The FT’s UK general election model explained

The Financial Times’ UK general election model shows how headline polling numbers translate into seats in the House of Commons.

To do this, the model seeks to work out what the current polls imply about the 632 constituencies in Great Britain and how they will vote on July 4. The headline projection will be updated every day as more national and local polling is released.

Underpinning the model is an attempt to track which voters have changed their minds since 2019, and which have not.

The FT has used data from the British Election Study to observe how national changes in polling led to local changes in political support between 2019 and 2024.

You are seeing a snapshot of an interactive graphic. This is most likely due to being offline or JavaScript being disabled in your browser.

By default, the national vote shares on the model are set to the FT’s polling average. The current central projection sees Labour winning a decisive victory, with around 450 seats to the Conservatives’ 100, a landslide majority of approximately 250 seats.

A particular complexity for modelling is tactical voting. This model does not explicitly try to guess at how tactical voters will split, which could be a significant problem for Prime Minister Rishi Sunak’s Conservatives.

As election day approaches, new data including constituency-level polls will allow future versions of the model to incorporate local information that should pick up potential tactical voting.

You are seeing a snapshot of an interactive graphic. This is most likely due to being offline or JavaScript being disabled in your browser.

At the moment, the Liberal Democrats, the biggest likely beneficiaries of tactical voting, are projected to be the third-largest party with around 60 seats.

The SNP, another party exposed to tactical voters, is fourth on around 15 seats. The Greens are projected to hold on to their one seat, and Plaid Cymru will hold on to two seats in Wales.

What is striking is that in many constituencies, Labour’s projected vote share is scarcely above where it was in 2017 under Jeremy Corbyn, when it won 40 per cent of the vote nationwide. Yet thanks to the collapse in Tory support, hundreds of seats are projected to change hands.

Our model gives similar projections to other sophisticated models available, such as MRPs, a type of survey analysis that combines large polls with demographic data to estimate local-level outcomes.

You are seeing a snapshot of an interactive graphic. This is most likely due to being offline or JavaScript being disabled in your browser.

If you disbelieve the polls, or think they will narrow as the election approaches, the model’s assumptions can be changed.

Were the Conservatives to perform 5 percentage points better than their current polling, our model projects them to hold around 200 seats. A further 7 points would put them on 280 seats and potentially deny Labour an outright majority. 

Shifting them in the other direction, the electoral abyss looms. Were the Conservatives to fall 5 points to 16 per cent, our model would give them just 18 seats. Such a result is not unheard of in a first-past-the-post system; in Canada in 1993, the ruling Progressive Conservative party collapsed to 16 per cent of the national vote and mustered just two seats.