doing with those prognostications? Well, not bad on the overall picture, but with some interesting caveats on the detail.

The first thing to say is that there has not been much of an overall 'Corbyn bounce'. That may not be surprising, given his absolutely torrid first week in charge, during which his project looked on several occasions like blowing up before it had even begun. But it is of note. There's a teeny, tiny bit of a lift discernable in the four national polls and one marginal survey we've seen. But it's well within the margin of error, and therefore it might not be there at all. ComRes showed a 0.5 per cent swing to the Conservatives from their last poll; YouGov, a 1.5 per cent move to Labour; ICM, the same; and Opinium, a 0.5 per cent swing to Labour since the General Election. An ICM marginals poll, taken in Conservative seats with slender majorities, showed a 2 per cent swing to Labour. All that left Labour a long way back - eight points behind the Conservatives if we take the YouGov figures, when they led by one point the day after Ed Miliband was elected leader in 2010. This should worry you a great deal if you are cheering Mr Corbyn on to ever-greater heights. Every Labour leader in the past fifty years has got a polling 'bounce' on taking up the reins - except James Callaghan and Michael Foot. And look what happened to them. Even Michael Howard and (especially) David Cameron gave their parties a lift when they first became leader. Now? Almost nothing. So our first presumption - that there would be something of an early Labour bounce - seems to have gone awry.

We also have quite a lot of evidence that the policies Mr Corbyn is most associated with are often very, very unpopular. Rail and energy nationalisation are popular, as are rent caps and more taxation on the rich. Fair enough - those causes have for many years been seen as pretty good ideas by a majority of the public. But the numbers on national security - Trident, the military, patriotism - and on the economy are dire. A majority in the ComRes data actually think Labour's leader is a positive danger to national security. The overall picture, away from specific policies, is just awful. A majority of voters trust Mr Corbyn only on the NHS, and that by a mere seven points: far worse figures than Mr Miliband's even at his nadir. He is underwater on the economy by 17 percentage points, even at the start of his leadership. Leftists are enthused, something that may well help Labour survive in an era of local campaigning and Individual Voter Registration. But worried centrists and swing voters alienated. The really worrying thing about this is how UKIP voters have responded. ORB data which only covers what voters are thinking of doing, or might do, has to be treated with caution. But it's revealing. 20 per cent of past UKIP voters are now more likely to vote Labour; but 42 per cent are now more likely to vote Conservative. Last week we thought this might be where Mr Corbyn might find some of the new votes Labour desperately needs. This sort of finding undermines even that thin hope.

Overall, Mr Corbyn is deeply divisive. Particular policy positions are popular: the overall offer, and philosophy, certainly is not. So Labour has learned nothing from the failures of the 2015 General Election, in which it made precisely this error of focusing on a 'retail offer' rather than a deeply resonant series of stories and emotions. Mr Corbyn starts off on YouGov data with a negative rating of -18: 38 percentage points below where Mr Miliband began as leader (that -18% figure is backed up by the ComRes numbers reweighted for those actually likely to vote). Those numbers alone ought to frighten Labour headquarters out of its wits, for personal ratings are more likely to fall than rise in the medium term. His numbers are somewhat above the truly abysmal that Mr Miliband's plumbed: given how far ahead of Mr Corbyn the latter started, similar numbers are probably not far away.

And what about attracting actual votes from among those who voted for other parties in 2015? Well, here there's some solid votes in ballot boxes to go on. Last week there were two council by-elections in Haringey, one in Ayr, one in rural North Yorkshire, and one in Cambridgeshire. Labour seems to have attracted votes from the Greens to increase their large majorities in the first two contests, remained in third place in Ayr East, and got nowhere in Yorkshire and Cambridgeshire. The move from the Greens, combined with numbers buried on page three of the ORB report showing 'others' (often Greens) shifting over to them, seems concrete and well-evidenced. But Labour stood still in Yorkshire and Cambridgeshire - while the Liberal Democrats pushed onwards with their local revival - and went backwards in Ayr. Labour's votes may be collecting exactly where the party doesn't need them: in liberal, multi-ethnic areas of big cities, where they're already well ahead. That bodes ill for them.

However, that Scotland result is interesting. That was about a seven point swing since the last time the seat was contested, a similar move from an Edinburgh (Leith) by-election the week before. Pretty bad. But nowhere near as bad as Labour's polling leading up to Mr Corbyn's election (opens as PDF), which implied a 14-15 per cent swing to the SNP since the Scottish local elections held in 2012. This, taken with the Scottish subsample in the YouGov figures - 46% to 28% for the SNP - might imply something of a (very limited) recovery by Scottish Labour. The two figures for swing that we can derive from those two sources aren't that far apart that they're incompatible. The problem with a post-2012 swing on that scale? It only means that Labour will be buried by an SNP landslide next May, on similar lines to 2011, rather than utterly, utterly wiped off the face of the planet. Scottish Labour's terrified of coming third next year, behind the Conservatives - in our view rightly. These first three data points - just two, perhaps unrepresentative by-elections and a tiny unweighted chunk of a UK poll - only hint at something less than oblivion. And in a Westminster contest, such a small swing back to Labour might give them back just three or four seats. Nowhere near enough to make a real difference. We doubt any of this 'hopeful' news comes as much of a consolation for them.

To sum up: no polling bounce; personal and ideological unpopularity; divisiveness; votes piled up or won back where Labour can't really use them. Not a great overall picture. Diving deeper, there are some crumbs of comfort in Scotland, which may help Labour avoid humiliation in next year's Holyrood election, but probably wouldn't in a general, UK-wide contest. There are definitely signs of Green voters coming over to Labour - though that won't help the party win many seats. And our theory about UKIP voters - that they might be brought back by an apparently old-fashioned, full-blooded, nostalgic socialist ticket - looks thinner by the day. One thought in parentheses is this: perhaps voters are rather more ideologically- and policy-driven than we thought. Certainly Mr Corbyn's leftward shift seems, initially, to be attracting back some SNP defectors, and alienating English UKIP supporters. Still, it's very, very early days. We need to know more. Much more. We need properly-weighted and balanced national polls from Wales and Scotland. We need qualitative and quantitative numbers on UKIP voters. We could do with lots more by-elections to confirm what we think we've seen already. But the first straws in the wind are deeply suggestive.

The real danger here is that Labour is going to be misled by what happens in next May's elections. If they attract a great deal of Green support (and second preferences), as well as left-leaning ex-Liberal Democrats, they may win the Bristol and London Mayoralties. If they can at least staunch the political bleeding in Scotland, they may avoid a total wipeout at the hands of the SNP; there is very little chance of them losing actual power in Wales, even if they do badly, because Plaid Cymru remains too weak to mount a sustained challenge, and there are no alternative coalitions available to replace Labour altogether. The English local election results, which are likely to be pretty grim, may get lost under a tide of relief at winning back City Hall by the Thames, and dodging total extinction at Holyrood.

But all the while, the storm clouds will be building up. Green defections, and SNP second thoughts, are not enough for Labour to hold the line and remain standing even where it landed in 2015. These numbers amount to a warning that things are even worse than we thought last week.

The increasingly thin ranks of left-leaning Britons comfort themselves by scrolling through Facebook and Twitter, reading comments written by their friends, for their friends - about how principled Labour's new leadership is, how silly the Prime Minister looks while beset by revelations about his youth (which he does), how social media lets them get the truth out (to each other), and how new lefttish alliances are now imaginable. It is a comfort blanket, twisted ever tighter while the political skies darken outside, and the ominous feeling of oncoming winter deepens.


  1. Firstly, for someone who is a historian of contemporary Britain, I'm shocked at how much weight you're giving polls when we all know how badly they performed in the last election. Not to mention, 1 week in, with 1 week of awful coverage in the press about completely farcical issues (I can only presume a bias/Westminster agenda). Is it any wonder he is polling badly?

    And you base your whole article on this and what looks like your own opinion? Don't you sometimes wonder how much better Corbyn might be doing if we didn't have such a useless misquoting right-wing media? E.g. Corbyn doesn't do his top button up; outrage. But Cameron wants to bomb the Middle East; meh.

    I love politics, but I sometimes feel people miss the bigger picture!

  2. Thanks for the comments. It is of course that the polls got GE2015 pretty wrong, but before that at GE2001 and GE2005 they became progressively more accurate. But one of the points we're making in this blog post is that we don't need just polls - we've got local by-elections (in which the Conservatives continue to run ahead, overall) to go on also. On the press, yes, this might well be the case - and we'll blog about this in the future - but this wasn't about the press, but about popularity in the political system as it is right now. He probably would do rather better without facing such a hostile press, but that isn't going away. Meanwhile, *all* the data we have points to Mr Corbyn being pretty unpopular. We're on the lookout for evidence to the contrary, and when it comes in, we'll throw it into the analysis.

  3. These analyses are some of the more rigorous anti-Corbyn arguments I've read, but still miss the point in significant ways I think. As much as it flatters the writer to frame the debate as empirical, rational psephologists vs. naive, deluded Corbynites, I think there are fundamentally different political philosophies in play, each valid by their own terms and invalid by the other's. For example, the methodology here is rooted in a set of assumptions - that the political can be reduced to analysing electoral/opinion poll/focus group data with a view to winning the next election at any cost - against which "Corbynmania" is obviously a reaction. The latter is banking on a social movement geared towards changing the very terms of politics in this country in the longer term, a notion that might be "utopian" in the sense of being undeniably hard to achieve, but cant be somehow disproved before the fact by this kind of short-termist microanalysis of the little available polling data. I think a more interesting argument might result if rather than endlessly trying to prove in Blairite fashion that Corbyn will be "unelectable" in 2020, the moderates or centrists or whoever took seriously the Corbyn phenomenon as intellectually justifiable on its own terms (which it is).