BMG Research has undertaken exclusive polling for The i newspaper. The polling covers topics including vote intention, leader satisfaction, immigration and the Autumn Budget. Key findings are:


  • The Labour party lead has increased to 17 points. A consistent 43% say they would vote for Labour, yet only 27% saying they would vote Conservative, a decline of 3 points since October.
  • Reform UK are recording their highest share of the vote at 11%, an increase of 4 points since our last polling in October. The highest share of the vote for Reform UK corresponds with the greatest proportion who consider immigration the most important issue facing the UK today (12%) in 2023 and three fifths who consider immigration too high (63% when presented with official ONS migration figures as context, 58% when asked without these figures).
  • The Conservatives have not achieved the polling bounce they were looking for in the Autumn Budget. Of those who have seen something about the Autumn Budget, only 17% say it would make them more likely to vote for the Conservatives. In comparison, 56% say the Autumn Budget would make no difference on how they vote and 24% say it would make them less likely to vote Conservative.
  • Both Labour and the Conservatives seem more divided than they were during conference season, with 62% of the GB public saying the Conservatives are divided (+10 since October) and 43% saying the Labour Party are divided (+15).
  • For the Conservatives, the view that they are more divided likely stems from the fallout of the removal of Suella Braverman as Home Secretary and their failings on immigration (see above). Yet for Labour, this division is likely a result of the resignations over the vote for a ceasefire in Israel-Gaza.  In line with this, voters are more likely to think that Sunak has responded well to the Israel-Gaza conflict than Starmer (33% c.f. 24%).
  • Sunak’s net satisfaction is at the lowest we have recorded. 21% are satisfied with the job Rishi Sunak is doing as Prime Minister (cf. a high of 31% in November / early December 2022), and 54% are dissatisfied (cf. 34% in November / early December). The Prime Minister’s net satisfaction is now more in line with Boris Johnson at the end of his premiership in August 2022 (-35).
  • Meanwhile, Keir Starmer’s net satisfaction has declined slightly (29% satisfied and 35% dissatisfied), with those dissatisfied with Starmer at the highest level we have recorded.

On the Autumn Budget, Immigration and Voting Intention:

“Rishi Sunak will have been hoping for a polling boost following the Autumn Statement last week. But if anything, the reverse has happened and his electoral prospects look even bleaker.

If our numbers were repeated in a general election, we’d be looking at a 1997-style landslide victory for Starmer, with the Conservatives dropping well below 200 seats—a remarkable reversal since 2019.

Labour’s lead is particularly notable given our methodology accounts for some undecided former Conservative voters likely returning to the fold. This makes Labour’s position looks all the stronger.

The voting intention picture is indicative of the Conservative’s weakness on issues where they have historically been stronger. Despite the Chancellor’s Autumn Statement, Labour has extended its lead on economic issues this month.

More strikingly, there’s a noticeable revolt brewing on the right. For the first time, the Reform Party is hitting double digits in our polls, a significant jump from their low single-digit presence a year ago.

This will have been fuelled by Suella Braverman’s resignation and the latest net migration figures, with Labour’s lead on the issue of immigration also increasing.

All these factors combine to leave Rishi Sunak in a precarious position, steering through a turbulent political landscape as public confidence in him and his party continues to wane.

An article by The i on Labour Party Conference and voting intention can be found here.

Methodology, fieldwork dates, and a full breakdown of these results can be found here.

As a sample of the population was interviewed, the results are subject to a margin of error around various estimates. This means that, given the random nature of the sampling process, we can be confident that the actual result lies somewhere within the margin of error. For example, with a sample of 1,000 we can be 95% certain that the actual value will fall 3% either side of the result. For a more detailed breakdown of results from this poll, or any other results from our polling series, please get in touch by email or phone.

Share this article: