Polling day · Thursday, 18 June 2026
Makerfield by-election
A 1.4pp toss-up between Reform UK and Labour. Well inside the model's confidence interval. Wigan borough · North West · E14001350.
Central forecast (probability-weighted projected vote share)
- Reform UK 41.1% Robert Kenyon
- Labour 39.8% Andy Burnham
- Conservative 7.0%
- Green Party 5.9%
- Lib Dem 2.9%
- Restore 1.8% Rebecca Shepherd
- Other 1.5%
Trigger
Why this seat is being contested
Josh Simons (Labour, MP for Makerfield since GE2024) announced his resignation on 14 May 2026 so that Greater Manchester mayor Andy Burnham could seek a return to Parliament. Labour's National Executive Committee cleared Burnham to stand on 15 May 2026. The Prime Minister has indicated Downing Street will not block the candidacy.
It is the first Westminster by-election since 1965 (Leyton, Patrick Gordon Walker) to be triggered specifically to seat a political figure from outside Parliament. Bookmakers and commentators frame the result as a precursor to a Labour leadership challenge to Sir Keir Starmer.
Swing from GE2024
How the model expects each party to move on 18 June
Coloured number is the swing vs the 4 July 2024 general election result. Blend = 85% Scenario A (Burnham stands) + 15% Scenario B (Burnham withdraws).
Two scenarios
The Burnham-on-the-ballot question
The headline forecast hinges almost entirely on whether Burnham appears on the ballot. Two independent forecasts were published on 15 May 2026: Survation and Britain Elects (Ben Walker). Neither is a fieldwork-based constituency poll. Survation transfers the Burnham effect from their Gorton & Denton post-by-election poll (n=501, 9 Mar 2026) re-weighted for Makerfield demographics, while Britain Elects applies a North West regional Burnham favourability adjustment to the GE2024 baseline. Both agree Labour leads by 3pp with Burnham as candidate; they disagree on absolute levels (Survation Lab 45 / Ref 42; Britain Elects Lab 39 / Ref 36).
Burnham stands
P = 85%- Labour 42.0%
- Reform UK 39.0%
- Conservative 7.3%
- Green Party 5.8%
- Lib Dem 2.9%
- Restore 1.5%
- Other 1.5%
Burnham withdraws
P = 15%- Reform UK 53.0%
- Labour 27.0%
- Green Party 6.0%
- Conservative 5.0%
- Restore 4.0%
- Lib Dem 3.0%
- Other 2.0%
Scenario A blends both forecasts: 50/50 mean of Survation (Lab 45 / Ref 42) and Britain Elects (Lab 39 / Ref 36) for the major parties, with Survation's minor-party splits scaled to the residual. Scenario B is Survation's non-Burnham scenario verbatim (Britain Elects didn't publish one).
Probability framing. Our model uses 85% as P[Burnham appears on the ballot], reflecting NEC clearance + Burnham's same-day acceptance + Downing Street's confirmation of no block. Note this is a different quantity to Survation's published 67%, which is P[Labour wins | Burnham stands] from their 10,000-simulation run.
1 May 2026 ward signal
What the council elections said about this seat
Reform UK won all nine Wigan wards inside the Makerfield Westminster boundary on 7 May 2026. Aggregated across 31,562 valid votes (turnout 37.3%), Reform led Labour by 25.9pp:
Survation's note references 8 wards / Reform 7-of-8. That's the pre-2023 Wigan boundary layout, before Hindley split into Hindley + Hindley Green. Our 9-ward aggregate uses the current (May 2023+) boundaries, which is what was actually voted on 7 May 2026.
| Ward | Turnout | Reform | Labour | Green | Con | LD |
|---|---|---|---|---|---|---|
| Abram | 33.1% | 56.1% | 24.2% | 11.3% | 4.5% | 4.0% |
| Ashton In Makerfield South | 38.5% | 46.4% | 32.5% | 12.7% | 8.4% | n/a |
| Bryn With Ashton In Makerfield North | 37.8% | 52.2% | 24.1% | 11.8% | 6.8% | 5.2% |
| Hindley | 35.3% | 52.3% | 21.4% | 10.7% | 4.0% | 3.6% |
| Hindley Green | 38.9% | 52.3% | 32.7% | 7.2% | 4.8% | 3.0% |
| Orrell | 43.0% | 39.9% | 24.2% | 11.6% | 19.4% | 4.9% |
| Pemberton | 31.1% | 54.2% | 24.5% | 12.0% | 4.9% | 4.4% |
| Winstanley | 42.4% | 49.7% | 31.0% | 8.2% | 6.4% | 4.6% |
| Worsley Mesnes | 37.4% | 50.9% | 3.0% | 9.5% | 25.2% | 3.2% |
| Aggregate | 37.3% | 50.2% | 24.3% | 10.5% | 9.5% | 3.7% |
Candidates
Who is contesting the seat
| Party | Candidate | Status |
|---|---|---|
| Labour | Andy Burnham | selected (subject to formal NEC endorsement): Greater Manchester mayor since 2017; cleared by NEC 15 May 2026. |
| Reform UK | Robert Kenyon | expected reselection (GE2024 runner-up; elected Wigan councillor 1 May 2026): Came second in Makerfield at GE2024 with 31.8%. |
| Conservative | to be confirmed | confirmed contesting: Kemi Badenoch confirmed Conservative candidate; ruled out non-aggression pact with Reform. |
| Green Party | to be confirmed | confirmed contesting: Green Party of England and Wales confirmed selection underway. |
| Lib Dem | to be confirmed | expected to contest |
| Restore | Rebecca Shepherd | selected: Restore Britain's first Westminster candidacy outside founder Rupert Lowe's Great Yarmouth seat. The party, which brands itself locally as 'Place Name First' (Great Yarmouth First in its home borough, founded Dec 2025, registered 4 Mar 2026), won all 9 Great Yarmouth divisions of Norfolk County Council and 1 of 39 Great Yarmouth Borough Council seats on 7 May 2026. Lowe himself holds the Westminster seat as Restore Britain having been elected for Reform UK in July 2024 and gone Independent before founding Restore Britain in November 2025. |
| Official Monster Raving Loony Party | Howling Laud Hope | expected |
Refreshed nightly from Democracy Club Candidates API. Final list confirmed at SoPN publication.
120-year historical anchor
What this seat has always done
Makerfield (and predecessor Ince) has returned a Labour MP at every general election since 1906, a 120-year unbroken streak. GE2024 produced the smallest Labour majority since 1931 (13.4pp); the 2026 Reform sweep of every constituent ward extends that compression. The seat voted 66% Leave at the 2016 referendum.
| Year | Winner | Lab share | Runner-up | Majority |
|---|---|---|---|---|
| 2024 | Josh Simons | 45.2% | Reform UK 31.8% | 13.4pp |
| 2019 | Yvonne Fovargue | 45.1% | Conservative 34.4% | 10.7pp |
| 2017 | Yvonne Fovargue | 60.1% | Conservative 31.3% | 28.8pp |
| 2015 | Yvonne Fovargue | 51.8% | UKIP 22.4% | 29.4pp |
| 2010 | Yvonne Fovargue | 47.3% | Conservative 18.8% | 28.5pp |
| 2005 | Ian McCartney | 56.6% | Conservative 19.3% | 37.3pp |
| 2001 | Ian McCartney | 62.4% | Conservative 20.7% | 41.7pp |
| 1997 | Ian McCartney | 65.8% | Conservative 18.8% | 47.0pp |
| 1992 | Ian McCartney | 61.6% | Conservative 27.3% | 34.3pp |
| 1987 | Ian McCartney | 52.2% | Conservative 29.7% | 22.5pp |
| 1983 | Roger Stott | 45.0% | Conservative 30.0% | 15.0pp |
Sources: Wikipedia, House of Commons Library general election briefings. Predecessor seat Ince held by Labour 1906-1983.
Per-ward indicators
Demographics, 2024 council baseline, and 2024 → 2026 Reform swing
The nine Makerfield wards are demographically homogeneous on ethnicity (all 90%+ White British, 0.6-1.8% Asian; these levers have no variance here and don't predict anything inside the seat). The wards diverge sharply on educational attainment, deprivation, retired-share and social-rent. The table below joins Census 2021 demographics to the 2024 council baseline and the 2026 result, then computes the Reform swing.
| Ward | IMD | % no-quals | % degree | % retired | % social rent | Reform 2024 | Reform 2026 | Swing |
|---|---|---|---|---|---|---|---|---|
| Abram | 2.9 | 24.7% | 20.3% | 18.5% | 17.3% | n/c | 56.1% | n/a |
| Ashton-in-Makerfield South | 6.3 | 18.6% | 26.4% | 26.3% | 7.1% | n/c | 46.4% | n/a |
| Bryn with Ashton-in-Makerfield North | 5.1 | 21.2% | 23.9% | 28.2% | 17.4% | n/c | 52.2% | n/a |
| Hindley | 3.3 | 25.5% | 20.8% | 22.5% | 23.4% | 13.2% | 52.3% | +39.1pp |
| Hindley Green | 5.8 | 20.9% | 23.5% | 24.9% | 9.2% | 15.5% | 52.3% | +36.9pp |
| Orrell | 7.6 | 15.8% | 33.4% | 28.6% | 8.2% | 13.8% | 39.9% | +26.1pp |
| Pemberton | 2.4 | 28.6% | 17.9% | 19.5% | 35.6% | n/c | 54.2% | n/a |
| Winstanley | 8.9 | 15.0% | 30.3% | 24.8% | 1.1% | 19.8% | 49.7% | +29.9pp |
| Worsley Mesnes | 4.5 | 25.0% | 20.6% | 24.9% | 21.1% | 23.6% | 50.9% | +27.4pp |
n/c = "no contest": Reform UK did not stand under their own party label in this ward at the 2024 council elections, so a clean 2024 → 2026 swing is undefined. In several of these wards the 2026 Reform candidate stood in 2024 as an Independent (notably David Bowker in Abram, 31.0% as Independent in 2024, 56.1% as Reform in 2026, so the de-facto swing is closer to +25.0pp than the headline-blank suggests). IMD = Indices of Multiple Deprivation 2019, average decile across LSOAs in the ward (1 = most deprived, 10 = least). Demographics from Census 2021 (TS021/030/054/066/067/004). 2024 turnout was ~23%; 2026 turnout ~37%, so the swing in wards where Reform did stand also captures higher mobilisation of the Reform-leaning electorate.
What the demographics predict
Running a multiple regression of Reform 2026 vote share on (no-quals %, IMD decile, retired %, social-rent %) across the nine wards yields R² = 0.716. That's 72% of the cross-ward variation in Reform vote explained by those four demographic axes alone. Single-predictor Pearson correlations:
| Predictor | Pearson r (with Reform 2026 %) | Direction |
|---|---|---|
| degree % | -0.853 | Strong negative |
| retired % | -0.737 | Strong negative |
| no quals % | +0.725 | Strong positive |
| imd decile | -0.704 | Strong negative |
| owned outright % | -0.629 | Moderate negative |
| social rented % | +0.539 | Moderate positive |
The strongest single predictor is the degree-holder share (r = -0.853). Reform wins big in low-education wards (Pemberton, Hindley, Abram, all 25%+ no-quals, Reform 52-56%) and collapses in graduate-rich wards (Orrell 33% graduates → Reform 39.9%). Asian % and White British % are not in the regression because all nine wards cluster too tightly to discriminate.
Which wards over- and under-performed?
Residuals (actual Reform 2026 share minus what the four-predictor regression expected). Positive = Reform did better than the demographic fundamentals predicted (likely candidate-quality or hyper-local effect); negative = the opposite.
| Ward | Actual | Predicted | Residual |
|---|---|---|---|
| Bryn with Ashton-in-Makerfield North | 52.2% | 47.1% | +5.0pp |
| Winstanley | 49.7% | 47.3% | +2.4pp |
| Hindley Green | 52.3% | 51.9% | +0.5pp |
| Abram | 56.1% | 55.7% | +0.4pp |
| Pemberton | 54.2% | 54.1% | +0.1pp |
| Hindley | 52.3% | 53.1% | -0.8pp |
| Worsley Mesnes | 50.9% | 53.0% | -2.0pp |
| Ashton-in-Makerfield South | 46.4% | 48.5% | -2.1pp |
| Orrell | 39.9% | 43.3% | -3.4pp |
Bryn-with-Ashton-North over-performs the model by +5pp, the biggest residual in the seat. Worth flagging as a candidate / local-organisation effect rather than fundamentals. Conversely Orrell (-3.4pp) is where Reform are weakest relative to its high-IMD / high-degree profile. That's the kind of ward Burnham's personal vote should pull hardest in.
Comparator seats. The seven closest Lab-vs-Reform 2024 marginals
How similar seats moved on 1 May 2026
Curtice noted Makerfield was the seventh-most marginal Labour-vs-Reform seat in the country at GE2024 (13.4pp Lab majority). The six tighter Lab-Reform seats plus the next two are below. Where the host council voted on 1 May 2026 we surface the LAD-wide aggregate as a directional read for the parliamentary seat; Welsh seats and councils outside the 2026 cycle are flagged.
| Constituency | Region | GE2024 Lab maj | 1 May 2026 host council | Reform 2026 | Labour 2026 | Reform lead |
|---|---|---|---|---|---|---|
| Llanelli | Wales | 3.7pp | Welsh, different cycle + Plaid factor | n/a | n/a | n/a |
| Amber Valley | East Midlands | 8.4pp | Borough not in 2026 cycle | n/a | n/a | n/a |
| Montgomeryshire & Glyndwr | Wales | 8.8pp | Welsh, different cycle | n/a | n/a | n/a |
| Great Grimsby & Cleethorpes | Yorkshire & Humber | 13.1pp | north-east-lincolnshire | 45.9% | 14.2% | +31.7pp |
| Kingston upon Hull East | Yorkshire & Humber | 13.1pp | Hull elections in different cycle | n/a | n/a | n/a |
| Bradford South | Yorkshire & Humber | 13.3pp | bradford | 22.6% | 21.5% | +1.1pp |
| Barnsley South | Yorkshire & Humber | 13.5pp | barnsley | 42.0% | 32.3% | +9.7pp |
| North Durham | North East | 14.1pp | Durham UA in different cycle | n/a | n/a | n/a |
| Makerfield (this seat) | North West | 13.4pp | wigan | 50.2% | 24.3% | +26.5pp |
What the comparators tell us
- Great Grimsby & Cleethorpes (NE Lincs): Reform 45.9 / Lab 14.2, a 31.7pp Reform lead, winning 14 of 15 wards. The closest demographic twin to Makerfield (ex-fishing / ex-heavy-industry, 90%+ White British, low graduate share) and the most extreme Labour collapse in the comparator set.
- Barnsley South (Barnsley MBC): Reform 42.0 / Lab 32.3, a 9.7pp Reform lead. Reform won the council outright (15 majorities nationally include Barnsley). Labour held up better than in Grimsby thanks to stronger local-government incumbency and ex-NUM organisational depth, but still a 9.7pp deficit.
- Bradford South (Bradford MBC): Reform 22.6 / Lab 21.5 / Green 21.4, a 1.1pp Reform lead but a four-way split. Bradford South is a structural outlier in this comparator set because its high Pakistani-heritage population dampens Reform's ceiling and splits the non-Reform vote between Lab + Green + independents. Makerfield does not have this dynamic, so the comparison should be read with caution.
- Llanelli, Amber Valley, Hull East, North Durham, Montgomeryshire & Glyndwr: host councils did not vote on 1 May 2026 (Welsh elections separate; Hull / Amber Valley / Durham in different cycles). No direct 2026 council readout available.
Across the three comparators with usable data, Reform's average council lead is 14.2pp. Makerfield's 26.5pp sits above that mean (nearly double the comparator average) and is closer to Grimsby's 31.7pp than to Barnsley's 9.7pp or Bradford's 1.1pp. The local-election fundamentals are worse for Labour in Makerfield than in most of its Lab-Reform-marginal peer group. Only Burnham's personal vote closes the gap, which is exactly what Survation and Britain Elects model.
Burnham personal vote. Calibration
Where the +20.7pp Labour uplift comes from
In the 1 May 2024 Greater Manchester mayoral election, Burnham scored 60.2% first-preference vote across Wigan borough, the local authority that contains every Makerfield ward. National Labour's brand on the same day stood at 33.7%. That gives Burnham a measured +26.5pp personal-vote uplift over the Labour brand in this electorate.
The model discounts that to +20.7pp (78% of the mayoral uplift) to reflect that a Westminster by-election is more partisan than an executive mayoral race and that the Starmer government's record is also on the ballot. Applied to the 24.3% ward baseline, this lands Labour at 45.0%, within rounding of Survation's published 45%.
Full methodology trace (11 steps)
- 120-year historical anchor : Makerfield and its predecessor Ince have returned a Labour MP at every general election since 1906. GE2024 produced the smallest Labour majority since 1931 (13.4pp). Treated as a soft prior; the contest fundamentals are dominated by the 1 May 2026 ward signal and the Burnham personal-vote shock.
- GE2024 baseline : Actual 4 July 2024 result in Makerfield. Labour 45.2% / Reform 31.8% / Con 10.9% / LD 6.8% / Green 4.4%. Reform was the runner-up, 13.4pp behind.
- 1 May 2026 ward signal : Aggregated weighted vote share across the nine Wigan wards inside the Makerfield PCON boundary, contested on 7 May 2026. Reform won every ward. Reform 50.2% / Labour 24.3%, a 25.9pp Reform lead, identical in sign to John Curtice's published 22.9pp on a slightly different 8-ward perimeter. Average turnout 37.3%.
- Burnham personal-vote uplift (calibration) : Burnham scored 60.2% first-pref in Wigan borough at the 1 May 2024 GM mayoral, vs 33.7% national Labour brand on the same day, a +26.5pp uplift. Applied at 78% strength (+20.7pp) to discount for partisan Westminster vs executive-mayoral framing. Cross-checked against the Survation Burnham scenario, which independently implies a +20.7pp uplift on Labour's 24.3% ward baseline.
- Survation 15 May 2026 pre-poll forecast : NB: Survation explicitly state 'this is a pre-poll forecast, not a poll'; no Makerfield fieldwork has been done. Burnham effect transferred from Survation's Gorton & Denton post-by-election poll (fieldwork 9 March 2026, n=501) and re-weighted for Makerfield's older, more White British, more Leave-voting electorate. Two scenarios from 10,000 simulations: Burnham stands → Lab 45 / Ref 42 (3pp Lab lead, P[Lab wins]=0.67); Burnham withdraws → Lab 27 / Ref 53 (26pp Ref lead, P[Lab wins]≈0).
- Britain Elects 15 May 2026 adjusted forecast : Ben Walker (Britain Elects / Britain Predicts) published a second adjusted forecast on 15 May. Method: GE2024 baseline + North West regional Burnham favourability (+20pp uplift over national; 50-60% in GM vs 30-40% nationally). Burnham takes ~5pp off Reform and ~4pp off Green. Result with Burnham as Labour candidate: Lab 39 / Ref 36, the same 3pp Lab lead as Survation but 6pp lower absolute levels for both major parties (implying a larger minor-party residual).
- Burnham-on-ballot probability (separate from P[Lab wins | Burnham]) : The two quantities are distinct: Survation's 0.67 is P[Labour wins | Burnham stands]; we need P[Burnham appears on the ballot]. Post-NEC clearance (15 May 2026), Burnham's same-day acceptance, and Downing Street's confirmation of no block, we set the latter at 0.85. Would only fall back on a fresh Code-of-Conduct issue or Burnham declining to formally resign as GM mayor (statutorily required within 6 weeks of taking up a Commons seat).
- Restore Britain entry effect : Rupert Lowe's Restore Britain (formed November 2025) will field its first Westminster candidate. Modelled as 1-4% of the vote drawn principally from Reform's right flank. Already incorporated in Survation's two scenarios; no further adjustment applied.
- Scenario A: Burnham stands (Survation + Britain Elects blend) : Major-party shares are a 50/50 mean of the two published forecasts (Survation Lab 45 / Ref 42; Britain Elects Lab 39 / Ref 36). Minor-party splits taken from Survation and scaled to the residual. ±6.0pp 1-sigma uncertainty on the winning margin.
- Scenario B: Burnham withdraws (contingency) : Anchored on Survation non-Burnham scenario (Britain Elects didn't publish one). Reform ceiling (~53%) consistent with the 1 May ward result (50.2%) plus by-election protest amplification. ±8.0pp 1-sigma.
- Probability-weighted blend : Final central forecast = 0.85 × Scenario A + 0.15 × Scenario B, renormalised.
Model version: makerfield-by-election-v1 · Generated 2026-05-26T09:50:05.960Z ·
Confidence: medium.
Stepped down from medium-high to medium after the per-ward + comparator analysis (see makerfield-2026-06-18.analysis.json). Reasons to weight Reform's side of the toss-up more heavily: (a) Makerfield's Reform-vs-Lab council lead (26.5pp) is nearly double the comparator average (14.2pp across Grimsby, Barnsley, Bradford); (b) the within-seat regression shows 72% of cross-ward Reform variance is explained by education + deprivation + retired + social-rent, and Burnham's brand is strongest in graduate-rich wards (Orrell, Winstanley) where Reform are already weakest, so the uplift narrows rather than reverses the gap in the no-quals wards (Pemberton, Hindley, Abram) where Reform are 52-56%; (c) two of the closest demographic comparators (Grimsby +31.7pp Reform, Barnsley +9.7pp Reform) bracket Makerfield's local-elections aggregate exactly. The 1.4pp Reform-leaning toss-up remains the central estimate but a 3-5pp Reform lead is now more probable than the bare blended figure suggests.