Keir Starmer left little doubt about his intentions this conference season when he spoke from a podium in Doncaster with a red brick wall as his backdrop.
The Labour leader was introduced by Ruth Smeeth – who lost her Stoke-on-Trent seat to the Conservatives in December – and his address was targeted not at the Labour intelligentsia watching on Twitter, but rather at people like Smeeth’s former constituents, from the parts of Britain that once made up the party’s “Red Wall”.
Starmer focused on issues of family and identity – typically safe territory for the Tories. This was a leader of a party conscious of the electoral Everest it faces in the years to come.
Some of the long-held Labour seats lost in 2019, where the Conservative majority is in the hundreds, could be easily won in a tight future election. However, many of the seats Labour lost in December were not narrowly won, but by 4,000-, 5,000- or 6,000-vote majorities.
The question of Scottish independence could make that climb even steeper. England has not always been Labour’s natural home. In 2005, though the party won a third consecutive general election victory, in England it came second in the popular vote after the Conservatives. Labour’s image problem in both rural and small-town England is nothing new, and losing Scotland will only add to the challenges facing an already humbled party.
There have been some signs of recovery over the course of the year. Labour has substantially narrowed the overall Conservative lead, with the parties tied in some polls. Boris Johnson now trails Starmer in some surveys on perceived competence, and on the question of who would make the best prime minister. Such figures would suggest a close result in an election.
But the reality is more complex.
Labour is currently ranging between six and eight points up on its 2019 performance, according to the Britain Elects poll tracker. The Conservative Party, by contrast, is down by four points.
Much of Labour’s gains can be attributed to two strands of voter movement. The first is the number of adults who say they would now vote Liberal Democrat – a figured that is much lower than the 2019 results. Polls suggest a significant portion of those who voted for Jo Swinson last December have shifted to backing Starmer.
The second strand is a sustained fall in the confidence of Conservative voters. Surveys now show that since the heights the party attained in polling in March and April, the number of 2019 Conservative voters who are now unsure as to how they’d vote an election has grown month on month. One YouGov poll, which puts the two main parties on 40 per cent each, found nearly one in five Conservative voters from 2019 were unsure about what they’d do with their ballot.
These two strands do not necessarily suggest that Labour will take the lead, nor that the party is poised to win an election. Often, when there isn’t a general election approaching, voters for the incumbent party either register for the protest choice (be it Ukip in 2012-16, or the Lib Dems in 2001-10), say they are undecided or that they wouldn’t vote, or refuse to declare their voting intention.
What these Labour “gains” mean, then, is not that the party has gained Conservative voters, but that it has been successful in both keeping its base more enthused than that of the incumbent, and in attracting former Lib Dem voters.
Further analysis of the most recent polls* shows these national swings are not uniformly replicated in each of the UK’s nations and regions. In southern England, for instance, Labour is up by more than seven points – which, if replicated in a general election, would represent Labour’s best performance in the South since 1997.
Though this might be indicative of a broadening of Labour’s appeal, a seven-point gain (and a 5.6-point swing) wouldn’t win Labour many seats. There are only seven seats in southern England where the party lost by fewer than ten points in 2019.
In Wales, support for Labour is virtually unchanged from 2019. It’s clear the jump in support for Plaid Cymru, up from ten per cent in December last year to 15 per cent in September, is sapping whatever gains Labour is making from the Lib Dems there.
Where Labour needs to build support in order to make significant net gains is in those northern Red Wall marginals.
Red Wall voters are not a homogenous collective, however much some commentators group them under broad labels such as “left behind”. There is, however, one common theme that runs through those places that swung Conservative in December. It is their affinity not to a particular party, but to Brexit.
Brexit has upended the party divisions that have shaped this country since the English Civil War. Where it was once Cavalier versus Roundhead, or Labour versus Conservative, that shaped political opinion, now Leave versus Remain is changing how Britons vote.
From what data is available it appears that the tribalism of Leave and Remain, though seemingly reduced since 2019, remains a significant issue, and it is this that is holding Labour back.
The data shows the Leave-heavy Red Wall seats won by the Conservatives last year are, albeit marginally, still blue. Ian Simpson of the Electoral Reform Society analysed Opinium surveys taken between the start of June and the end of August and found that 46 per cent of voters in the constituencies that were won by the Conservatives in 2019 would vote Tory again now – down just two points on the last election. The number who would vote Labour is up six points, at 44 per cent.
While the polls have moved on since this analysis, and the gap has narrowed further, it’s clear that the hold the Tories have on the Red Wall seats remains a huge challenge for Labour. But it is one the party must overcome to have a realistic chance of winning a general election. The electoral maths of the UK no longer works in Labour’s favour. Even if vote shares tied nationally, Labour would still lose based on the regional swings the polls are showing. The Red Wall may not be the only place Labour seeks to win back seats, but without the north, the party cannot win an election.
*where the subsamples are of a satisfactory size and provide consistent results.