I treat this as a high-variance but mostly two-candidate race between Labour’s Andy Burnham and Reform UK’s Rob Kenyon. The official notice fixes polling day at 18 June 2026, and the official statement of persons nominated confirms a crowded ballot including Labour, Reform UK, Conservatives, Liberal Democrats, Greens, Restore Britain, and several smaller or independent candidates. In the 2024 general election, Labour won Makerfield with 45.2% to Reform’s 31.8%, a 13.4-point margin. (wigan.gov.uk)

The main bearish signal for Labour is the local-election collapse. PollCheck’s aggregation of the eight Wigan wards wholly inside Makerfield puts Reform on 49.8% and Labour on 26.9% in the 7 May 2026 local elections, with Reform winning every one of those eight wards. The official ward pages show especially large Reform leads in Abram, Bryn with Ashton-in-Makerfield North, Hindley, and Worsley Mesnes. Nationally, Labour is also weak: the latest YouGov Westminster poll, fielded 31 May-1 June 2026, had Reform on 27% and Labour on 18%. (pollcheck.co.uk)

The main bullish signal for Labour is that this is not a generic Labour candidacy. Survation’s pre-poll model said that with Burnham as candidate Labour would lead Reform 45% to 42% and have a 67% chance to win, whereas without him Reform would win clearly. Survation’s first constituency poll, fielded 18-22 May among 504 adults in Makerfield, found Burnham on 43%, Kenyon on 40%, Restore Britain on 7%, Lib Dems 4%, Greens 3%, Conservatives 2%, and others 1%. The same poll found Reform ahead by 11 points on a generic Westminster question, which is strong evidence that Burnham’s personal vote is what turns this into a toss-up. Burnham’s popularity is unusually resilient relative to Labour’s brand: YouGov says he is the most popular Labour figure with the public and the only senior politician it tested without a negative net rating, while Survation says he is the only Labour leadership contender with net positive favourability across all 2024 voter groups and voting-intention groups. (pollcheck.co.uk)

There are also smaller tactical effects that probably help Labour at the margin. Restore Britain is on the ballot and was already polling 7% in Survation’s constituency poll; I infer that this mostly hurts Reform rather than Labour, and John Curtice separately told Al Jazeera that Restore’s support is the vote share Reform really needs. Meanwhile, the Greens have signalled a scaled-back campaign, reducing the risk that anti-Reform tactical voting fragments too badly. Current bookmaker prices still make Labour favourite but far from safe: Oddschecker today shows Labour generally around 4/11 to 1/3, with Reform mostly 12/5 to 3. (survation.com)

My synthesis is: absent Burnham, Reform would probably be on course for a solid win; with Burnham, the race becomes Labour-leaning but still close. Because the ballot is crowded and the top two are still overwhelmingly likely to be Labour and Reform, most of the probability mass should sit on modest winning margins rather than landslides. I therefore center the forecast a little above the 3-point Survation poll lead, to reflect that the poll was taken very early and largely before Burnham’s formal campaign launch, plus the subsequent signs of a softer Green challenge and continued right-fragmentation via Restore Britain. But I keep a fat lower tail because national conditions remain hostile to Labour and the May local-election signal was extremely bad. My median forecast is about 5 points, with mean about 5.3 points; roughly half the mass is between about 2.5 and 8 points, and outcomes above 12 points require a late tactical squeeze or campaign shock. (survation.com)

Makerfield votes on Thursday 18 June 2026. The official statement of persons nominated lists 14 candidates, including Andy Burnham for Labour and Robert Kenyon for Reform UK. In the 2024 general election, Labour won Makerfield with 18,202 votes (45.2%) to Reform’s 12,803 (31.8%), a 13.4-point margin. (wigan.gov.uk)

The strongest quantitative evidence is Survation’s work. Its pre-poll estimate, published on 15 May 2026, said that without Burnham Reform would be likely to win 53% to 27%, but with Burnham the race tightened to Labour 45% and Reform 42%, with Labour winning roughly 67% of simulations. Then Survation’s actual constituency poll, fielded 18-22 May, found Burnham on 43%, Kenyon on 40%, Restore Britain on 7%, the Liberal Democrats on 4%, the Green candidate on 3%, and the Conservatives on 2%. In the same polling exercise, a generic Westminster ballot in Makerfield had Reform ahead of Labour 45% to 34%. My read is that the seat is structurally pro-Reform at the moment, but Burnham’s personal vote is large enough to turn it back into a toss-up. (pollcheck.co.uk)

I put real weight on Burnham’s personal vote because he is unusually well known locally: in the 2024 Greater Manchester mayoral election he won 44,927 of 67,975 valid votes in Wigan district, or 66.1%. But I also put real weight on Reform’s local strength and on Kenyon as a non-token challenger: he was Reform’s 2024 runner-up in Makerfield and then won the Bryn with Ashton-in-Makerfield North council seat on 7 May 2026 with 1,770 votes to Labour’s 816. (wigan.gov.uk)

The main uncertainty is how much of Burnham’s personal vote survives in a Westminster by-election fought in a much harsher national climate for Labour. Survation explicitly said its constituency poll was early, conducted before Burnham’s campaign launch and during a volatile candidate picture. The 14-candidate ballot also means there is plenty of room for vote-splitting. Restore Britain’s 7% in the only published constituency poll is especially important; I infer that this fragments the anti-Labour vote to some extent, though late tactical squeezing toward Kenyon could reduce that effect. (survation.com)

So I treat this as overwhelmingly a Labour-versus-Reform top-two race, and I model the resolution amount mainly as the absolute Labour-Reform gap. My median is in the mid-3s and my mean is about 4 points. Most of my probability mass is in roughly the 1-8 point range, with a smaller but still meaningful tail into the low teens if either Burnham’s personal vote or a late anti-Labour squeeze proves stronger than the current polling implies. A near-photo-finish remains very plausible, because both Survation’s model and its actual constituency poll landed on a 3-point Labour edge. (pollcheck.co.uk)

Makerfield votes on 18 June 2026. The official statement of persons nominated dated 26 May confirms Andy Burnham for Labour and Kenyon for Reform on the ballot, alongside a large minor-party field, so this is now clearly a live Labour-versus-Reform by-election rather than a hypothetical Burnham candidacy. My central view is that the race is genuinely close and that the most likely margin of victory is in the mid-single digits. (wigan.gov.uk)

The historical baseline is still Labour-favourable. At the 4 July 2024 general election, Labour won Makerfield with 18,202 votes to Reform’s 12,803, a 45.2% to 31.8% split, on 40,263 valid votes. That was a 13.4093-point margin of victory. (wigan.gov.uk)

But the short-term environment is much worse for Labour than that 2024 result suggests. The latest YouGov GB voting-intention poll, fielded 31 May-1 June and published 2 June, put Reform on 27% and Labour on 18%. At local level, Reform also beat Labour heavily in a string of Wigan ward contests on 7 May that are inside or overlap materially with Makerfield: Abram 1958-844, Ashton-in-Makerfield South 1572-1100, Bryn with Ashton-in-Makerfield North 1770-816, Hindley 1832-750, Orrell 1621-983, Winstanley 1881-1174, Worsley Mesnes 1711-847, and Hindley Green 1878-1172. That is consistent with a seat that would ordinarily be highly vulnerable to Reform in the current climate. (yougov.com)

The reason I do not just extrapolate those ward results is the Burnham effect. In the 2024 Greater Manchester mayoral election, Burnham took 44,927 of 67,975 votes in Wigan district, or 66.0934%, which is about 20.8857 points above Labour’s 2024 parliamentary share in Makerfield. Nationally, YouGov says Burnham remains Labour’s most popular senior figure, and Ipsos likewise finds him the most popular alternative Labour leader. Most importantly, Survation’s constituency poll of Makerfield, fielded 18-22 May among 504 adults, found the named-candidate by-election vote at Labour 43% and Reform 40%, while the same poll found Reform ahead by 11 points on a generic Westminster ballot. Survation also says the poll was conducted very early in the campaign, largely before Burnham’s launch, and that the headline margin of error is about ±5.4 points, so I treat it as directional rather than definitive. (wigan.gov.uk)

Market pricing broadly agrees that Burnham is favoured but far from safe. Oddschecker quoted Labour at 4/11 and Reform at 3/1 on 27 May; normalising those quoted prices gives roughly a 69.6% Labour win probability and a 23.7% Reform win probability, with the remainder on long-shot others. Because Burnham is now officially nominated, the very bearish non-Burnham branch in Survation’s pre-poll forecast is no longer the live downside case; the real question is how much of Burnham’s personal vote survives contact with a very anti-government local and national mood. (oddschecker.com)

So my synthesis is: Labour is a modest favourite to hold, but probably not by much. The top two are very likely Labour and Reform, given Survation’s party levels and the betting market, while the crowded ballot makes it unlikely that either of them runs away to 50%+. My inference is a three-scenario race: a close Burnham-favoured contest as the main case, a smaller but real Reform-overperformance case, and a less likely Burnham squeeze/tactical-consolidation case. That leads me to centre the margin distribution around about 5 points, with most probability mass between roughly 1 and 10 points and only modest tail risk beyond that. In other words, my median-style read is that the Makerfield by-election is most likely to be decided by about five percentage points. (survation.com)

Makerfield votes on 18 June 2026. Wigan’s official notice shows a crowded ballot including Andy Burnham for Labour, Rob Kenyon for Reform UK, Rebecca Shepherd for Restore Britain, Michael Winstanley for the Conservatives, Jake Austin for the Liberal Democrats, Sarah Wakefield for the Greens, and several smaller-party or independent candidates. That matters because a crowded field usually keeps the winner’s share down, and Restore Britain in particular can split the anti-Labour or right-leaning vote. (parliament.uk)

At the 2024 general election Labour won Makerfield with 45.2% to Reform’s 31.8%, a 13.4-point margin. But the political environment has deteriorated sharply for Labour: Survation’s pre-poll model, using the 2024 general election result, May 2026 Wigan council results, and constituency demographics, estimated that with an ordinary Labour candidate Reform would win Makerfield 53% to 27%. The same note says Reform won seven of the eight Makerfield wards used in the model. Since the 2024 general election Labour has already lost one parliamentary by-election to Reform in Runcorn and Helsby and another to the Greens in Gorton and Denton, which is strong evidence that the current by-election environment is punishing for the governing party. (election.news.sky.com)

Burnham materially changes the race. Survation’s 15 May model estimated that with Burnham as Labour candidate the race tightens to Labour 45% and Reform 42%, with Labour winning roughly two thirds of simulations. Survation’s first actual Makerfield constituency poll, a telephone survey of 504 adults fielded 18-22 May, put Burnham on 43%, Kenyon on 40%, Restore on 7%, Liberal Democrats on 4%, Greens on 3%, and Conservatives on 2%. The same poll found Reform led Labour by 11 points on a generic Westminster question in Makerfield, implying a large Burnham personal vote. It also found that 82% of 2024 Reform voters stayed with Reform but 13% had moved to Restore, while 83% of 2024 Labour voters stayed Labour. (cdn.survation.com)

Why not simply forecast a 3-point margin, then? First, the poll was taken more than two weeks before the 18 June vote and its headline margin of error is about ±5.4 points. Second, national conditions remain awful for Labour: YouGov’s 31 May-1 June Westminster poll had Reform on 27% versus Labour on 18%, and Ipsos in late May also had Reform ahead of Labour by 6 points. Third, by-elections can move late as tactical voting and campaign salience kick in. Against that, Burnham is unusually strong locally: in the 2024 Greater Manchester mayoral count for Wigan district he took 44,927 votes out of 67,975 valid votes, and Forward Democracy’s tactical-voting site is already recommending Labour in Makerfield to stop Reform. Those offsetting forces make a single-digit result more likely than either a photo finish or a landslide. (survation.com)

My synthesis, as an inference from the evidence above, is that Labour remains a slight favourite because Burnham’s personal vote looks real and because Restore Britain reduces Reform’s clean path to consolidation, but the anti-government mood is strong enough that a Reform gain is still very plausible. I put Labour around 58%, Reform around 40%, and everyone else collectively around 2% to win. Conditional on either Labour or Reform winning, the most likely margin is in the low single digits; there is a meaningful secondary cluster in the high single digits if tactical squeezing or vote-splitting strengthens; and there is only a small tail into the mid-teens or above. That gives me a median margin near 5 percentage points and a mean a little above 6. (cdn.survation.com)

As of June 3, 2026, Wigan Council says the Makerfield parliamentary by-election will be held on Thursday, 18 June 2026. The official statement of persons nominated lists 14 candidates, including Andy Burnham for Labour and Co-operative, Rob Kenyon for Reform UK, Michael Winstanley for the Conservatives, Sarah Wakefield for the Greens, Jake Austin for the Liberal Democrats, Rebecca Shepherd for Restore Britain, plus several smaller-party and independent candidates. Given the 2024 result and the only constituency poll so far, Labour and Reform are overwhelmingly likely to be the top two, so the margin is best modeled as the absolute Labour-Reform gap. (wigan.gov.uk)

In the 2024 general election, Labour won Makerfield with 18,202 votes, 45.2%, against Reform’s 12,803, 31.8%, a margin of 13.409333631373716 percentage points. But the recent local-election signal is much worse for Labour. Using the eight full Wigan wards that Boundary Commission data places wholly inside Makerfield and noting that the two small partial wards account for only 2.7667054380072402% of the constituency electorate, I get a rough 7 May 2026 council-election baseline of Reform 49.784731702194684% and Labour 26.903286779376245%, a Reform lead of 22.88144492281844 points. I treat that as a warning sign rather than a direct forecast, because council elections had much lower turnout and did not feature Burnham as a parliamentary candidate. (members.parliament.uk)

Burnham clearly narrows that gap. Survation’s pre-poll forecast on 15 May projected a Burnham scenario of Labour 45% to Reform 42% and said Labour held the seat in roughly two thirds of simulated outcomes. Survation’s first actual constituency poll, fielded 18-22 May among 504 adults in Makerfield, found Burnham on 43% and Kenyon on 40%. The same poll said Reform led Labour by 11 points on a generic Westminster ballot, implying Burnham’s personal vote roughly erased the anti-Labour baseline without making the seat safe. Survation also found 13% of Reform’s 2024 voters moving to Rebecca Shepherd’s Restore Britain, and it noted that the poll was conducted largely before Burnham’s formal campaign launch on 22 May, which is one reason to leave some upper tail for Labour. (cdn.survation.com)

Other current signals cut both ways. Nationally, YouGov’s 31 May-1 June 2026 poll had Reform on 27% and Labour on 18%, confirming the difficult national backdrop for Labour. On the other hand, the betting market on 27 May still made Burnham the favorite, with Labour at 4/11 and Reform at 3/1. I therefore lean to a Labour hold, but only narrowly. (yougov.com)

I also want a wide distribution because recent by-elections have been extremely volatile. Runcorn and Helsby in May 2025 was decided by just 6 votes, which works out to a 0.01837390904915021-point margin, while Gorton and Denton in February 2026 was won by the Greens by 11.957407508013256 points. Those two cases are a reminder that 2025-26 by-elections have produced both photo finishes and double-digit surprises. (parliament.uk)

My synthesis is a small Labour edge in signed terms, around 2 points, driven by Burnham’s candidate effect and some tactical or personal-vote squeezing of smaller parties, but with a substantial Reform-gain path because the local and national environment is plainly hostile to Labour. Because the resolution variable is the absolute margin rather than the signed lead, that translates into an absolute-margin distribution centered higher than 2 points: both a narrow Labour hold and a narrow Reform gain usually produce margins of a few points, not a perfect tie. The code below represents that by modeling the signed Labour-minus-Reform lead as a four-component normal mixture: a narrow Labour-hold scenario, a wider Labour-hold scenario, a Reform-gain scenario, and a knife-edge scenario. This yields a distribution centered around roughly 4-5 percentage points, with most mass in the low-to-mid single digits but meaningful tails into double digits. (survation.com)

What will be the margin of victory in the Makerfield by-election?

• My most likely prediction for the margin of victory in the Makerfield by-election is 1.50 percentage points, with a 50% chance of falling between 0.50 and 4.50 percentage points.
• Recent constituency-specific polling shows a very tight race, with Andy Burnham (Labour) leading Robert Kenyon (Reform UK) by only 0.8 to 3.0 percentage points.
• While Reform UK dominated the May 2026 local elections in the area, Andy Burnham’s personal popularity (having won 66% in the 2024 mayoral election in Wigan) provides a significant ‘personal vote’ boost for Labour.
• The presence of the “Restore Britain” party, polling at approximately 7%, is likely to split the right-wing vote, potentially helping Burnham secure a win but narrowing the overall margin between the top two candidates.
• Market forecasts on Polymarket favor a Burnham victory (73% probability) but show high uncertainty regarding the margin, with significant weight on both “less than 3%” and “9% or more.”
• Economic sentiment in the UK is extremely low, with consumer confidence at -23.0, which may suppress turnout or fuel anti-incumbency sentiment against the Labour candidate.

Key figures

Historical context

The Makerfield constituency has been a Labour stronghold since its creation in 1983. However, it is an area with high ‘Leave’ sentiment (65% in the 2016 referendum) and a demographic profile (97% white, older, fewer graduates) that has become increasingly receptive to Reform UK. In the 2024 General Election, Labour held the seat with 45.2% of the vote against Reform’s 31.8%, a 13.4-point margin. This by-election is historically significant as the first time since 1965 (Leyton) that a seat was vacated specifically to allow a figure not currently in Parliament to seek a leadership path. The May 2026 local elections marked a turning point, where Reform UK won 24 of 25 seats in the Wigan borough, indicating a total breakdown of the traditional ‘Red Wall’ Labour alignment in the region.

Tailwinds

• The “Burnham Factor”: Personal popularity estimated to provide a 20-percentage-point boost over a standard Labour candidate.
• Right-wing vote splitting: The “Restore Britain” party (7% in polls) and other minor candidates are drawing votes primarily from Reform UK.
• Tactical voting: Potential for Green and Liberal Democrat voters to switch to Burnham to prevent a Reform UK victory.
• High household income and home ownership (74%) in the area compared to the national average may provide a more stable environment for Labour than other “Red Wall” seats.

Headwinds

• Reform UK’s massive momentum in May 2026 local elections, where they secured 50% of the vote across Makerfield wards.
• Extremely low national consumer confidence (-23.0) and business confidence (-65.0), which typically penalizes the incumbent party (Labour).
• Voter disillusionment with the seat being used as a “stepping stone” for Burnham’s national leadership ambitions.
• Reform UK’s highly motivated base compared to a potentially demoralized Labour turnout in industrial areas.

Detailed reasoning

My analysis of the Makerfield by-election margin of victory is centered on the conflict between two powerful, opposing electoral forces: the collapse of the national Labour brand in industrial heartlands and the significant ‘personal vote’ premium of Andy Burnham. In the 2024 General Election, the margin was 13.4 points. However, by May 2026, local election results indicated a catastrophic shift, with Reform UK winning all eight wards in the constituency and outpolling Labour by roughly 25 percentage points.

Direct polling since Burnham’s candidacy was announced suggests he has successfully neutralized much of that Reform momentum, but the race remains extremely tight. Three independent polls (Survation, Convergent, and JL Partners) all point to a margin of between 0.8 and 3.0 percentage points. My forecast places the most likely margin at 1.50 points to reflect this consensus of a ‘knife-edge’ contest.

The probability distribution is weighted heavily toward these low-single-digit outcomes, but it maintains a significant ‘right-tail’ to account for a possible ‘Burnham surge.’ Expert analysis from More in Common suggests Burnham provides a personal boost of up to 20 points over a standard Labour candidate. If this personal popularity effectively consolidates the ‘progressive’ vote (Liberal Democrats and Greens) while Reform UK’s vote is splintered by the ‘Restore Britain’ party (which is polling at 7%), the margin could widen into the 6-9% or even 9%+ range.

Market data from Polymarket supports this dual-peak possibility: while the winner market is liquid and strongly favors Burnham (73%), the margin-specific brackets show substantial uncertainty, with 35% of the market expecting a margin of 9% or more, while others cluster around the ‘less than 3%’ mark. My prediction reconciles these by centering on the tightest polling results while allowing for a wider HDI that covers the possibility of a more comfortable win. The absolute nature of the margin calculation also means that even a narrow Reform upset would result in a small positive value, further concentrating probability mass at the lower end of the scale.

Key uncertainties

• Voter turnout: By-elections often have low turnout; a “motivated” Reform base versus a “disillusioned” Labour base could swing the margin by 5+ points.
• The “Restore Britain” impact: Canvassing data suggests they could overperform their 7% poll rating, which would drastically change the Reform vs. Labour gap.
• Tactical consolidation: If progressive voters do not consolidate behind Burnham, the margin could narrow to less than 1 point or result in a Reform victory.
• Polling error: Small sample sizes (n=369 to 504) in constituency polls carry a high margin of error, typically +/- 4-5 points.

Conclusion

• The Makerfield by-election is a high-stakes, two-horse race where Andy Burnham’s personal popularity must counteract a massive localized surge for Reform UK.
• While Burnham is the heavy favorite to win, the margin of victory is expected to be significantly narrower than the 13.4 percentage point gap seen in 2024.
• Decision-makers should monitor “Restore Britain” and turnout levels; a split on the right or low Reform engagement could unexpectedly widen Burnham’s margin.

Question: what will the Labour–Reform margin be in the Makerfield by‑election? Forecasters agree the single best numeric anchor is a Survation constituency poll (18–22 May) showing Burnham/Labour 43% vs Reform 40%, but they strongly weight that single poll with wide uncertainty. Key contrary evidence is May 2026 Wigan local results in Makerfield wards where Reform led Labour by ~22.9 points on low turnout, and recent North West by‑election precedents (Runcorn, Heywood, Gorton/Denton) show large anti‑Labour swings are possible; betting markets and Burnham’s named‑candidate effect temper that, keeping Labour the favorite. The final ballot and vote‑splitting — notably Restore Britain and a collapsed Conservative vote — are central margin drivers: if Reform consolidates anti‑Labour votes it can overtake Labour, but if Restore/Greens siphon enough right‑ or left‑leaning votes Burnham can win comfortably. Forecasters therefore place most mass on a close, single‑digit margin, while allowing meaningful tails for either a Reform local‑style landslide or a clearer Burnham personal‑vote victory depending on turnout and late campaign movement.

Forecast and Reasoning

Forecast: Probability distribution:

• 1.00% chance of value below 0.1
• 5.00% chance of value below 0.6
• 10.00% chance of value below 1.4
• 20.00% chance of value below 2.4
• 30.00% chance of value below 3.4
• 40.00% chance of value below 4.6
• 50.00% chance of value below 6.0
• 60.00% chance of value below 7.4
• 70.00% chance of value below 9.3
• 80.00% chance of value below 12.3
• 90.00% chance of value below 16.5
• 95.00% chance of value below 22.0
• 99.00% chance of value below 34.0

Forecast Reasoning

Key evidence across analysts is consistent: (1) the only constituency poll cited (Survation, 18–22 May) has a very tight top-two gap (about 3 percentage points between Labour/Burnham and Reform/Kenyon), and (2) recent local-election/topline indicators suggest the seat can look substantially more Reform-leaning under different turnout/composition, increasing uncertainty. With ~15 days to go, there is limited time for many more high-quality constituency polls, so residual uncertainty remains large.

Because the question resolves to the winner’s margin (always non-negative), the distribution should place substantial mass on small margins (a genuine risk of a near-tie given the poll’s sampling error plus by-election-specific turnout error), but also retain a meaningful right tail: by-elections can produce lopsided results if one side’s turnout/coordination is much better, if minor-party support (e.g., Restore Britain / small parties) collapses asymmetrically, or if late campaign dynamics produce a bandwagon effect.

I synthesize the four forecasts by keeping the center near the poll-implied “single digits” (medians in three forecasts are ~5–6pp, one is ~8pp) while adopting a moderately fat upper tail (closer to Forecasters 3/4 than Forecaster 2). Net: median about 6pp, with a wide 10–90 interval (~1.4–16.5pp) reflecting by-election volatility and the fact we are relying heavily on one poll plus indirect signals.

An amateur forecast would likely anchor too tightly to the single Survation topline gap (~3pp) and understate uncertainty, or else over-extrapolate from either the previous general-election margin (13.4pp) or recent local-election strength into a confident “blowout” call. This forecast improves by (a) explicitly treating the outcome as the winner’s absolute margin (so both ‘Labour wins’ and ‘Reform wins’ scenarios contribute to small-but-nonzero margins), (b) widening the distribution to reflect by-election turnout/polling-miss risk, and (c) retaining a realistic but not dominant upper tail for asymmetric minor-party consolidation or turnout-driven runaways. Confidence in improvement: moderate, because the evidence base is still thin (notably, one main constituency poll).

Forecast rationale (numeric):

— Iteration 1 — Across the forecasts, the core picture is a close but likely Labour-leaning contest, with the expected margin of victory clustered around 6–7 percentage points.

Main factors driving the estimates

• Polling signal: The main named-candidate poll shows Burnham ahead by about 3 points, which anchors the race as competitive rather than secure.
• Market/implied probabilities: Prediction markets are treated as a useful aggregate of available evidence, pointing to a median margin near 7–7.5 points.
• Local political context: Reform’s strong local showing, including winning all 8 wards in May, suggests meaningful opposition momentum and keeps the race tight.
• Candidate quality: Labour’s use of Andy Burnham is seen as a major asset, with a plausible personal-vote boost that could widen the margin.
• By-election volatility: All forecasts stress that turnout, late campaign shifts, and local dynamics can produce outcomes substantially different from headline polling, so the distribution has a wide tail.

Areas of consensus

• The race is not expected to be a landslide by default, but neither is it likely to be a pure knife-edge.
• There is substantial probability mass around small-to-moderate margins.
• The result is modeled with fat tails, meaning both a very close race and a more decisive win remain plausible.

Main disagreement / uncertainty

• The biggest uncertainty is how much Burnham’s presence offsets Reform’s local momentum.
• Some reasoning gives more weight to the poll and local Reform strength, while other reasoning leans more heavily on market aggregation and Labour’s candidate advantage.
• There is also uncertainty about whether the final margin will reflect a near tie or a clearer Labour win, but the central estimate stays in the mid-single digits.

Overall synthesis

The collective reasoning points to a tight by-election with a likely Labour victory and a margin around 6–7 points, while acknowledging a meaningful chance of both a much closer outcome and a more comfortable Labour win.

— Iteration 2 — Across the forecasts, the main expectation is a Labour victory by a low-to-mid single-digit margin, centered around about 6 percentage points.

Core reasoning patterns

• Polling anchor: The Survation named-candidate poll showing roughly a 3-point Labour lead is the main numerical baseline.
• Local Labour advantage: Burnham’s strong local profile is treated as a meaningful boost to Labour beyond the poll.
• Market reinforcement: Bookmaker and prediction-market signals broadly support a Labour win, with some evidence that the margin could be larger than the poll suggests.
• Vote splitting / turnout effects: Possible right-of-centre fragmentation, lower Reform turnout, and by-election-specific turnout swings are used to explain why Labour could outperform the narrow poll lead.
• By-election volatility: All forecasts widen the range substantially because UK by-elections can produce surprising swings, especially with a fragmented field.

Areas of consensus

• The race is expected to be competitive but Labour-favored.
• A near-tie or very narrow outcome is plausible, but not the central case.
• The distribution is right-skewed, meaning modest Labour wins are more likely than large ones, but there is a meaningful chance of a bigger-than-expected result.

Main differences in emphasis

• Some forecasts lean more heavily on the tight-poll scenario, keeping the median close to 3–6 points.
• Others give more weight to Burnham’s local strength and market-implied upside, pushing the average a bit higher.
• Tail assumptions vary: all allow for upset/close-race outcomes, but some assign more probability to double-digit Labour wins and a few even leave room for a small Reform win.

Overall synthesis

The collective view is that the Makerfield by-election is most likely to be a Labour hold by around 6 points, with uncertainty driven by poll error, turnout, local candidate effects, and fragmented opposition voting.

— Iteration 3 — Across the forecasts, the main consensus is that the by-election is likely to be close but not necessarily razor-thin, with a central margin in the mid-to-high single digits — roughly 6–7 percentage points. The models generally treat Labour/Burnham as favored, but not securely so.

Main reasoning patterns

• Polling anchor: All forecasts lean heavily on the Survation named-candidate poll, which shows Labour only about 3 points ahead. That serves as the baseline for a competitive race.
• Reform strength: They also incorporate evidence that Reform is unusually strong locally, including broader local-election or generic-ballot signals, which keeps the race tight.
• High-variance by-election dynamics: The forecasts emphasize that by-elections can move sharply because of turnout differences, late swings, postal votes, and campaign effects.
• Candidate-specific effects: Burnham’s profile/celebrity advantage is treated as a factor that could enlarge the Labour margin beyond the poll lead, especially if Labour mobilizes strongly.
• Vote splitting / opposition fragmentation: Some scenarios assume anti-Labour vote splitting could either help Labour win more comfortably or, if Reform consolidates, keep the margin very tight.

Areas of agreement

• The race is competitive.
• The expected outcome is not a landslide by default, but a modest Labour win is the central case.
• There is substantial uncertainty, with meaningful probability of both a near-photo-finish and a low-double-digit win.

Areas of disagreement

• The forecasts differ mainly on how wide the upper tail should be:
  • some put more weight on a modest 6–8 point win,
  • others allow more room for a double-digit or even larger Labour victory if Burnham’s personal vote and turnout advantages materialize.
• They also vary in how much they trust polling versus market-implied probabilities, but both sources point to a wide distribution rather than a single precise number.

Bottom line

The collective picture is a tight, volatile contest, with the most likely result being a Labour/Burnham victory by around 6–7 points, but with enough uncertainty that the final margin could plausibly be much smaller or meaningfully larger.

Here is a synthesis of the key reasoning and shared logic across the forecasters’ rationales:

Key Factors Shaping the Makerfield By-Election

The “Burnham Premium” vs. The National Labour Brand Forecasters universally highlight the tension between Labour’s deeply unpopular national brand and the towering personal popularity of their candidate, Andy Burnham. Recent local elections in the overlapping council wards demonstrated a massive swing toward Reform UK. However, constituency-specific polling suggests that Burnham’s strong local profile as Mayor of Greater Manchester almost completely mitigates this generic deficit, turning what would otherwise be a comfortable Reform victory into a dead heat.

Enthusiasm Gap and Turnout Dynamics By-elections historically suffer from lower turnout, which frequently punishes the incumbent party. Forecasters emphasize a severe “enthusiasm gap” in this race: Reform UK supporters are highly motivated to cast their ballots, while Labour’s traditional base is suffering from mid-term apathy. This differential turnout strongly favors the insurgent campaign and threatens to erase Burnham’s localized appeal.

The “Parachute” Penalty The by-election was explicitly engineered to bring Burnham into Parliament, a political maneuver that forecasters note voters often punish. Drawing parallels to historic anti-incumbent backlashes (such as the 1965 Leyton by-election), rationales warn that local resentment over a manufactured vacancy could further depress Labour turnout and drive protest votes toward Reform UK.

Tactical Voting and Minor Party Squeezes The absolute margin of victory will be heavily influenced by whether third-party voters consolidate behind the frontrunners. Labour’s margin relies on squeezing progressive voters (Greens and Liberal Democrats) who want to tactically block Reform. Conversely, Reform UK’s ceiling depends on its ability to absorb right-wing voters from the collapsed local Conservative base and the newly formed splinter party, Restore Britain.

Electoral Volatility Because this race pits intense local populist momentum against a formidable individual candidate brand, rationales generally expect a razor-thin margin of victory resulting from a fierce tug-of-war. However, forecasters uniformly maintain that a blowout victory for either side remains highly plausible. If the enthusiasm gap and parachute penalty collapse Labour’s base, Reform could win decisively; conversely, if Burnham’s ground campaign successfully mobilizes his loyal supporters and completely squeezes the minor parties, Labour could secure a surprisingly comfortable margin.

SUMMARY

Question: What will be the margin of victory in the Makerfield by-election? Final Prediction: Probability distribution:

• 10.00% chance of value below 1.193333
• 20.00% chance of value below 2.42
• 40.00% chance of value below 4.793333
• 60.00% chance of value below 7.523333
• 80.00% chance of value below 11.173333
• 90.00% chance of value below 14.6

Total Cost: extra_metadata_in_explanation is disabled Time Spent: extra_metadata_in_explanation is disabled LLMs: extra_metadata_in_explanation is disabled Bot Name: extra_metadata_in_explanation is disabled

Report 1 Summary

Forecasts

Forecaster 1: Probability distribution:

• 10.00% chance of value below 1.05
• 20.00% chance of value below 2.1
• 40.00% chance of value below 4.25
• 60.00% chance of value below 6.75
• 80.00% chance of value below 10.5
• 90.00% chance of value below 13.8

Forecaster 2: Probability distribution:

• 10.00% chance of value below 1.1
• 20.00% chance of value below 2.3
• 40.00% chance of value below 4.75
• 60.00% chance of value below 7.45
• 80.00% chance of value below 11.2
• 90.00% chance of value below 14.4

Forecaster 3: Probability distribution:

• 10.00% chance of value below 1.43
• 20.00% chance of value below 2.86
• 40.00% chance of value below 5.38
• 60.00% chance of value below 8.37
• 80.00% chance of value below 11.82
• 90.00% chance of value below 15.6

Research Summary

The research summarizes that the Makerfield by-election (scheduled for 18 June 2026) was triggered by Labour MP Josh Simons’ resignation to allow Andy Burnham to stand. The 2024 general-election baseline: Labour 45.2% (18,202 votes) vs Reform UK 31.8% (12,803), a 13.4 percentage-point Labour margin on 52.5% turnout. Recent data show a much tighter race: a Survation poll (18–22 May 2026) gave Labour 43% and Reform UK 40% (±5.4% MOE), while May 7, 2026 local election results in the Makerfield area showed Reform UK 49.8% vs Labour 26.9% (an approximate 23-point local swing). The research also notes Labour’s poor by‑election record since 2024 (losses including Gorton & Denton and a 6‑vote loss in Runcorn & Helsby), a Labour leadership crisis tied to Burnham’s candidacy, and large betting/prediction-market activity (Polymarket ~ $2M volume) with late‑May bookmaker odds implying roughly 65–71% for Labour and ~22–25% for Reform UK.

Quantitative markers highlighted for margin forecasting include the 13.4‑point 2024 baseline, the Survation 3‑point lead for Labour (within the poll’s error margin), the 23‑point local‑election swing toward Reform UK, and betting‑market odds that still favor Labour despite the local results and recent by‑election losses. The research lists turnout uncertainty, Burnham’s personal profile, vote mobilization, and minor‑party vote shares as key uncertainties affecting the ultimate margin.

Sources used (as cited in the supplied research; no verbatim URLs were provided in the material):

• Survation poll (May 18–22, 2026)
• May 7, 2026 local election results for Makerfield wards
• 2024 Makerfield general election constituency results (vote counts and percentages)
• Polymarket market (reported ~$2M trading volume)
• Late‑May bookmaker odds summaries (Labour 4/11–7/20; Reform UK 3/1–10/3)
• News reporting on Andy Burnham’s candidacy and Labour leadership crisis
• Historical by‑election results cited (Gorton & Denton; Runcorn & Helsby)

RESEARCH

Report 1 Research

I’ll help you gather comprehensive information about the Makerfield by-election. Let me conduct several searches to find relevant news, prediction markets, and historical data.

Forecast — Post 407

Modal: 0-12.5 (87.0%) • partial_shift • 6m2s

Interpretation

Reading: strict

The resolution criteria provide an explicit, unambiguous mathematical formula (first minus second, divided by total, times 100) with a worked example, leaving no room for interpretive latitude. The only non-trivial rules (use initial official release, ignore post-election career of winner) are clearly stated and narrow.

Time bounds

• observation_window: Official vote totals from the initial official release of the Makerfield by-election result, scheduled 18 June 2026
• resolution_date: 2026-06-19

Edge cases identified

1. Election scheduled for 18 June 2026 but resolution date is 19 June 2026 — if the result is not declared until after midnight or later, resolution still waits for it per the fine print.
2. Rounding: the formula produces a continuous value; the example given (13.4) is shown to one decimal place, suggesting results will be reported to at least one decimal point.
3. If there is a tied result (zero margin) or a very narrow win, resolves to 0.0 or near-zero.
4. The fine print specifies only the initial official release counts — any subsequent correction or recount is ignored.
5. Andy Burnham is the expected Labour candidate; if he withdraws or a different candidate stands for Labour, the same formula still applies to whoever finishes first and second.
6. The constituency voted Labour in 2024 with a 13.4pp margin; the by-election context (triggered for Burnham) may produce a very different margin.

Research (3/3 variants, shared evidence pool)

Total evidence registered (shared pool): ?

Research Brief

Evidence confidence: medium

Scenario 1: Narrow Labour (Burnham) win — margin in the 1–6pp range [high evidence]

Conditions favoring

Survation poll (3pp lead) and the JL Partners mini-poll both suggest a tight Labour lead; turnout differential narrows the margin further; Burnham’s personal premium is real but modest; Restore Britain vote-split delivers the seat to Labour on a minority of the vote; differential turnout partially offsets the Burnham premium.

Conditions against

If Burnham motivates higher Labour turnout than Kenyon, the margin could be larger; if Restore Britain collapses tactically toward Reform on polling day, the margin disappears.

Scenario 2: Moderate-to-comfortable Labour win — margin in the 6–12pp range [medium evidence]

Conditions favoring

Raw unweighted Survation data shows a 10pp lead (source 14); Burnham’s campaign launched after fieldwork and could boost turnout; Polymarket assigns 49% combined to Burnham winning by 6pp+ (sources 16, 23); level forecast median/mean in several models sits around 5–8pp; bookmakers at 74% implied Labour win probability could reflect wider margin expectations.

Conditions against

Turnout-filtered polling consistently shows only a 3pp lead; Reform swept local elections across the constituency; national Labour polling at historic lows; differential turnout favours Reform.

Scenario 3: Very narrow Labour win or near-tie — margin under 2pp (Labour wins but barely) [medium evidence]

Conditions favoring

Survation lead is within margin of error; Kenyon has 90% vs 74% turnout certainty; Reform swept all wards at May 2026 local elections; Convergent analysis suggests Burnham only at parity with Reform even after reclaiming all Labour defectors (source 11); Polymarket assigns ~10% to Labour winning by under 3pp (source 23).

Conditions against

Polling consistently shows Labour above Reform on a named-candidate basis; the Burnham premium is well-documented and substantial; Restore Britain vote split structurally advantages Labour.

Scenario 4: Reform UK (Kenyon) win — negative margin for Labour [medium evidence]

Conditions favoring

Restore Britain vote-split could collapse on polling day (Lowe voters switching to Reform); historical base rate of Labour losing northern by-elections (Runcorn, Gorton & Denton); 1965 Leyton parallel of Labour losing the ‘get-minister-into-Parliament’ seat; Reform already leads nationally; Reform swept Makerfield wards in May 2026.

Conditions against

Multiple polls show Labour ahead on named-candidate basis; Polymarket assigns only ~24% to Kenyon winning; betting markets at ~26% Reform implied; the Burnham premium is substantial and broadly documented.

Key uncertainties

• Whether Restore Britain’s 7% vote share holds on polling day or collapses toward Reform (the single biggest structural swing factor)
• Differential turnout materialisation: whether Kenyon’s 90% certainty-to-vote vs Burnham’s 74% translates into actual polling day behaviour
• How much Burnham’s campaign launch (22 May, after Survation fieldwork) has moved the numbers — no post-launch named-candidate poll yet available
• Whether national Labour unpopularity (-43 net approval for Starmer) suppresses Labour turnout beyond what the Survation filter captures
• Extent of tactical voting: whether Labour/Lib Dem/Green voters coalesce behind Burnham, or right-wing voters coalesce behind Kenyon (or Restore Britain)

Forecast variants (3/3)

Variant 0: inside_view (inside_view_v1)

Inside View: Causal Chain Analysis

The key causal mechanism here: Burnham (Labour) has a documented ~14pp personal premium over generic Labour in Makerfield (Survation source 3). This translates a constituency where Reform leads Labour by 11pp on a generic ballot into a narrow 3pp Labour lead on a named-candidate basis. The causal chain is: Burnham’s name recognition + personal brand → overcomes national Labour unpopularity (-43 Starmer net approval, source 17) → narrow Labour lead despite Reform sweeping every ward at May 2026 locals (source 5/7).

Mapping to bins:

The margin question resolves to a number in [0, 100] but by construction, values above ~25pp are implausible in a competitive multi-party race. The bins that matter are 0-12.5, 12.5-25, and then the negative scenario is also captured in 0-12.5 (since the formula uses absolute margin if we interpret strictly, but actually the question says “margin of victory” which would be a positive number by definition regardless of who wins).

Wait — re-reading the resolution criteria: “take votes for first-place candidate minus second-place, divide by total, multiply by 100.” This is always positive (or zero). So the question resolves to a positive margin regardless of which party wins.

Scenario weighting:

• Scenario A: Narrow Labour win, 1–6pp (relative_evidence_strength: HIGH). Survation 3pp lead, JL Partners corroborating, turnout-filter-adjusted polling median is ~3-5pp. This is the most evidence-backed scenario. Causal mechanism: Burnham personal premium + right-wing vote split (Restore 7%) delivers narrow Labour win. This lands in 0-12.5pp (bin_0).

• Scenario B: Moderate Labour win, 6–12pp (relative_evidence_strength: MEDIUM). Raw unweighted Survation shows 10pp gap; Burnham campaign post-launch momentum; level forecast mean ~7.5pp. Polymarket assigns 49% to Burnham winning by 6pp+. This also lands in 0-12.5pp (bin_0).

• Scenario C: Very narrow win, <2pp (relative_evidence_strength: MEDIUM). Differential turnout risk, Runcorn precedent. Still 0-12.5pp (bin_0).

• Scenario D: Reform wins (relative_evidence_strength: MEDIUM). Polymarket ~24%, betting ~26%. If Reform wins, margin would still likely be 0-12.5pp (Runcorn was 0.0%). Still bin_0.

• Burnham comfortable win 12.5-25pp: Possible if the raw unweighted 10pp gap materializes plus campaign momentum, but Survation turnout filter already accounts for differential motivation. Level forecast p95 is ~12-17pp. This could reach 12.5-25pp (bin_1) with some small probability.

Probability allocation:

Virtually all plausible outcomes fall in 0–12.5pp (bin_0):

• Labour wins by 1-12pp: the central mass (high + medium evidence scenarios)
• Reform wins by any margin: almost certainly 0-12pp range (very competitive, likely under 5pp if Reform wins)
• P(bin_1, 12.5-25pp): possible if raw Survation gap + campaign surge materializes, or if Restore Britain collapses heavily toward Labour; level forecast p95 ~12-17pp suggests maybe 10-15% chance
• P(bin_2+): near-zero given the competitive nature of this race

The question as measured is always positive. Given:

• Overwhelming evidence concentration in 0-12.5pp range
• Level forecast p5 = -4pp (which if Reform wins, still means margin ~0-5pp in absolute terms)
• Level forecast p95 = 12-17pp
• All four scenarios produce margins predominantly in 0-12.5pp

I assign approximately 82% to bin_0, 15% to bin_1, 3% spread across others.

Variant 1: outside_view (outside_view_v1)

Reference class: UK parliamentary by-elections in single-seat constituencies (recent decade). Base rate: historically, most UK by-elections are decided by margins under ~12.5 percentage points — I anchor the base probability for the 0–12.5pp bin at ~70%. Adjustment: +17 percentage points (final bin_0 = 87%) because the available constituency-level evidence strongly points to a close result rather than a landslide. Key evidentiary inputs: the only named-candidate constituency poll (Survation, fieldwork 18–22 May 2026) shows Labour (Burnham) +3pp (source 2); turnout modelling in that poll narrows the lead further (Peter Kellner notes the turnout filter cuts the lead substantially, source 14); prediction/level-forecast models and Polymarket markets place the implied median margin in the low-to-mid single digits (sources 19, 23, 26). The Research Brief explicitly marks the “Narrow Labour win — 1–6pp” scenario as the highest evidence scenario (relative_evidence_strength = high), so the distribution concentrates mass in 0–12.5pp (bin_0). I widen tails because confidence in evidence is medium (brief’s stated confidence_in_evidence = medium) and by-elections are volatile: historical precedents include extremely close northern by-elections (Runcorn & Helsby decided by 6 votes, source 21) and big upsets (Gorton & Denton, source 18), so I assign non-negligible probability to larger margins or shocks. Allocation of remaining mass: 12.5–25 (bin_1) = 8% to reflect model p95s that sometimes reach ~12–17pp (sources 26, 27) and betting-market optimism; small probabilities across bins 2–7 (total 5%) capture low-probability landslides or major collapse scenarios (e.g., Restore Britain collapse to Reform, or an unusually large Burnham surge). I follow the Interpretation’s strict reading: margin is computed from the initial official release (first minus second, divided by total votes ×100), and I treat the bins exactly as defined. Evidence cited inline: Survation named-candidate poll (source 2), turnout-filter analysis (source 14), Polymarket margin/winner markets (sources 16, 23), level forecasts (sources 19, 26, 27), and by-election precedents Runcorn & Helsby and Gorton & Denton (sources 21, 18). The distribution intentionally places most mass in bin_0 to reflect the brief’s high-evidence narrow-win scenario and the outside-view base rate; tails remain to reflect medium confidence and known upside/downside risks.

Variant 2: contrarian_v2 (contrarian_v2)

Adversarial reading

The implicit consensus in the brief is a very tight race with Labour (Burnham) winning by a low-single-digit margin. The high-evidence scenario (Survation 3pp named-candidate lead, JL Partners corroboration, level forecasts centered ~4-5pp, Polymarket concentration in the 3-6pp band) and the medium-evidence scenarios (raw unweighted 10pp gap, moderate 6-12pp win) both place the outcome comfortably inside 0-12.5pp.

Counter-evidence found (or absence)

Searched for counter-evidence; the only material push toward higher margins comes from the raw (unfiltered) Survation sample and a subset of the level-forecast p95 tails reaching 12-17pp. These are tagged medium evidence strength and are offset by the turnout filter, Reform’s local-election sweep, and the Restore Britain spoiler dynamic. No high-evidence pathway to a margin above 12.5pp is present.

Distribution implications

The high-evidence narrow-win scenario receives the bulk of probability mass inside bin_0. The medium-evidence moderate-win scenario adds limited mass still inside bin_0 with only a thin spillover into bin_1. The Reform-upset and near-tie scenarios are also contained inside bin_0. This produces an 82/17/1 split across bins 0-2 rather than the naive even spread that would result from treating all four scenarios as peers.

Consolidate

Strategy: logit_mean • Drafts: 3/3

Reviewers (3/3)

anchoring → shift_lower (low magnitude, medium confidence) anchoring_v1

The draft rationale correctly identifies that virtually all plausible outcomes fall in 0–12.5pp (bin_0), and the allocation of ~84% to bin_0 is broadly defensible. However, there is a modest optimistic anchoring issue: the draft places the modal expectation closer to the optimistic end of the Labour-win scenario by treating the Survation 3pp lead as the central anchor without adequately weighting the downward pressure from differential turnout, the post-local-elections Reform sweep, and the JL Partners mini-poll showing Labour at only 31.2% vs Reform 30.4%. The level forecast median of ~4–5pp is itself anchored on the Survation named-candidate poll, which was conducted largely BEFORE Burnham’s 22 May campaign launch — meaning it may not fully capture post-launch dynamics in either direction. More critically, the bin_1 allocation (~12.6%) is the main anchoring concern: the draft uses phrases like “raw unweighted Survation shows 10pp gap” and “Polymarket 29% on 9%+ band” to justify a non-trivial mass in bin_1, but these are optimistic inputs — the turnout filter, differential motivation, and by-election volatility all cut against reproducing anything close to the 2024 13.4pp margin. The 12.6% on bin_1 feels slightly high given the weight of evidence pointing to a tight race.

Flagged concerns

• optimistic_anchor_on_best_case_input: The draft cites the raw unweighted Survation sample (10pp gap) and Polymarket’s 9%+ band (29%) to justify ~12.6% in 12.5–25pp (bin_1). However, Peter Kellner’s analysis explicitly shows the turnout filter halves Burnham’s lead from 10pp to 3pp — the unweighted figure is not a valid standalone forecast input. Using it to anchor the bin_1 allocation overstates the probability of a comfortable Labour win.
• insufficient_downshift_for_by_election_volatility: The level forecasts (sources 19, 26, 27) are anchored on the Survation 3pp lead, which predates Burnham’s 22 May campaign launch. The JL Partners mini-poll (early June, post-launch) showed Labour 31.2% vs Reform 30.4% — a tighter picture. The draft does not explicitly discuss whether the forecast should shift slightly toward lower margins given this later (and tighter) data point.

ceiling → shift_higher (low magnitude, medium confidence) ceiling_v1

The key structural ceiling here is the three-way split of the vote: with Burnham ~43%, Kenyon ~40%, and Shepherd (Restore Britain) ~7%, the arithmetic ceiling on the winning margin is tightly constrained. Even in a Labour best-case scenario where Burnham gets, say, 50% and Kenyon remains at 40%, the margin is only 10pp. The brief explicitly states “right-wing vote totals approximately 49% (Reform 40% + Restore 7% + Con 2%) against Labour’s 43%” — meaning the ceiling on a Labour win margin is roughly 43% minus whatever the second-place finisher gets, and in most plausible scenarios that second-place finisher will be at 35-45% of the vote. The raw polling ceiling of a 10pp gap (Survation unweighted) represents the structural upper bound for a Labour win. The draft’s assignment of ~12.6% to 12.5–25pp (bin_1) appears too generous: reaching 12.5pp would require Burnham to be at ~49%+ while Kenyon stays at ~37% or lower, which is barely within the raw unweighted Survation numbers and requires the campaign to have gone very well for Labour. The draft handles this mostly correctly (84% in bin_0) but bin_1 at 12.6% is slightly high given the structural ceiling. The level forecasts’ own p95 of 12-17pp represents an extreme tail, not a median scenario, meaning only ~5% of outcomes should be above 12.5pp from those models alone.

Flagged concerns

• structural_ceiling_vote_share: Three-party vote split imposes a hard ceiling on the winning margin. With Labour at ~43%, Reform at ~40%, and Restore Britain at ~7%, even a full collapse of minor parties toward Labour cannot push the first-place margin above ~15-16pp (if Kenyon stays near 37%). The brief states the right-wing combined vote is ~49% vs Labour’s 43%, so the arithmetic ceiling on a Labour win margin is constrained to roughly the gap between Burnham’s share and Kenyon’s, with very limited room above 12.5pp. Draft assigns 12.6% to 12.5-25pp (bin_1), which overstates the probability of exceeding the structural ceiling.
• level_forecast_p95_misread: The draft rationale cites level forecast p95s of 12-17pp as supporting bin_1 mass. However, a p95 outcome by definition occurs only ~5% of the time, so citing it to justify ~12.6% in bin_1 double-counts the tail probability. The proper structural reading is that bin_1 should receive roughly 5-8% mass, not 12.6%.

math → redistribute (medium magnitude, medium confidence) math_v1

The draft’s probability map sums to 1.0 exactly and is internally consistent with the stated causal logic. However, the rationale makes a critical strategic error: it treats the margin resolution as always positive (stating “the question resolves to a positive margin regardless of which party wins”), then assigns ~24% cumulative probability to scenarios where Reform wins, yet places virtually all of that (20pp+ of probability mass) into 0–12.5pp (bin_0). This is mathematically contradictory. If Reform wins by, say, 5pp, the margin is still 5pp and falls in bin_0. But if Burnham wins by 5pp, that also goes in bin_0. The draft conflates “both outcomes fit bin_0” with “we can’t distinguish them.” In reality, the Polymarket market structure (sources 16, 23) explicitly prices Burnham 9%+, 6–9%, 3–6%, and <3% margins as distinct outcomes, implying forecasters ARE attempting to forecast the margin conditional on the winner, not just the absolute value. This suggests the question may resolve to a signed margin (Labour = positive, Reform = negative) or that the margin does encode directional information that the draft ignores.

Flagged concerns

• resolution_interpretation: Rationale claims ‘margin resolves to a positive number regardless of who wins’ but the Polymarket market (sources 16, 23) clearly prices both ‘Burnham wins by X%’ and ‘Kenyon wins by Y%’ as distinct outcomes. If the margin is always absolute-value positive, the market structure wouldn’t distinguish direction. This suggests either: (a) the question resolves to a signed margin (Lab +/Reform -), or (b) the resolution criteria language is ambiguous about whether negative values are possible. The draft doesn’t resolve this ambiguity, leading to probability mass that blurs Labour wins and Reform wins into a single bin.
• scenario_to_bins_mapping: The draft assigns ~24% probability to a Reform win (Scenario D) yet doesn’t materially adjust the bins to reflect this. If Reform wins by, say, 3pp (bin_0), and Burnham wins by 3pp (also bin_0), the draft’s 82% bin_0 allocation is correct in aggregate but masks that 24% of that mass is ‘Reform wins’ vs 58% is ‘Burnham wins.’ This is not false per se, but it obscures the directional uncertainty. If the resolution is signed (±), the draft should explicitly state this and calibrate bin_0 to reflect the two-sided distribution.
• level_forecast_reconciliation: The rationale cites level forecast p5 = -4pp (source 27) as supporting the 24% Reform-win probability. But it then says ‘if Reform wins, still means margin ~0–5pp in absolute terms,’ folding both outcomes into bin_0. This is only valid if the resolution is unsigned. If signed (which Polymarket’s structure suggests), a -4pp median would imply ~50% of the p5–p95 interval is negative (Reform leads), not a minority. The draft should clarify: does the level forecast’s -4pp floor support a ~24% Reform upset, or does the consensus forecast assume Burnham always wins?
• bin_allocation_consistency: The draft assigns 0.127 to 12.5–25pp (bin_1), citing level forecast p95 in the 12–17pp range (sources 26, 27) as support. This is internally consistent (p95 at ~12–17pp does suggest ~10–15% in the 12.5–25pp bin), but the rationale doesn’t explicitly connect the p95 tail to this bin allocation; it relies on verbal cues rather than showing the quantile-to-bin mapping. A clearer audit: ‘p95 ~13pp ⇒ P(X > 12.5pp) ≈ 5–10%’ would strengthen it.

Reconcile: partial_shift

Reviews consumed: 3

The draft places ~84% on 0-12.5pp (bin_0) and ~12.6% on bin_1, which is well-supported by the evidence. The anchoring reviewer flagged that the draft slightly overweights bin_1 by leaning on the raw unweighted Survation 10pp gap (which Kellner explicitly says is reduced by the turnout filter), and suggested shifting lower into bin_0 — I partially accepted this. The ceiling reviewer independently flagged the same issue from a structural vote-share angle: with Lab ~43% vs Reform ~40% and a right-wing split, the arithmetic ceiling on a Labour margin makes >12.5pp a tail event (~5-8%), not 12.6%; I accepted this convergent signal and trimmed bin_1 modestly. The math reviewer raised a resolution-interpretation concern (signed vs unsigned margin) and pushed for MORE mass in bin_1; I reject this because the resolution criteria explicitly state “votes for first minus votes for second / total × 100” — this is structurally non-negative, and Polymarket’s directional pricing reflects market segmentation, not a signed resolution. With two reviewers converging on “shift slightly lower / tighter to bin_0” and the math reviewer’s directional concern rejected on resolution-reading grounds, I made a small partial shift: bin_0 to ~0.87, bin_1 to ~0.10, with tails preserved for by-election volatility (Runcorn/Gorton precedents).

Rejected critiques

• [math] Resolution may be signed (Lab + / Reform -), so Reform-win mass shouldn’t collapse into bin_0; suggested redistributing ~16% to bin_1 and more to higher bins.

  • Why rejected: The resolution criteria explicitly compute margin as (first-place votes minus second-place votes) / total × 100 — structurally non-negative regardless of which party wins. The worked example confirms this. Polymarket’s separate Burnham-vs-Kenyon margin buckets reflect market design (traders care about who wins), not a signed resolution. Both a 3pp Burnham win and a 3pp Kenyon win correctly resolve to ~3.0 in bin_0. The math reviewer’s directional push toward higher bins is therefore not supported by the resolution rules.

• [math] Bin_1 should be increased to ~16% to reflect level-forecast p95 of 12-17pp and resolution ambiguity.

  • Why rejected: The ceiling reviewer independently showed the structural arithmetic caps a Labour-win margin near the raw 10pp gap, making >12.5pp a true p95 tail (~5-10%). Doubling bin_1 would double-count optimistic tail inputs that are already reflected in the level forecast’s p95.

Final Distribution

Evidence Sources (36)

1. PollCheck Makerfield By-Election page — The 2024 Makerfield general election result: Josh Simons (Labour) 45.2% (18,202 votes), Robert Kenyon (Reform UK) 31.8% (12,803 votes), total votes 40,263, giving a margin of 13.4 percentage points.
2. Makerfield by-election: Burnham premium translates to small lead – Survation — Survation constituency poll (fieldwork 18–22 May 2026, N=504 adults, 369 likely voters), published 26 May 2026: Burnham (Labour) 43%, Kenyon (Reform UK) 40%, Shepherd (Restore Britain) 7%, Austin (Lib Dem) 4%, Green 3%, Winstanley (Con) 2% — giving Labour a 3-point lead. Margin of error ±5.4pp.
3. Survation Makerfield by-election poll — On a generic Westminster ballot (no candidate names), Reform UK leads Labour by 11 points in Makerfield — the “Burnham premium” accounts for a 14-point swing from generic to named-candidate polling.
4. Survation Makerfield initial estimate PDF — Survation modelled estimate (15 May 2026): Without Burnham, Reform would win Makerfield ~53% to Labour 27%. With Burnham as candidate, it tightens to Labour 45% vs Reform 42%, with Labour winning in ~67% of simulated outcomes.
5. The Independent – Burnham beats Reform by small margin in first Makerfield by-election poll — Reform UK won every ward in Makerfield constituency at the May 2026 local elections, showing Reform’s strong local position before the by-election.
6. Reform and Restore Britain lock horns in Makerfield byelection — Restore Britain (led by Rupert Lowe, expelled from Reform UK) is standing Rebecca Shepherd in Makerfield, polling at 7% as of late May 2026. Reform insiders accuse Lowe of splitting the right-wing vote and potentially handing Burnham the seat, with Restore Britain on 7% in the Survation poll — enough to influence the margin.
7. Liverpool Echo Makerfield by-election poll coverage — At the 2026 Wigan local elections (held before the by-election), Reform UK swept the board winning 24 of 25 available seats, signaling strong local support for Reform in the Makerfield area.
8. Will Restore stop Reform winning in Makerfield? - New Statesman — Restore Britain (Rupert Lowe’s party, described as far-right “deport them all”) is polling 7% in Makerfield, creating a potential spoiler effect that could squeeze Reform’s margin — but Labour doorstep reports say voters are mentioning Restore as often as Reform.
9. 1965 Leyton by-election - Wikipedia — The 1965 Leyton by-election (the last time a by-election was called specifically to get a minister into Parliament) resulted in Patrick Gordon Walker (Labour’s Foreign Secretary) LOSING to the Conservative candidate — a cautionary historical parallel for the Burnham situation.
10. Odds cut on Labour to win Makerfield by-election as Reform lose ground — Bookmaker odds as of late May 2026: Labour (Burnham) at 7/20 (71% implied win probability), Reform UK at 10/3 (22%), Restore Britain at 14/1, per Oddschecker data.
11. Andy Burnham facing fight on two flanks in ‘prime minister by-election’ — Convergent poll (10,000-voter analysis, reported 1 June 2026): Burnham starts 22 points behind Reform’s Kenyon on local election benchmarks; even reclaiming all Labour defectors from both left and right brings him only level with Reform at 37%; a 3pp lead for Burnham is within the margin of error.
12. Lines.com Makerfield by-election margin of victory — Prediction market (Lines.com, ~May 29 2026): The Makerfield by-election margin market sits at 19.5% probability for Burnham winning by 9+ percentage points, reflecting market uncertainty about the size of any Labour lead.
13. Makerfield By-Election: Right-Wing Vote Split Could Hand Victory to Labour’s Andy Burnham — Sir John Curtice warned that the Restore Britain party’s presence (7% in the poll) splits the right-wing vote, potentially delivering a Labour victory despite a majority of voters opposing Labour; right-wing vote totals ~49% combined (Reform 40% + Restore 7% + Con 2%) vs Labour 43%.
14. Left Foot Forward - Andy Burnham beats Reform in first Makerfield by-election poll — Peter Kellner analysis: Survation’s turnout filter halves Burnham’s lead because only 74% of Burnham’s supporters say they’re certain to vote vs 90% of Kenyon’s; on the raw unweighted Survation sample, Burnham is at 47% and Kenyon at 37%, a 10-point gap.
15. Our survey says Tories don’t do tactical voting – that might matter in Makerfield — Conservative Home article (2 June 2026) noted the most recent available poll (Survation) has Labour 43%, Reform 40%, Restore 7%, with the race decided by just 3 percentage points; Conservative candidate at 2% which — if shifted to Reform — could tighten the race further.
16. Polymarket Makerfield by-election margin of victory — Polymarket (as of ~June 3 2026, liquidity $51,100): “Will Andy Burnham win by 9%+ margin?” YES at 35.5%; “Will he win by 6-9%?” at 15.5%; “Will he win by under 3%?” at 11.5%; “Will Reform’s Kenyon win by 6%+?” at 4.45% — implying expected Burnham win in the 3-6% range as the modal outcome.
17. Can Andy Burnham save the Labour Party? - Opinium — Opinium (June 1 2026): National Labour voting intention stands at just 20%, behind Reform UK at 27%. Starmer’s net approval is -43. Only 56% of 2024 Labour voters say they will vote Labour again at the next election.
18. UK Parliament election results - Gorton and Denton — Gorton and Denton by-election (26 Feb 2026): Green Party gained the seat from Labour with a majority of 4,402 votes (12.0% margin) on a 47.5% turnout — Labour, which had won with 50%+ in 2024, came third behind Reform UK in a multi-party contest.
19. Level forecast model — Level forecast modelling the margin of victory: anchored on the 3pp polling lead (Survation, late May 2026), with 0.5pp daily volatility and slight positive drift as Burnham’s campaign progresses (16 days to election): projected median margin ~4.6pp, p5 at 1.3pp, p95 at 7.9pp — consistent with a tight race unlikely to reproduce the 2024 13.4pp margin.
20. Polymarket – Makerfield by-election Winner — Polymarket (3 June 2026) Makerfield winner market: Andy Burnham (Labour) at 73%, Robert Kenyon (Reform UK) at 26%, with $2M volume.
21. UK Parliament - Runcorn and Helsby by-election result — Runcorn and Helsby by-election (1 May 2025): Reform UK gained the seat from Labour by just 6 votes, a margin of 0.0%, on 46.2% turnout — illustrating that tight by-elections in Labour-held northern English seats can produce near-zero margins.
22. Convergent Research / The National coverage — UK by-elections typically have lower turnout than general elections, which historically helps the party with the most motivated/enthusiastic base. Reform UK’s more reliable turnout (vs Labour’s 38% abstention rate among local election supporters) could narrow the margin further relative to polls, as differential turnout is a key swing factor.
23. Polymarket – Makerfield by-election: Margin of Victory (updated via Exa) — Polymarket margin-of-victory market (most recent data from ~3 June 2026): Burnham 9%+ = 29%, Burnham 6–9% = 20%, Burnham 3–6% = 17%, Burnham <3% = 10%, Kenyon <3% = 16%, Kenyon 3–6% = 4%, Kenyon 6%+ = 4%, Other = 7%. Implied median margin is Burnham winning by ~5–6 percentage points.
24. Winners and losers - First magazine for councillors — In the first quarter of 2026 (through ~April 2026), 68% of local by-election defences were lost by the defending party — a record rate; since the May 2025 local elections, Labour has lost almost eight in ten of its local by-election defences.
25. Polymarket Makerfield by-election margin of victory market — Polymarket margin-of-victory market breakdown suggests the modal scenario is a narrow Burnham win (3-6pp range), with meaningful probability of both a larger win (9%+ at 35.5%) and a Reform upset (Kenyon winning at ~20% combined). The Polymarket-implied median margin for a Labour win is approximately 5-7 percentage points.
26. level_forecast tool (based on Survation poll 43% vs 40%) — Level forecast from current Survation poll (Labour +3pp), assuming ~1.5pp/day volatility in campaign swings and a small positive drift (Burnham campaign effect growing): projected margin at election day has mean 7.5pp, p25 3.6pp, p75 11.4pp, p5 -2.1pp, p95 17.1pp — suggesting a wide distribution centered around a moderate Labour win.
27. Level forecast model - own calculation — Level forecast for Makerfield by-election margin: anchored at +3pp Labour lead (Survation, May 2026), with a single-period σ=5pp reflecting polling uncertainty and by-election volatility; projected median +4pp (Lab wins), p5=-4.2pp (Reform wins), p25=+0.6pp, p75=+7.4pp, p90=+10.4pp, p95=+12.2pp.
28. Andy Burnham launches Makerfield by-election campaign — Andy Burnham launched his Makerfield campaign on 22 May 2026, branding it a ‘change election’ and urging voters to help ‘reshape Labour,’ positioning himself explicitly as an anti-Starmer candidate — which creates a unique dynamic where Labour voters may be motivated by a desire to change the Labour leadership.
29. Makerfield parliamentary by-election - 18th June 2026 (Wigan Council) — The Makerfield parliamentary by-election is officially scheduled for Thursday 18 June 2026, confirmed by Wigan Council (the returning authority).
30. 2026 Makerfield by-election - Wikipedia — Wikipedia states the 2026 Makerfield by-election is due to be held on 18 June 2026, triggered by the resignation of MP Josh Simons.
31. Makerfield Parliamentary by-election - Voter information (Wigan Council) — Wigan Council’s voter information page for the Makerfield by-election confirms polling day is Thursday 18 June 2026 and provides key voter deadlines, with no indication of any delay or postponement.
32. Election result for Makerfield (Constituency) - MPs and Lords - UK Parliament — In the 2024 general election, the Makerfield constituency had a registered electorate of 76,641 and a turnout of 52.5%, with 40,263 total votes cast (Labour 18,202; Reform UK 12,803; Conservative 4,379; Lib Dem 2,735; Green 1,776; English Democrats 368).
33. Survation – Makerfield by-election poll — The Survation poll (fieldwork 18–22 May 2026, n=504 total / 369 likely voters) — the only named-candidate constituency poll published before June 3 — showed Burnham (Lab) 43%, Kenyon (Reform) 40%, Shepherd (Restore Britain) 7%, Lib Dems 4%, Greens 3%, Conservatives 2%. The 3pp Labour lead was described as a “Burnham effect,” as Reform led Labour by 11pp on the generic Westminster ballot in the same poll. The fieldwork was conducted largely PRIOR to Burnham’s 22 May campaign launch.
34. The Guardian – Cost of living and high streets among top priorities for Makerfield voters, poll shows — A JL Partners focus-group/mini-poll of 112 Makerfield residents commissioned by 38 Degrees (fieldwork approx. early June 2026, ~2 weeks before 18 June polling day) showed Labour 31.2%, Reform 30.4%, Greens and Conservatives both 10.7%, Lib Dems 3.6%, others 13.4%. This is not a standard named-candidate poll (open-ended questions, small n), but it corroborates a narrow Labour lead post-campaign-launch.
35. Makerfield (constituency) - Wikipedia — Wikipedia lists the Makerfield constituency electorate as 76,517 (2023 figure), confirming the ~76,600 registered voter base for the 2026 by-election context.
36. The London Economic – Odds cut on Labour to win Makerfield by-election as Reform lose ground — Betting markets (via Oddschecker, as of ~28 May 2026) showed Labour odds at 7/20 (implied 74% chance of winning), up 11 points, with Reform at 10/3 (implied 22%), down 8 points — a significant shift toward Labour vs. the close polling numbers, potentially reflecting the Restore Britain vote-split effect.

wooley_rhino • run a31b0e92 • $2.83 • 6m2s

Extracted Percentile Values: {0.1: 1.1, 1.0: 1.3, 5.0: 1.8, 10.0: 2.5, 15.0: 3.0, 20.0: 3.5, 25.0: 4.0, 30.0: 4.5, 35.0: 5.0, 40.0: 5.5, 45.0: 6.0, 46.0: 6.1, 47.0: 6.3, 48.0: 6.4, 49.0: 6.6, 50.0: 6.8, 51.0: 7.0, 52.0: 7.2, 53.0: 7.4, 54.0: 7.6, 55.0: 7.8, 60.0: 8.5, 65.0: 9.5, 70.0: 10.5, 75.0: 12.0, 80.0: 14.0, 85.0: 16.0, 90.0: 19.0, 95.0: 24.0, 99.0: 35.0, 99.9: 50.0}

Model’s Answer: Looking at this Makerfield by-election margin of victory forecast, I need to carefully weigh the available evidence.

Key Data Points

2024 baseline: Labour won with a 13.4 percentage point margin (45.2% vs 31.8%)

2026 local elections (Wigan area): Reform led Labour 47% to 25% - a dramatic swing

Survation constituency poll (May 18-22, 2026): Burnham (Labour) 43%, Kenyon (Reform) 40% - a 3pp Labour lead with MoE ±5.4pp

Betting/prediction markets: Labour/Burnham at ~74-75% probability to win; Reform/Kenyon at ~24-26%

Analysis

The race has clearly tightened dramatically from 2024. The key dynamics are:

• Burnham’s personal popularity provides a “Burnham premium” that converts what would otherwise be a Reform lead into a narrow Labour edge
• Reform has surged in the local area based on 2026 local elections
• The poll shows only a 3pp lead with high uncertainty (sample size N=504)
• This is a high-stakes “prime minister by-election” which could affect turnout in unpredictable ways

Low outcome scenario (1-4 pp): If Reform’s local election momentum carries through, turnout is low (favoring motivated protest voters), or late-breaking issues hurt Burnham, we could see a knife-edge result. The constituency poll’s margin of error explicitly allows for a Reform lead. Likelihood: ~25%

High outcome scenario (15+ pp): If Burnham’s campaign successfully consolidates anti-Reform tactical voting from Conservative/Lib Dem voters, and Labour loyalists turn out in force for the leadership stakes, margins could approach or exceed 2024 levels. Likelihood: ~15%

Key Assumptions to Scrutinize

1. Polling accuracy: Single constituency poll with large MoE could be significantly off (~35% chance of major miss)
2. Turnout patterns: By-election turnout is unpredictable and could favor either side (~30% chance of unexpected pattern)
3. Late campaign shifts: Two weeks remain for events to shift the race (~25% chance of material shift)

Given the tight polling, high uncertainty, and structural volatility in UK elections (41% of 2026 local wards won by <5pp), I’m centering my distribution around 6-7 pp with significant weight on both tails.

Percentile .1: 1.1 Percentile 1: 1.3 Percentile 5: 1.8 Percentile 10: 2.5 Percentile 15: 3.0 Percentile 20: 3.5 Percentile 25: 4.0 Percentile 30: 4.5 Percentile 35: 5.0 Percentile 40: 5.5 Percentile 45: 6.0 Percentile 46: 6.1 Percentile 47: 6.3 Percentile 48: 6.4 Percentile 49: 6.6 Percentile 50: 6.8 Percentile 51: 7.0 Percentile 52: 7.2 Percentile 53: 7.4 Percentile 54: 7.6 Percentile 55: 7.8 Percentile 60: 8.5 Percentile 65: 9.5 Percentile 70: 10.5 Percentile 75: 12.0 Percentile 80: 14.0 Percentile 85: 16.0 Percentile 90: 19.0 Percentile 95: 24.0 Percentile 99: 35.0 Percentile 99.9: 50.0

Makerfield remains anchored to a Labour structural baseline that has weakened through boundary change and partisan realignment.

By-election mechanics—lower turnout, tactical incentives, and local salience—limit how national swings convert into a margin.

A high-profile local candidate creates a substantial personal vote that shifts the contest away from baseline party shares.

A consolidated challenger on the right has absorbed protest and realignment energy but is exposed to vote splitting and reputational shocks.

If the challenger consolidates right-wing support and minor parties hold, the resolved margin compresses to low single digits.

If the high-profile candidate pulls cross-party support while opponents fragment, the margin expands to mid-single digits for the winner.

Unresolved sensitivities include turnout composition, late movement among Restore/Green/LD/Conservative voters, and tactical transfers.

Residual uncertainty is driven by late campaign events, differential mobilization, and whether local voting patterns replicate recent municipal results.

Forecast (median margin of victory): 4.2013 percentage points

• Polling says a close race. Multiple constituency polls put Labour only slightly ahead (roughly low single digits), so a small top-two gap is most likely.
• Local election results pull toward Reform. In May 2026, Reform won all the Makerfield wards decisively, with Labour much weaker there—this limits how large a Labour win can be.
• The big swing factor is Andy Burnham’s personal vote. If Burnham adds a lot compared with “generic Labour,” Labour can win by several points; if that boost is smaller than expected, Reform can narrow or even take the lead.
• Right-wing vote splitting adds uncertainty—but also shapes the gap. How the right vote parcels out between Reform, Conservatives, and Restore Britain affects whether Labour’s lead translates into a margin, or whether Reform consolidates.
• Recent by-elections suggest Labour often underperforms expectations. This keeps the outcome slightly lower than what a straightforward “Labour + national environment” story would imply, reinforcing a modest margin rather than a blowout.