BMG Research has undertaken exclusive polling for The i newspaper. The polling covers topics including vote intention, benefits and defence.
Analysis by Robert Struthers, Head of Polling at BMG
Keir Starmer may be facing his most vulnerable moment as Prime Minister, but his polling, while hardly rosy, is showing some signs of improvement. Labour has recovered to the dizzy heights of 27%. That might not sound great, but it is five points higher than immediately after the local elections and Labour now trailing Reform by 3 points rather than 10.
We are likely still years away from a general election, but these figures show Reform falling from being able to form a commanding majority shortly after the locals, to likely just short of one now: showing Labour still has some fight left in it.
Starmer’s personal ratings remain stable, and he is still very unpopular: almost as unpopular as Rishi Sunak was before his heavy defeat. But again, compared to shortly after the local elections, they show signs of moderate improvement.
Where has this recovery come from? A couple of months ago, just 6 in 10 of those who voted Labour 12 months after the election said they would back the party again. This is now up to 7 in 10 – a modest recovery. Fewer of his 2024 coalition indicate they would switch to the Lib Dems, Reform, or Greens. These are small shifts but are coming from all direction directions.
This could hint that dropping the unpopular winter fuel allowance policy has helped, but it’s worth stressing Labour is still losing 3 in 10 of its 2024 voters. And this is off the back of an election victory with an already low vote share that only translated to a majority because of the quirks of the electoral system. So, while Labour may look steadier this month, the ground beneath them is still far from firm.
An article by The i on vote intention can be found here.
An article by The i on benefits can be found here.
An article by The i on foreign policy 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.