Polling 24/09/2026:

On the hub’s own measure, the Conservatives were the largest party in the ballot box on 24 September. Across five council by-elections they took 26.3% of the votes cast.

Labour was second on 22.8%. Reform UK was third on 19.2%.

The Greens were on 16.4%, the Liberal Democrats on 13.2%, and independents on 2.2%. The SNP, Plaid Cymru, Restore Britain and Your Party did not register a vote in this set of wards.

That is a local snapshot. It is not a forecast of a general election. The same day, the big national polls still had Labour narrowly ahead, and short of a majority.

What the hub counted

The sheet is titled “Voting Intention based on BY ELECTIONS”. The 24 September row is built from five contests and about 8,700 votes. Shares are the raw total, not an average of the five wards and not weighted to look like Britain.

Seats told a different story from the vote pile. The Conservatives gained one and held two. Labour held one and lost one. The Greens gained one and lost one.

• Queens Park, Brighton & Hove. Green gain from Labour. Green 1,139 (50.1%), Labour 802 (35.2%), Reform UK 160 (7.0%), Conservative 97 (4.3%), Liberal Democrat 74 (3.3%).

• Blockley, Cotswold. Conservative gain from Green. Conservative 431 (54.0%), Liberal Democrat 230 (30.1%), Reform UK 59 (7.4%), Green 54 (6.8%), Labour 14 (1.8%).

• New Bradwell, Milton Keynes. Labour hold. Labour 1,023 (54.2%), Reform UK 441 (23.4%), Green 175 (9.3%), Conservative 160 (8.5%), Liberal Democrat 87 (4.6%).

• Wombourne North, South Staffordshire. Conservative hold. Conservative 864 (50.1%), Reform UK 645 (37.4%), Labour 115 (6.8%), Green 68 (3.9%), Liberal Democrat 33 (1.9%).

• Bray, Windsor & Maidenhead. Conservative hold. Conservative 745 (36.1%), Liberal Democrat 719 (34.9%), Reform UK 368 (17.8%), Borough First independents 193 (9.4%), Labour 38 (1.8%).

  
  

The night before, in Cranbrook, Sissinghurst & Frittenden on Tunbridge Wells council, the Liberal Democrats gained the seat from the Tunbridge Wells Alliance. That contest is not in the 24 September pool.

Why these shares swing

A by-election pool is only as typical as the wards on the ballot. Queens Park is a Green-Labour fight in Brighton. Blockley and Wombourne North are Conservative ground. Bray was a near-tie between the Conservatives and the Liberal Democrats. New Bradwell was a clear Labour hold, with Reform second.

Put those five together and the Greens look strong, because one safe-looking Green ward casts a lot of votes. Reform looks weaker than in national polls, because it was a poor third in Brighton and a minor candidate in Blockley, even while taking 37.4% in Wombourne North.

The hub’s own sheet shows how fast this moves. The next logged week, 1 October, six by-elections produced Labour 35.2%, Reform UK 30.6%, Conservative 22.8%, Restore Britain 7.1%. Same method. Very different map.

The national picture on the same day

National voting intention that week was a three-way scrap, with Labour usually first.

The Independent’s 24 September round-up put the average at Labour 26%, Reform UK 23%, Conservatives 20%. On those numbers, Andy Burnham would be the largest party in a hung parliament, not a prime minister with a majority of his own.

Two MRP polls, which estimate each seat rather than just the national vote, were the reason commentators that day played down an early election.

Electoral Calculus, for Restore Britain, published on 24 September. Find Out Now polled 5,549 adults in Great Britain from 4 to 11 September. With its tactical-voting model, the shares were Labour 25%, Reform UK 22%, Conservative 20%, Restore Britain 7%, Liberal Democrat 10%, Green 9%.

The seat estimate was Labour 288, Conservative 165, Reform UK 84, Liberal Democrat 61, SNP 18, Green 7, Plaid Cymru 7, Restore Britain 1 (Great Yarmouth). Labour would be 38 seats short of a majority. Labour plus the Liberal Democrats would have a majority of 48.

Without that tactical-voting assumption, the same poll had Reform UK on 127 seats and Labour on 261. The model’s own finding was that strong anti-Reform tactical voting could cost Reform dozens of seats. More than a third of Labour voters said they would vote Conservative to keep Reform out.

YouGov’s larger MRP, fieldwork 23 August to 18 September, sample about 17,000, put Reform UK third on seats: 123, behind the Conservatives on 130. On that map, Labour and the Liberal Democrats together would still be short, so a Burnham government would also need the Greens, the SNP or Plaid Cymru.

Survation’s MRP, 20 August to 4 September, sample about 8,000, put Reform second, which would make Nigel Farage leader of the opposition. It had Labour and the Liberal Democrats combined on a majority of about 40.

Ordinary polls in the days around 24 September sat in the same band. Opinium for the Observer, 23–25 September, had Labour 28%, Reform UK 24%, Conservative 17%. Survation, 22–23 September, had Labour 27%, Reform UK 24%, Conservative 21%. Find Out Now on 23 September had Labour 24%, Reform UK 23%, Conservative 19%, Green 13%. BMG for the i Paper, 23–24 September, had Labour 28%, Reform UK 23%, Conservative 20%.

YouGov for the Times and Sky News was the outlier in the weekly trackers. On 20–21 September it had Labour 23%, Reform UK 21%, Conservative 21%. A week later, 27–28 September, it had Reform UK 25% and Labour 24%.

What it meant for an early election

Taken together, 24 September did not show a government that could call an election and expect to win it outright.

The by-election pool flattered the Conservatives and the Greens, and understated Reform relative to the national polls. The national polls flattered no one with a majority. Labour’s lead over Reform was real but small, about three points in that day’s average. Under first past the post, a lead of that size, split across five parties in England, does not produce 326 seats.

Burnham, prime minister since 20 July, was being told by those MRPs that he would need at least one partner. On the Electoral Calculus map, the Liberal Democrats would do.

On the YouGov map, he would need a wider deal. The Independent’s conclusion that day was blunt: even if tactical voting broke his way, he would be asking for trouble, and Labour would still be on course to lose more than 100 seats and its majority.

The hub figure is the cleaner local read. The Conservatives won the day’s votes, and the day’s seats, in a set of wards that suited them. It is not evidence that Kemi Badenoch had overtaken Labour in the country. For that, the national polls of the same week still had Labour first, Reform second, and nobody over the line.