The Economist’s final prediction points to a Tory wipeout in Britain

With the cut-off date for postal-voting applications already past, the Tories are polling at historically awful levels.

It is not simply that other parties — Labour, the Liberal Democrats and Reform UK — are taking chunks out of their support. It is the way they are doing it.

A new mega-poll conducted by The Economist in partnership with WeThink, a research outfit, suggests that they are hacking away at even the safest Tory seats.

Between May 30th and June 21st, WeThink asked 18,595 adults how they intend to vote in the general election. The results suggest Labour has a 20-percentage-point lead over the Conservatives, by 42 per cent to 22 per cent.

Reform UK is on course for 14 per cent of the vote, the Liberal Democrats 11 per cent and the Green Party 6 per cent.

This is a dramatic turnaround from the 2019 election, when the Tories led by 12 points, and would be the largest swing between the main parties in modern history.

But even that does not tell the full story.

Because of the mega-poll’s large sample size, The Economist is able to use it to analyse local and demographic trends in voting intention.

Using a statistical technique called multilevel regression and poststratification (MRP), we have produced estimates of voting intentions for each constituency in Britain.

These MRP estimates suggest that the Labour Party is on course to win 465 seats, 263 more than it won at the 2019 election.

The Tories are heading for just 76 seats, the lowest number in the party’s history.

The Liberal Democrats are estimated to win 52, their best result since 2010; the Scottish National Party is on track to lose 19 seats in Scotland, though it would remain the largest party there.

Reform UK and the Green Party are on course to take three seats each.

Close readers will notice that these tallies differ from the predictions of our election model.

During the same period that the mega-poll was being conducted, that model was giving a central estimate of around 185 mps for the Tories, a very heavy defeat but not a rout.

The key difference lies in the “efficiency” of each party’s vote—how many seats they can expect to win for a given number of votes.

Our prediction model, which is based on national and regional polls, assumes the distribution of votes for each party across seats would be similar from election to election.

Our MRP poll shows that the Tories are underperforming in marginal constituencies.

This is because opposition parties and voters are highly co-ordinated.

Labour has explicitly targeted “hero voters” who supported the Conservatives in 2019.

These are people who are older, working-class, Leave-voting and overrepresented in marginal constituencies in the north of England.

The Liberal Democrats may have rather similar policies but they have targeted very different voters—wealthier, more educated and more likely to have voted Remain and to live in the south.

The MRP poll shows that Labour’s vote share has risen by 11 percentage points in competitive seats where they face the Conservatives, compared with eight points in constituencies where the Lib Dems are the Tories’ main rivals.

The Liberal Democrats, meanwhile, have gone backwards in most of the country compared with 2019 but, crucially, have increased their vote share by three percentage points in seats where they are competitive with the Conservatives (see chart 1).

This dynamic on its own is enough to make the Tories susceptible to substantial losses.

But the installation of Nigel Farage as leader of Reform UK on June 3rd has made even Conservative MPs with huge majorities vulnerable.

Reform UK draws its support overwhelmingly from the Tories.

The loss of white, non-graduate Leavers seems to be causing the Tory vote share to tumble in former strongholds: our MRP poll estimates that it will fall by over 30 percentage points in seats where they won more than 60 per cent of the vote in 2019.

Reform gains the votes; Labour reaps the rewards.

Many MRP polls have been published since Mr Farage became leader of Reform UK; all agree that the Conservatives are heading for a landslide defeat.

But the range of outcomes is large: four of these recent mega-polls have shown Tory seat totals of between 53 and 155.

One reason for that great variation is the knife edge that the Conservatives sit on.

In our MRP estimates, we find that they win or lose by a margin of less than 5 per cent in 91 constituencies.

In these extreme electoral circumstances, the difference between a heavy defeat and a total rout is a very small number of votes.

According to our own MRP, if Reform UK were to take just three additional percentage points from the Conservatives, the Tories’ seat total would fall to only 45.

MRPs are not foolproof.

Like all polling, an MRP poll is a snapshot and has a substantial margin of error due to statistical variation.

Like all modelling, the technique also relies on a number of subjective assumptions — which variables should be included, how turnout is estimated, and so on.

These assumptions vary between pollsters, and it is difficult to know which set will produce the most accurate estimates.

The best approach, therefore, is to pool information from across a full range of forecasts. We have updated our prediction model accordingly, to incorporate the results of all the published MRPs, our own included.

This “blended” election model, which will be updated between now and July 4th, is our best guess at the final result.

At the time of writing its central estimate is that Labour will have 429 MPs, the Conservatives 117, the Liberal Democrats 42 seats, the SNP 23 and Reform UK two.

Pollsters can get things wrong; we will do our own post-mortem after the election.

But with only days to go, for this sort of result even to be a possibility is remarkable.