General election: How voter loyalty that lasted 100 years is vanishing – making 2019 impossible to call

In the 1900 general election, 99 per cent of the votes cast in Great Britain went to Conservative, Labour and Liberal candidates. In the 2005 election, the equivalent figure was 95 per cent. Whatever the variable fortunes of the individual players within the party political establishment, one thing was crystal clear: outsiders were unwelcome in general elections.

However, five years later, non-Conservative/Labour/Lib Dem candidates achieved a record 10 per cent of the vote in Britain; these included 920,000 votes for Ukip, 564,000 for the BNP and 491,000 for the SNP. This 2010 election result left the political establishment unshaken and only slightly stirred; but the 2015 election was totally unprecedented – a full 23 per cent of all votes cast were not for the three establishment parties, including 3.86 million for Ukip, 1.45 million for the SNP and 1.15 million for the Greens.

Two years later, many thought that our politics had returned to business as usual. There was a great surge of voters back to the Conservatives and Labour in the 2017 election, who recorded the highest joint vote percentage share in any election since 1970. Along with a number of others, I expressed my scepticism about the firmness of the foundations on which this remarkable Conservative and Labour recovery was built.

It only took the 2019 European elections to reveal that 2017 was just so much Scotch mist. For an election that should not have been held and which, when it was, was only supposed to elect MEPs to serve for a few months (if at all), it was politically devastating. The Brexit Party that had only been formed less than two months before polling day took almost one-third of Britain’s votes, along with 29 of its 70 seats.

The Conservatives and Labour could not muster even a quarter of the votes between them. The Tories came fifth with their lowest level of support in any national election since 1832; and Labour came third with its lowest vote share since 1910 (when it contested just 56 out of 569 British constituencies).

The real significance of the European parliament election was the speed with which its explosive outcome transferred into subsequent Westminster voting intention polls. As a result, the October 2019 polls sampled before the 12 December election date was settled showed the Conservatives averaging 36 per cent (down seven points on the 2017 election), Labour on 25 per cent (down 16 points), the Lib Dems on 18 per cent (up 10 points), the Brexit Party on 11 per cent (up nine points on Ukip’s 2017 share) and the Greens and other parties up two points each.

None of these latest poll figures is set in concrete. But I think it would take great courage to claim that they will return during the course of this election campaign to the 84 per cent share of the total vote that the Conservatives and Labour jointly achieved in 2017.

The scale of this astonishing and unprecedented electoral volatility was recently analysed by the British Election Study (BES) team. The BES has captured voting behaviour at every general election since 1964, through a series of substantial surveys of voters. The study is currently supervised by academics from Manchester University and Nuffield College, Oxford. Last month they presented some remarkable data that will feature in Electoral Shocks: The Volatile Voter in a Turbulent World (Fieldhouse et al, Oxford University Press, December 2019).

In 2015, some 43 per cent of voters supported a different party from in 2010; and in 2017, 33 per cent changed their vote from 2015. Overall, across all three of the last elections between 2010 and 2017, half of people (49 per cent) did not vote for the same party each time.

In 2017 there was the largest switching of votes from Conservative to Labour and vice versa ever recorded by the BES. The authors estimate that fully 11 per cent of voters switched between the two main parties. Even in the immense Labour landslide of 1997, when a record 28 per cent of all seats changed hands, the comparable figure was 9 per cent.

Underpinning these developments is another set of BES data that tracks the collapse in strength of party identification over past decades. In 1964, 48 per cent of Conservatives described their party identification as “very strong”; the figure among Labour supporters was 51 per cent. In 2017, the comparable figures were 14 per cent and 23 per cent respectively.

In the face of all this we still demand predictions for the outcome of the 12 December election. I suppose that is inevitable; but it really would take a heart of stone not to have sympathy for those trying to find traces of order in such electoral chaos.

David Cowling is a political analyst and senior visiting fellow at King’s College London. He will be commenting on the election for ‘The Independent’ throughout the campaign