6 April 2021

A shock poll from Hartlepool is less revelatory than it seems

The headline result is a nightmare for Labour in general – but it doesn't tell us much that we didn't already know.

What to make of a new Survation poll that shows that the Conservatives are on course for victory in the Hartlepool by-election?

Survation’s headline result is close to what would be the worst possible scenario for Labour in general – the party increasing its share of the vote from the 2019 general election, but losing the seat due to the Conservatives gaining votes from the Brexit Party – because it suggests that Labour will have to do significantly better in terms of winning votes than it did in 2019, simply to do as poorly as it did at gaining seats in 2019.

In addition to the general undesirability of drawing snap judgements about any single poll, however, any analysis of what the Hartlepool by-election result “means” needs to be taken alongside the local elections that will take place the same day. In Hartlepool that includes elections for the local council and the Tees Valley metro-mayor.

If, say, Labour wins control of Hartlepool Borough Council, Labour’s Tees Valley mayoral candidate Jessie Joe Jacobs finishes ahead of the Conservative incumbent Ben Houchen, but Paul Williams fails to win the parliamentary seat, then that would suggest that the Labour leadership had bungled badly in its choice of candidate. That’s not an existential crisis but one that would require a change of approach in the leader’s office. If Williams overperforms or simply does as badly as the Labour party in the area, then that would suggest the problem is more fundamental, and therefore more worrying.

These hypotheticals are all wildly premature, though, because constituency polling in the United Kingdom is notoriously difficult and has been wrong more often than it has been right. Even when it has called the “headline” result correct, it has tended to be way off as far as the detail is concerned.

Equally importantly, when constituency polls get it wrong, they have tended to do so in fairly predictable ways. They tend to over-count the most politically engaged voters: whether that be significantly overestimating the strength of the Liberal Democrats in Conservative constituencies that voted to remain in the European Union or overstating the size of Vernon Coaker’s personal vote in Gedling. In both cases, constituency polling picked up a real phenomenon: the Conservative party did lose Remain votes to the Liberal Democrats, and Coaker did overperform compared to the Labour party nationally and regionally. They simply overestimated the scale. And you can construct a perfectly plausible case in any direction that this poll might be wrong for similar reasons: it could be wildly overstating how much of the Brexit Party vote will go to the Conservatives, it could be hugely overestimating the extent to which the smaller progressive parties will see their votes squeezed by Labour (no other party of the centre or the left has a vote share above the margin of error in this poll). It could be overstating both the level of support for the Conservatives and Labour.

That makes it hard to draw any valuable lessons from Survation’s latest poll, because the trends it captures (the Conservative Party gaining votes due to the end of the Brexit Party as a meaningful electoral force, Labour gaining votes on 2019) are ones we already know exist. What we don’t know is which one of those trends is going to be more important in May 2021 in Hartlepool, let alone July 2023 or May 2024, the two most likely dates for the next general election.