Epistemological nightmares

This week: some number crunching on UK electoral history, some notes on a list of lists of lists, and, finally, that new rail map of Birmingham.

I’ve been, you’ll have noticed, increasingly bullish about Labour’s chances of winning a decent majority at the next general election. Keir Starmer may not have either the policies or charisma to inspire the kind of response that – though it’s easy to forget this, given everything that followed – greeted the arrival of Tony Blair in the mid 1990s. But there’s ample evidence of a widespread desire to throw the Tories out of office, and of the sort of tactical voting that will – thanks to First Past the Post – translate into bountiful Labour gains, no matter how underwhelming the voters find the Labour leader.

Set against that, though, there’s just one niggle. To win a majority after its catastrophic 2019 result, Labour would need to gain 124 seats. This is not easy – it’s something the party has only managed three times in its history, in 1929, 1949 and 1997 – and the first of those was so long ago, coming at the tail end of another party system and being the first election held under the universal franchise, that I’m not convinced it even counts.

This doesn’t mean it’s not possible – precedent has been a pretty poor guide to the future these past few years, and it’s hardly as if furiously anti-Tory voters are going to sit on their hands just because past election results say they should. But it is, nonetheless, a worry, that Keir Starmer – Keir Starmer – would need to do as well as Clement Attlee or Tony Blair, just to get a majority of 1.

Labour general election election results since 1945, ranked.

What else can a sort of differential approach to electoral history teach us? Let’s find out.

1. The Tories have only lost more than 10 points in vote share twice in recent history – again, in 1945 and 1997. As things stand, their polling average is somewhere around 27%, down 16 points on the last election result. So either they’ll recover some ground, or they’re on course for the biggest collapse in vote share in the party’s modern history. There is no third option.

Tory general election election results since 1945, ranked.

2. The 1945 election really was a hell of an achievement for Clement Attlee’s party – the only time Labour’s vote share gain made it into double figures; 239 seat gains in a sitting; the Tories’ worst decline in both vote share (down 13.7 points) and seats (down 219) – and all it took was the Second World War.

That’s the sort of result some polls have Labour on course for now, and yes, put like that it is a bit easier to understand the cynicism.

3. That 1997 would be the next worst Tory, and next best Labour, election, fits the narrative too, and is – given the absence of either a decade without elections or an existential war – a better comparator for the next election. 

What’s odd, though, is that it’s not actually the best election for Labour in terms of the increased vote share: that was 2017, when it gained 9.6 points, only 0.4 less than in 1945. Okay, it didn’t win, in large part because that was also the Tories’ second best election by vote share (they gained 5.5 points) – all of which is presumably what happens when the vast majority of the electorate simultaneously go “Oh god not that lot”, but can’t agree on which of two lots they mean. But the oddity, the fact it doesn’t fit the received narrative, feels worth noting nonetheless.