Rachel Reeves and Angela Rayner could both lose their seats, the poll suggests (Image: PA)

Ed Miliband is also vulnerable (Image: Getty)

Reform would unseat 12 cabinet ministers, the Greens three and an independent one. Among the high-profile Labour losers are eight senior figures who won their seats in the July 2024 general election: Chancellor Rachel Reeves (Leeds West and Pudsey), Foreign Secretary Yvette Cooper (Pontefract, Castleford and Knottingley), Energy Secretary Ed Miliband (Doncaster North) and Defence Secretary John Healey (Rawmarsh and Conisbrough), all defeated by Reform; Chief Secretary to the Treasury Darren Jones (Bristol North West), Business Secretary Peter Kyle (Hove and Portslade) and Northern Ireland Secretary Hilary Benn (Leeds South), defeated by the Greens; and Health Secretary Wes Streeting (Ilford North), losing to an independent. Former deputy prime minister Angela Rayner (Ashton-under-Lyne) is also projected to lose to Reform.

Prime Minister Sir Keir Starmer would hold Holborn and St Pancras comfortably, though with a significantly reduced majority, reports The Sunday Times. Deputy Prime Minister David Lammy clings on in Tottenham by just 1 %.

Tory leader Kemi Badenoch could also be at risk (Image: Getty)

Luke Tryl of More in Common said Reform remains “well over 200 seats clear of their closest challenger and the only party that could viably form a government”, despite a slight plateau since January (down 57 seats). Labour gains 16 seats on the earlier poll but still faces a historic collapse from 411 in 2024.

The Green surge marks the biggest change, threatening Labour in its London strongholds and among younger voters. Possible leadership contenders such as Mr Streeting and Angela Rayner are among those projected to fall, leaving few obvious successors to Sir Keir.

With Reform on 324 seats, it is the clear frontrunner but short of the 326 needed for outright control.

The survey leaves Sir Keir’s government under siege from both Reform nationally and the Greens on the left, especially in the capital.

Labour Left With Just 102 MPs in Bombshell New Poll – Reeves, Rayner and Miliband Gone

A dramatic polling projection once suggested that Labour could be reduced to little more than 100 MPs at a future general election, with several senior figures facing the possibility of losing their parliamentary seats. The figure of 102 MPs has since been used in political commentary to illustrate the scale of Labour’s potential losses, although it is important to distinguish an MRP projection from an actual election result.

The original More in Common model published in April 2026 projected Labour winning just 101 seats if a general election were held under the conditions measured by the poll. Labour would have lost 310 seats compared with its landslide result in the 2024 general election. Reform UK was projected to win 324 seats, while the Conservatives were projected to win 81 and the Greens 22.

The projection attracted particular attention because it suggested that Labour’s parliamentary strength could be dramatically reduced within a relatively short period. It also identified a number of prominent Labour politicians whose constituencies could become vulnerable.

However, the figures should not be interpreted as a prediction of a confirmed election result. An MRP model estimates constituency-level outcomes using polling data and demographic information. It can change substantially as national support changes, candidates are selected and local campaigns develop.

A dramatic reversal from 2024

The scale of the projected Labour decline is particularly striking because the party won a large parliamentary majority at the July 2024 general election.

The April More in Common model estimated that Labour could fall from 411 seats won in 2024 to just 101. That represented a projected loss of 310 MPs.

At the same time, Reform UK was projected to move from a small parliamentary presence to 324 MPs, just two seats short of the 326 normally required for a Commons majority. The Conservatives were projected to win 81 seats, while the Liberal Democrats were projected to take 62 and the Greens 22.

The model therefore depicted a highly fragmented political landscape in which Labour would no longer dominate the parliamentary map.

One of the most significant features was the pressure coming from several directions. Labour could lose seats not only to Reform UK but also to the Conservatives, Liberal Democrats, Greens, SNP and independent candidates.

That is important because Labour’s difficulties cannot simply be described as a single-party challenge. The model indicated that different parts of the country could experience very different political contests.

Senior Labour figures under pressure

The headline claim that Rachel Reeves, Angela Rayner and Ed Miliband could all be “gone” reflects the wider theme of senior Labour figures being vulnerable in constituency-level projections.

Reeves, Rayner and Miliband have all occupied important positions in Labour’s political story, although their parliamentary circumstances and political roles have not been identical.

Rachel Reeves became one of the central figures in Labour’s economic agenda after the party entered government in 2024. Her position as Chancellor made her especially important to the government’s argument that Labour could combine public investment with fiscal discipline.

Angela Rayner, meanwhile, became one of Labour’s most recognisable political figures and served as Deputy Prime Minister under Keir Starmer. Her future was also discussed during the political upheaval surrounding Labour’s leadership in 2026. ITV reported in May that Rayner had been cleared by HMRC of tax avoidance and was considering her political future, while Ed Miliband was among the figures mentioned in discussions about a possible Labour leadership contest.

Miliband has also remained a prominent Labour figure, serving as Energy Security and Net Zero Secretary. His name was mentioned as a possible leadership contender during the period of uncertainty surrounding Starmer, although people close to him denied that he was preparing a leadership campaign.

The political significance of these figures means that constituency-level threats involving senior politicians can generate much more attention than ordinary marginal-seat changes.

The model changed later

One of the most important qualifications to the “102 MPs” headline is that More in Common subsequently published another MRP model in July.

That later model projected Labour on 110 seats rather than 101. It was still an exceptionally large decline from the 2024 election result, but it demonstrated how quickly projections can move.

More in Common’s July model was based on fieldwork conducted between June 5 and June 28. It projected Reform UK on 293 seats, the Conservatives on 96, Liberal Democrats on 66, SNP on 37 and the Greens on 10. Labour’s projected total was 110.

This later result matters because it shows why individual MRP figures should not be treated as fixed forecasts.

The Labour projection improved by nine seats compared with the April model. Reform’s projection fell by 31 seats. The Conservatives increased by 15, while the Liberal Democrats gained four.

In other words, the political map was already changing between successive models.

Labour’s losses were geographically uneven

The July More in Common analysis offered a more detailed picture of what a heavily reduced Labour parliamentary presence could look like.

Almost half of Labour’s projected 110 seats were in London, where the party was expected to retain 48 constituencies. A further 23 were projected to be in the North West and 12 in Yorkshire and the Humber.

The picture elsewhere was much weaker.

The model projected Labour winning only one seat in the South West, five in Scotland and none in Wales. More in Common described this as the first time one of its models had projected a complete Labour wipeout in Wales.

The geographical distribution is significant because Labour’s traditional electoral coalition has historically depended on support across different regions of Britain.

A model showing Labour concentrated increasingly in major urban centres would therefore represent a substantial change in the party’s parliamentary position.

The Green challenge

Another important element of the earlier projection was the rise of the Green Party.

More in Common’s April model projected the Greens to increase their parliamentary representation from five MPs to 22. The party was identified as a significant challenger in several Labour-held urban constituencies.

The Times also reported that three Labour cabinet ministers were projected to lose their seats to the Greens in that model: Darren Jones, Peter Kyle and Hilary Benn. The wider projection suggested that 16 cabinet ministers could lose their seats to Reform, the Greens or independent candidates.

However, the later July model showed the Green projection falling from 22 to 10 seats.

That change illustrates the volatility of the polling environment. A party can appear capable of making major gains in one model and then see its projected seat total fall considerably in a subsequent exercise without an election having taken place.

What about Reform UK?

The original 101-seat Labour projection was closely connected to a particularly strong Reform UK model.

In April, More in Common projected Reform on 324 seats. By July, that projection had fallen to 293.

The July figure nevertheless remained substantially higher than Reform’s 2024 result, but the model also showed the Conservatives gaining ground and Reform losing some of its projected support. More in Common said Reform remained the largest party in its model but was further from a Commons majority than in April.

This demonstrates another important point: the Labour headline cannot be considered in isolation.

A projected Labour decline depends partly on where Labour’s lost support goes. If votes move to Reform in one constituency, the Conservatives in another and the Greens somewhere else, the parliamentary consequences can be very different.

MRP is not an election result

MRP, or Multilevel Regression and Poststratification, is designed to estimate results across individual constituencies using polling and demographic information.

It can be useful for understanding possible electoral patterns, particularly when ordinary national polling cannot show how votes might translate into seats.

But it remains a model.

The April More in Common analysis itself described its figures as projections based on the political conditions measured at the time. The July model also included a methodological warning: it did not incorporate some of the most recent by-election results.

That means constituency boundaries, candidate effects, campaign developments, turnout and changes in national opinion can all alter the eventual outcome.

Consequently, the claim that Labour “will” be left with 102 or 101 MPs would go beyond what the polling evidence establishes.

A warning rather than a final verdict

The significance of the projection is therefore less about one exact number and more about the scale of change it illustrated.

Labour entered government after the 2024 election with 411 MPs. An MRP model published less than two years later placed the party at just 101 seats. A later model moved that figure to 110.

Even though those figures are not election results, they demonstrate how dramatically Britain’s electoral landscape was being modelled during 2026.

For Labour, the challenge was not confined to one rival. Reform was making gains in many constituencies, the Conservatives were recovering in some areas, the Greens were challenging Labour in parts of urban Britain, while the SNP and Plaid Cymru remained important forces in Scotland and Wales.

The fate of individual senior politicians would ultimately depend on the constituencies they contested and the political environment at the time of an actual election.

The “102 MPs” headline therefore captures the dramatic nature of one polling scenario, but it should not be confused with a settled outcome. The subsequent movement in the More in Common models—from 101 Labour seats in April to 110 in July—shows precisely how much can change before voters actually cast their ballots.

What the polling does establish is that Labour’s 2024 dominance was not necessarily reflected in every subsequent electoral measure. The scale of the projected losses varied from model to model, but the underlying data showed a much more competitive and fragmented political landscape than the one produced by the 2024 general election.