Labour open up 33-point vote intention lead in wake of mini-Budget chaos
The latest YouGov/Times vote intention poll shows the Labour party on 54% of the vote, up nine points on their previous record high with YouGov on Monday. The Conservatives meanwhile have dropped to 21% of current vote intention, down seven points.
As well as being their record highest share ever in a YouGov poll, Labour’s 33-point lead is the highest figure the party has ever recorded in any published poll since the late 1990s.
Elsewhere, the Liberal Democrats have dropped down to 7% (-2) of vote intention, the Greens are on 6% (-1), and Reform UK are on 4% (+1).
There are three key factors in Labour's lead. First, the proportion of 2019 Conservative voters who now say they will vote Labour has risen to a sizeable 17%, doubling from 8% in our poll published on Monday. Only 37% of those who backed the Conservatives in the previous election currently intend to stick with the party.
In historical perspective, this 17% ‘Con to Lab’ switching figure far eclipses the 5% rate of switching we saw when, under Ed Miliband, Labour were consistently the Tories by more than 10 points in 2012. The current situation is more comparable to the around 16% of 1992 Conservative voters who switched to Labour in 1997 under Tony Blair, according to the British Election Study.
The second important factor in Labour's lead is that a further 26% of 2019 Conservative voters say they do not know how they would vote if an election were held tomorrow, and 9% say they would not vote at all. These figures have not changed significantly since Monday, but are higher than in the spring when the recorded ‘Con to don’t know’ figure was between 15-20% and Labour’s lead was between zero and five points.
The third important story in the growing Labour lead is the squeeze on support for the Liberal Democrats. While in July YouGov polling had Ed Davey’s party at 15% of vote intention, they have now slipped to just 7% in our latest poll. Fully half (50%) of 2019 Liberal Democrat voters now tell us they intend to vote Labour.
The direct transfer of voters away from both the Conservatives and the Liberal Democrats is the key reason for Labour’s record lead today. The context is of course very important – we are in the middle of a very negative news cycle for the government as they deal with the economic fallout from Friday’s deeply unpopular mini-budget, at the same time as Labour are enjoying increased coverage and exposure – including of popular policies such as the establishment of a public energy company – from their conference. Add to this long-term public frustration with the government’s handling of the cost of living and decreasing faith in the Conservative Party to handle the economy, and it all makes for a toxic public opinion environment for the government.
YouGov closely monitors underlying trends and patterns in responses to all its surveys, including for potential response bias and other sampling issues. The panel-wide response rate for the fieldwork dates was not significantly different from our September average, and the weight efficiency for this survey was 87%, with a weight standard deviation of 0.4 (indicating no major problems with reaching particular response groups).
In recent years, there have been good questions asked as to whether polls accurately sample voters at times of particular difficulty for a given party – is it the case that their voters will simply drop out of taking polls for a while during periods of poor electoral results or negative media environments?
As well as controlling for known vote intention correlates such as education and political interest, YouGov quota all our political polling by past vote to ensure that we have enough voters from each party in our samples at all times.
Further, while it is impossible to measure the intentions of people who are not in polls, we can use other methods to monitor the type of voters (and supporters) coming into polls at any given time and check whether this correlates with news and events cycles. Our tests today indicated that there was no significant difference in the supporter profiles of voters coming into this particular sample versus both our panel population averages nor the long-run average of panellists coming into political polls.