Are Britain’s voters fickle or stable? It depends on the way you look at them

It was a time of turbulence, it was a time of stability.

When the definitive account of the current election comes to be written, my candidate for its title is: A Tale of Two Islands.

To explain: In the past three months, support for each of the Britain-wide parties has fluctuated as never before. Individual polls have reported Conservative support to be as low as 27 per cent and as high as 47 per cent. Labour has ranged between 21 and 34 per cent, the Liberal Democrats between 12 and 23 per cent and the Brexit Party between 2 and 17 per cent.

Past elections have seen big  movements in the support for particular parties, but never before has every party experienced such volatility. The ranges narrow a bit if we look at a rolling average of the figures for the main pollsters, to iron out sampling fluctuations – but not much. This autumn, Britain’s electoral tribes have set new records for fickle behaviour.

However, if we look at the figures another way, a very different picture emerges. Instead of counting the support for each party separately, let us allocate each tribe to one of two islands – Brexit Island, where the Conservative, Brexit and tiny Ukip tribes live, and Rethink Island, where the Labour, Lib Dem, SNP, Green and Plaid Cymru tribes look warily at each other.

What this reveals a picture of remarkable stability. The population of Brexit Island since the summer has remained 48 per cent, plus or minus two per cent, while that of Rethink Island has remained at 51 per cent, again plus or minus two per cent. Statistically, the story is one of no change. Plenty of voters have changed tribes, but hardly any have moved between the two islands.

What, then, should we make of the latest polls reporting a narrowing of the Conservative lead? Could we see a repeat of what happened in 2017, when Labour surged, the Tories sagged and Theresa May lost her Parliamentary majority?

We should not get carried away. Taking an average of the six polls conducted towards the end of last week with the same polls a week earlier – and so comparing like with like – we find that Conservative support is unchanged, while Labour has gained two points at the expense of the Lib Dems. Once again, there has been a movement between tribes, but not between islands.

Furthermore, the trajectory of party support during the 2017 campaign no longer offers Labour much comfort. By this stage last time, Labour’s surge had already happened. The polls conducted a fortnight before election day reported much the same figures as their final, election day predictions. True, the polls mostly overstated the Conservative lead; but they probably did so throughout the campaign. There was little or no movement in the final ten days (or, more accurately, no NET movement – the small numbers of late switchers between parties cancelled each other out).

In short, there is little in the latest polls to undermine predictions of a clear Conservative victory.

That said, nothing is certain. For a start, the last two elections warn us against excessive faith in the polls. Currently they suggest a Tory lead of 11 per cent. They need to be just two points too high for the Conservatives and two points too low for Labour for the gap to be just seven per cent. Depending on seat-by-seat variations and possible tactical voting, that kind of lead on December 12 could make for a nerve-wracking election night.

However, if the average of the latest polls is about right, then there must be a material change between now and December 12 for Boris Johnson to be ejected from Downing Street. And that would require getting on for one million voters to change their minds late (which did not happen in 2017) and for many of those voters to move not just between tribes but between islands (which has not happened in the past three months).

That is not to say neither can happen. One of the big lessons from 2017 is that the past offers an unreliable guide to the future. This time could be different. How?

Tactical voting could make a difference; however, Deltapoll’s surveys of key seats suggests that Labour supporters are far more happy to switch to the Lib Dems than vice versa. Many Lib Dem supporters in Con-Lab marginals are torn between their opposition to Brexit and their dislike of Jeremy Corbyn. We shall see on election night whether Labour has done enough to overcome this aversion in the seats that matter most.

Another possible problem for the Tories is the large number of young people, and especially students at universities in marginal constituencies, who have recently joined the electoral register. Again, in two weeks’ time we shall know whether this has made a significant difference.

The big prize for Labour, however, would be to win back some of their pro-Brexit former supporters. Analysis of the huge 100,000 sample for YouGov’s MRP exercise last week suggests that of the 3.5 million voters who backed Brexit in the 2016 referendum and Labour in 2017, 1.5 million have fled to Brexit Island, with most of them joining the Conservative tribe.

Labour has lost the battle to persuade them of the virtues of its plan for a new referendum. There is, though, another battle that may not yet be lost. Deltapoll’s latest survey for the Mail on Sunday asked people which were the most important issues facing “you and your family”. Among the electorate as a whole, leaving the EU came second (chosen by 28 per cent) behind health (40 per cent). But among those who voted Leave in 2016 and Labour in 2017, Brexit comes fifth (just 18 per cent), behind health (48 per cent) but also the economy, welfare benefits and pensions (all between 19 and 23 per cent).

These figures suggest one way for Labour to defend its vulnerable seats in the North and Midlands: deploy its local campaigning not to defend the party’s stance on the EU but to focus single-mindedly on Labour’s traditional issues: the NHS and fighting inequality. And never to mention Corbyn.

This blog first appeared in The Article