The professor will see you now: tribalism
4 min read
Lessons in political science. This week: blocs
One electoral change has been unrelenting: the growing – in fact accelerating – refusal of the electorate to cast a ballot for either of the two governing parties.
Forgive me, father, for I have sinned – because those are not my words. The first sentence of this article was published in 1977, almost half a century ago.
I deploy this slightly hackneyed columnist’s trick not because anyone can doubt that British electoral politics is in a state of flux, but merely to make the point that what we are seeing has been a long time coming.
That 1977 sentence comes from one of the classics of the psephological canon – Ivor Crewe et al’s Partisan Dealignment in Britain, 1964-1974 – and it noted that the declining vote shares gained by the two major parties had begun in the 1950s. Labour and the Conservatives currently find themselves on the wrong end of a 70-year trend.
Crewe et al was one of the first major pieces of research to note voters’ declining sense of unquestioning tribal attachments to parties – what psephologists call partisan identification. In 1964, 71 per cent of people said they had a very or fairly strong Labour or Conservative identification, with 40 per cent saying ‘very’. By 2005 the latter figure had dipped below 10 per cent; in 2024 it was seven per cent.
Beneath the surface, though, there are still pockets of much stronger party identification. Supporters of Reform UK in 2024, for example, had levels of identification akin to those from the 1960s; there are also high levels of party identification among SNP, Plaid and Green supporters. But overall, whereas in 1964, 43 per cent of the total population supported a party ‘very strongly’, in 2024 it was 13 per cent.
This all leads to voters being much more willing to swap between parties. Over 40 per cent of people voted for a different party in 2024 than they did in 2019; the highest rate of vote switching since 1964.
(As an aside: you will often, as here, see 1964 used as the start date for such calculations. There is nothing special about 1964, apart from that it was when the British Election Study began; it is therefore the first election for which we have reliable panel data – that is, data that tracks the behaviour of individuals over time, which you need for statistics like this. What this sentence actually means is that 2024 saw the highest rate of vote switching since at least 1964.)
In partisan terms, this is obviously bad news for Labour and the Conservatives – who can no longer reply on deep reservoirs of support – and good news for almost everyone else. But it also has other electoral consequences, often largely unnoticed. Conversion becomes more important, not just mobilisation. Elections become less stable; outcomes previously seen as impossible become possible, plausible even. Safe seats become marginal. Landslides become more likely.
Voters aren’t entirely free-floating, though. In some ways what is remarkable about that vote-switching figure is that it is still below 50 per cent; despite five years of political turmoil and change, the majority of people stuck with the same party in 2024 as they did in 2019.
And perhaps more importantly, as new research recently published in Parliamentary Affairs shows, most of the switching that did occur took place within voting blocs that map on to attitudes to Brexit. Some 70 per cent of those who voted Leave in 2016 backed a right/conservative party in 2024; 84 per cent of 2016 Remain voters backed left-liberal parties in 2024. There was some movement between these blocs, but voters were roughly two-and-a-half times more likely to switch within blocs than between them. Voters may no longer swim in schools, but the parties fish in different ponds.
Further reading: I Crewe et al, Partisan Dealignment in Britain 1964-1974, British Journal of Political Science (1977), J Griffiths et al, The Brexit realignment amid electoral volatility: The role of party blocs in the 2024 General Election, Parliamentary Affairs (2025)
