Sunday, 22 January 2017

Labour will probably hang on in Copeland and Stoke

Now 2017's up and running properly, two pending byelections now due in February will give us some indication of how British electoral politics is looking. Labour is defending two seats - semi-rural Copeland in Cumbria (above), mainly famous for the Sellafield nuclear reprocessing plant, and the rather more urban Stoke Central in the potteries. So what's likely to happen?

Start here: nationally, Labour is in a mess. You knew this already, of course, but the party's cruel dilemma over how to approach the UK's exit from the European Union - no fault of Labour's leaders, in and of itself - has made things even worse than they might have been. Yesterday former Liberal Democrat leader Nick Clegg rightly warned that Labour's vote could be drawn away in two directions, from the Right by the United Kingdom Independence Party, able to call for a rapid and ragged Brexit, and from the Left by the Liberal Democrats themselves, ready and willing to pose as the party of Remain. In two seats that voted heavily to Leave the EU, Labour's liberal (but confused) stance on immigration and its at-least nominal commitment to Remain will hurt them. Jeremy Corbyn's deep unpopularity will hurt them. Their gradual cultural estrangement from their own voters will hurt them. Their own divisions will hurt them. There is even panicked talk - almost certainly overdone, if it is not deliberate expectation-managing spin - that Labour might come fourth in Copeland. You get the picture.

So should we expect Labour to lose these seats? Not so fast. It's actually incredibly difficult for sitting governments to win seats from the official Opposition. Not unreasonably, voters like to give any administration a bit of a risk-free kick, just to keep them up-to-date on who's boss. Most seats' holders will usually have a reservoir of well-liked and hard-working party activists to draw on for candidates, while parties without such a rich history in any seat will have a much sparser bench of talent. The government down in London might face any sort of popularity-eating problem in the few days leading up to any poll - think of the recent National Health Service beds crisis - and that would hit their poll on the day.

Note in this respect that the Conservatives - Labour's main challengers in Copeland - are not riding all that high in the polls compared to their General Election victory just a couple of years ago. On average, they stand at about 40 per cent - so they're only two per cent or so up on 2015. Labour have fallen a little bit more than the Conservatives have risen: on a six-poll rolling average they're attracting about 27 per cent of the vote, down about four per cent since the last national vote. The swing implied? Just over three per cent. The swing needed for the Conservatives to gain Copeland is 3.23 per cent. Even if we just took the opinion polls at face value - and they are absolutely excellent for a government in its second term - they wouldn't quite make it. Nor is there much sign in local by-elections that the Conservatives are storming ahead in real voters. Over the whole of last year, there was a small move from Labour to the Conservatives, of less than one per cent - nowhere near enough to give the Conservatives a gain here. And with a massive nine per cent or so swing needed for the Conservatives to grab Stoke Central - a seat in which they would have to come from third to first - that doesn't seem all that likely either.

And UKIP? UKIP's record in traditional British elections is awful. They can win seats on regional lists in European elections, taking advantage of an election and a system that enormously favours them, but they have never won a Westminster seat where they weren't taking advantage of a sitting MP defecting to them - not even Eastleigh in 2013, at a time when the Liberal Democrats should have been sitting ducks. UKIP has been experiencing a collapse of its vote in recent council byelections (perhaps because it relies on voters who are less likely to actually turn out than other parties). And it's gone absolutely nowhere in previous Westminster byelections, either since the last General Election or the referendum. New UKIP leader Paul Nuttall is standing in Stoke, and he may think that he's the answer to Labour (or ex-Labour) voters' prayers, but his ratings are very poor - even among those voters who have an opinion about him, since 'don't know' is easily the winner when pollsters ask people about him. At least he's turning up and putting his trousers where his mouth is, so to speak - unlike previous UKIP leader Nigel Farage, notoriously wary before he'd commit himself to likely defeat in byelections. But he's not likely to be as effective as many media boosters think he will be.

The history of these contests tells us a lot about how unlikely Labour losses are. The last time a governing party gained a seat was 1982 - at a time when Labour was even more split than it is now. And there have only been two such occasions in the last sixty years. Yes, Labour did lose a couple of contests (Bermondsey in 1983 and Greenwich in 1987) to the famous Liberal/ SDP byelection machine in the 1980s, but since the Liberal Democrats are very unlikely to win either of these seats, a Richmond Park-style shock seems even less likely than Conservative or UKIP victories. If we look at the data on all such contests since 1983, as polling expert Matt Singh does in this post about Copeland, you'd expect Labour's majority to increase, not decrease, even adjusting for present polling. Only using data from the recent Sleaford by-election - for a seat won in 2015 by the Conservatives and not Labour - can you make Copeland look competitive. If history holds - and we accept that we live in uncertain electoral times - these seats should stay Labour.

Nor does local intelligence from either Copeland or Stoke really speak to the kind of earthquake that Labour losses would require. The party is unpopular, sure: many of its traditional voters didn't like the Remain campaign much; in Copeland, it faces the additional headwind of Mr Corbyn's past opposition to nuclear power. But there are pluses for Labour too, not least local anger about the state of the local hospital in Copeland. Have we really heard the explosive detonation we'd expect if there were to be a total collapse of Labour support - have we felt the rushing drumbeat that will mark the moment the dam bursts? We don't think so.

Now perhaps all this doesn't matter all that much. With Donald Trump in the White House, likely to instigate a naval faceoff with China just as quickly as he can - and then to eyeball the Russians over Iran's nuclear programme - what happens in a couple of vacant British Parliamentary seats isn't really anyone's idea of the first order of business.

Nor are these byelections likely to change Labour's political trajectory. One gets the impression that Mr Corbyn could trail by twenty five points in the polls and lose hundreds of council seats in May - and he'd still be re-elected against any challenger. The results are also unlikely to matter much even in narrow electoral terms. If Labour hangs on to these seats, as we think they will, they probably still face a very severe drubbing in the next General Election. If they lose one (and especially both of them), as they still might, they probably face a defeat that cannot even be described as a catastrophe. Such a result would be a sign of impending cataclysm: the opening of a political black hole from which there may be no escape. But in a way: so what? European social democracy as a whole is in rapid retreat, and the speed of its recessional is not particularly important.

But the point remains that we will learn something from two Labour holds: that Labour's voters are not quite giving up the ghost just yet, and that although they dislike the party's leader and are happy to moan about him on the doorstep, the Conservatives and UKIP are seen as an unappealing alternative. And that tells us something else: that however bad things seem right now, Labour will probably still recover - one day.