Polling analysis: CON’s getting biggest swings where it matters most

Writing a thread on why Labour might do even worse than headline polling figures suggest feels uncomfortably like kicking a man when he’s down. However, if that’s what’s happening then it needs reporting and interpreting; I am only the messenger. And it is happening. Poll after poll has reported differential swings across regions, social groups and Brexit alignment.

I’ve therefore looked at all the polls published in March to see if we can make sense of what’s going on and where it would lead. The figures I’ve derived are – unless otherwise stated – based on an average of the five companies to have published polls in March’s own average figures for the month.

The national baseline is that the average headline figures for March (together with the swing since the 2015 GE) were:

Con 42.7 (+5.0)
Lab 27.1 (-4.1)
LD 10.2 (+2.1)
UKIP 9.8 (-3.1)

This represents a net 4.5% swing from Lab to Con – though in fact most of the actual movement is the result of voters switching to and from third parties.

This of itself would be bad enough for Labour. Electoral Calculus predicts a Conservative majority of 106 on those vote shares, using the current boundaries. The regional swings however, produce a multiplier effect.

Unfortunately for Labour, their smallest swings mostly come where it helps least. The one exception is Scotland. To describe this as a disaster zone for them is putting it kindly. The way things stand, Scotland has more chance of winning the World Cup than it does of electing a Labour government. Even after the losses of 2015 and 2016, the trend remains downwards, with the firms reporting an average support of just 12%. Two polls (those by ComRes and Mori), had Scottish Labour in fourth place, behind the Lib Dems and in single figures; Mori gave them just 6%. This might be bad news for Scottish Labour councillors but the good news as far as the general election goes is that the party can only lose one seat north of the border, which of course it would.

In England, however, the picture doesn’t pan out so well. A London-only YouGov poll published yesterday found support of Lab 37 / Con 34 / LD 14 / UKIP 9. (These are in fact similar figures to the aggregated London subsamples from the nationwide polls, which is reassuring as to their reliability for other regions). This represents a notional Lab-Con swing of just under 3%, so less than the national average – though in reality it’s more a secondary effect of a 7% Lab-LD swing, mostly in Inner London. Unfortunately, London has an unusually large number of close Lab-held marginals and even a swing of that size would sweep eight seats into the Con column, though in all probability this would be offset by some Con losses to the Lib Dems.

Elsewhere, both Labour and Tory support across the rest of the South – a huge area of the country, almost entirely Blue – is virtually unchanged. Polling for both parties puts them within 1% of where they were at the 2015 general election.

Instead, the areas where the swings are greatest are those which are packed with middle-target marginals: the Midlands, Wales and the North.

Not all firms disaggregate Wales from the Midlands, presumably because of its small population and – unlike Scotland – its tendency to swing more-or-less in line with England. But taking the three regions of Wales, the East Midlands and West Midlands together, produces a swing over that super-region of no less than 7.6%. Put another way, it turns the tidy 5.5% lead that Con had in the 2015 election into a monster 20.7% lead.

Likewise, while the polling is unusually inconsistent, even for subsamples, in the picture painted by the polls for the North, we can see something similar. In 2015, Labour held a 12.4% advantage; that’s now been turned into a lead of about 1.5% for the Conservatives: just shy of a 7% swing. In contrast to London, Labour holds only ten seats in the northern super-region with majorities of less than 3k but holds 22 with majorities of between 3k and 7k (not all over Con).

All this would be bad enough but two other factors are likely to multiply the differential even further (though we should be aware of some element of double-counting here).

Firstly, the swing to Con has come almost entirely from the working class. Unlike regional voting, the baseline of the 2015 election here is, for obvious reasons, less secure. Even so, Mori carried out a sizable post-election poll which is the best information we have. Using that, it seems that since the election, the ABC1 group has swung to the Tories by just 1.6%, whereas the C2DE group has swung by some 8%. As the Tory target seats as the party heads from comfortable win into landslide territory will tend to become increasingly working-class, the vote is again exactly where TMay needs it.

And secondly, returning to the top of the polling data, polls are consistently reporting that between a quarter and a third of the 2015GE UKIP vote has defected to the Conservatives. It was the small and medium northern and midlands towns which won it for Leave and while the match isn’t a perfect fit – the Labour seats are too diverse for that – it’s still likely to have a boosting effect to the swing to Con where it matters.

Is there any good news at all for Labour? Yes: two items. One is that the Tories are currently hitting the sweet spot. If the polls shift, they probably won’t be; the multiplier effect works both ways. And the other is that there is probably quite a simple solution to bringing about such a change.

David Herdson

Follow @DavidHerdson

Tweet