Inevitably, a poll that suggests the Conservatives could win fewer seats at the next election than the SNP grabs attention. As it should.
John Curtice www.independent.co.uk
According to a large 28,000-person poll conducted at the turn of the month by FindOutNow and statistically modelled by Electoral Calculus for The Telegraph, the Conservatives would win just 45 seats in an immediate general election, while the SNP would have 50.
However you look at them, these kinds of poll numbers will produce heavy seat losses for the Tories – if these poll numbers transpired at an election, the losses are bound to be greater than the conventional approach to estimating seat outcomes would suggest.
What is uncertain is how much greater the losses would be… not least because we have never had a party’s support drop by as much as 22 points at an election. However, we should not discount the possibility that they could be quite a bit more, even if not necessarily on the scale suggested by Electoral Calculus.
How could this be so? There is, after all, only a small difference between the estimate of national party support in this poll and the average of other polls conducted at the same time. The Conservatives are credited with 23 per cent of the vote, only a couple of points below other polls. At 48 per cent, the estimate for Labour is exactly in line with that for other polls.
Turning estimates of national support into anticipated outcomes in seats is far from straightforward. Under our first-past-the-post system, how many seats a party wins depends not just on its level support, but also on how that support is distributed across the country.
The conventional approach is to assume the geographical distribution of each party’s support will remain the same as at the last election – that is, each party’s support will go up or down in every constituency in line with the change nationally.
If we analyse the Electoral Calculus/FindOutNow poll in that way, the Conservatives, down 22 points on their 2019 tally, would be credited with 142 seats (still an all-time record low). Meanwhile, Labour, up 15 points on 2019, would have 406 seats, just a little below the record tally of 418 seats Tony Blair won in 1997, and well short of the 509 projected by Electoral Calculus.
However, at present there is a problem with this approach. There are 90 seats in which the Conservatives won less than 22 per cent of the vote in 2019 and thus where it is arithmetically impossible for the party’s support to fall by 22 points. Consequently, the party’s support must be down by more than 22 points in some places where it performed better last time. That is likely to include constituencies the party is trying to defend, resulting in a greater loss of seats than the conventional calculation would anticipate.
But where and by how much would Conservative support fall by more than 22 points? According to Electoral Calculus’s modelling, the answer is clear – the higher the Conservative vote last time, the more the party’s support would fall. In true blue Castle Point, the party’s support is projected to have fallen from 77 per cent to 42 per cent, a fall of 35 points – in sharp contrast to the picture in Liverpool Riverside, where it is estimated to have slipped from 7 per cent to 4 per cent.
On average, Electoral Calculus estimate the party’s vote in its 120 strongest seats is now as much as 30 points down on 2019 – and by 27 points in the 120 next strongest. It is those figures that explain why the poll’s estimated outcome of the Tory tally of seats is nearly 100 less than in the conventional calculation.
But is this expectation of a much sharper fall in Conservative support in the party’s strongest seats correct? After all, when the party last crashed to a serious defeat in 1997, its share of the vote was only down two points more in its safest seats than across the country as a whole.
However, at 11 points, the national drop in support for the party on the previous election in 1992 was half the fall in the Electoral Calculus poll – so there were far fewer seats in which it was mathematically impossible to match the fall across the country as a whole.
Closer to the Tories’ current position is the predicament of the Liberal Democrats in 2015, when their support fell across the country by 15 points. That translated into as much as a 20 point drop in their strongest seats. However, as compared with the position of the Conservatives now, there were nearly twice as many seats in which it was arithmetically impossible for the drop in Liberal Democrat national support to be replicated locally.
In truth, we are in unchartered waters. Given the unprecedented scale of the fall in Tory support, nobody can be sure what the outcome in seats would be if the current polls were reflected in the ballot boxes.
Maybe the party would lose rather less support in its strongest seats than Electoral Calculus anticipate. But, equally, perhaps voters would vote tactically for whichever of Labour or the Liberal Democrats could best defeat the Conservatives locally, something mid-term polls rarely pick up but which, as in 1997, could cost the Conservatives dear.
One thing though is certain. Unless the Conservatives can haul themselves out of the electoral doldrums, the party will be a much diminished parliamentary force after the next election.
John Curtice is a professor of politics at Strathclyde University, and senior research fellow for NatCen Social Research and ‘The UK in a Changing Europe’