The overhang trap: How National could lose power to its own coalition partners
- Grant McLachlan
- 12 minutes ago
- 30 min read
Seven predictions. One structural argument. And a documented set of conflicts of interest that the Wellington Press Gallery — distracted by a leadership crisis, by the Greens' sewage problems, by Chlöe Swarbrick's uncertain electorate — has been too busy to examine closely.
National is tracking to win more electorate seats than its party vote entitles it to. Winston Peters will hold the balance of power. He has done this before. He will do it again. And the machinery that has delivered this outcome — the funding networks, the polling house effects, and the factional warfare over who owns land — has been operating in plain sight throughout this parliamentary term.
Contents
The diversionary landscape: What the Press Gallery is watching instead
The 2020 precedent: Blue-rosette country and the sympathy vote
The polling picture: The Curia timing question
The seat arithmetic: Sainte-Laguë and the overhang
The final month: Why the polls will lie
The redrawn map: What the 2025 boundary review means
The electorates at risk: National's vulnerability matrix
The list that vanishes: Who does not come back
The caucus war: Factions, the seventeen, and a leadership in free fall
The property-industrial complex: Who funds the party and what they want
Tamaki: Van Velden goes, Christmas arrives, and it is not a gift seat
The seven predictions
These are the conclusions of the analysis below. Each is supported by the evidence that follows. They will be tested on the night of 7 November 2026.
National retains approximately 42–44 electorates on a party vote of 30–32 per cent, creating an overhang of 6–8 seats. Parliament expands to 126–130 seats. The right bloc cannot govern without New Zealand First.
National wins zero Wellington general electorates. This was true before the boundary changes. The new map confirms it structurally.
National's list seat count is zero. Nicola Willis, as Finance Minister and a list MP, faces the same compression as Maureen Pugh and Paul Goldsmith — the entire list is eliminated at 30 per cent with 43 electorates.
Chris Bishop loses Hutt South. The boundary shift left, his 1,332-vote majority, and the national swing combine to make this Labour's most achievable pickup from National's 2023 haul.
NZF polls 12–15 per cent, wins 14–17 seats, and holds the balance of power. Winston Peters will determine who governs. The 2017 precedent is the operating manual.
James Christmas wins the Tamaki electorate for ACT, not by gift but by contest against National's Mahesh Muralidhar and Labour's Max Harris.
Labour wins Tukituki. The Wattie's and McCain closures, the Cyclone Gabrielle recovery failure, Taine Randell's three-way vote split, and a national swing of 8 per cent make this the seat where economic devastation most directly translates into electoral consequence.
The diversionary landscape: What the Press Gallery is watching instead
The leadership drama is not the election. It is a distraction — and one that is being systematically amplified by a network of interests that benefit from its continuation.
The disgruntled caucus members, as commentator Liam Hehir characterised them, are non-ministers whose careers have stalled. The pressure they are generating isolates Luxon and his Christian faction — Simeon Brown, Chris Penk, the values-conservative wing that has been most publicly loyal — while creating the impression that the dirty politics faction (Mark Mitchell, Cameron Brewer, the northern Auckland property network recently rewarded by the April reshuffle) represents the pragmatic alternative waiting in the wings. ACT, New Zealand First, and the Taxpayers' Union all benefit directly from a National Party that appears in chaos and haemorrhages party votes to the right. Their polling is elevated precisely by voter dissatisfaction with National's governance; any restoration of confidence in National as a governing party would reduce it.
The profiles of some Press Gallery journalists who have been most active in breaking the leadership stories are consistent with long-term relationships within the factions that benefit most from the Luxon story's continuation. That is not an accusation of bad faith. It is a structural observation about who has access, whose calls get returned, and whose framing of the story — that this is a crisis for Luxon personally rather than a structural crisis in the party's relationship with its funding base — is the one that dominates coverage.
The Curia poll, timed — as documented below — to enter the media environment while mainstream polling is active, consistently shows National performing 2–3 points better than other firms. This dampens any impulse among right-bloc voters to return to National from ACT or NZF, sustaining the elevated minor party polling that generates the overhang. The leadership story and the polling house effect work together toward the same structural outcome: a National Party that holds many more electorates than its party vote entitles it to, and that is therefore more dependent on its coalition partners than at any point in the MMP era.
Three additional narratives complete the diversionary picture.
Labour appears stronger to the Wellington Press Gallery than the structural analysis warrants. The party vote polling shows Labour ahead and the left bloc approaching a majority. But this reading ignores the electorate-level dynamics documented below: party vote and electorate vote are two different signals, and National can hold electorates in the north while Labour wins the party vote there. If the left bloc forms a government on the back of a strong party vote, it faces an expanded parliament — and every additional National electorate seat held above proportional entitlement raises the majority threshold further.
The Greens are under pressure in their own Wellington strongholds. The Wellington Water sewage crisis — years of overflows, aging infrastructure, deferred investment under the Green-led Wellington City Council — has damaged the Green brand in the capital in ways that the party has not fully absorbed. Tamatha Paul holds Wellington North (formerly Wellington Central). The electorate has shifted right with the addition of Khandallah, Ngaio, and Wadestown — previously Ohariu territory — and the sewage narrative reinforces a damaging argument: that Green governance in Wellington has been ideological rather than operational. Wellington Bays (Julie-Anne Genter, formerly Rongotai) is more protected by its new left-wing boundary additions from Brooklyn and Mt Cook. But if the Green brand suffers across Wellington, the seat is not immune.
Chlöe Swarbrick's Auckland Central electorate is conventionally regarded as secure after the 2025 boundary change added Grey Lynn, Arch Hill, and Western Springs. The conventional analysis concludes this "all but guarantees" her re-election. That reading underestimates a quieter demographic shift. Auckland Central has been changing: newer residents, including Asian-Pacific communities and urban professionals whose social conservatism on specific issues does not align with Swarbrick's progressive brand. In an electorate where National won 40 per cent or higher of party votes in some mesh blocks in 2023, and where the inner-city demographic continues to evolve, the seat is not as fireproof as the boundary numbers imply. If Swarbrick is unseated, the Greens hold one fewer electorate. If both Wellington North and Auckland Central are lost, the Greens have no electorate seats and must clear 5 per cent on the party vote alone.
The diversionary effect of these three narratives — Luxon's leadership drama, the Green vulnerability story, Labour's apparently strong recovery — focuses media attention precisely away from the structural question at the heart of this election. While every political journalist in Wellington is counting caucus numbers, the electorate-level arithmetic of the overhang is being built, quietly, in Rodney, Northland, Tukituki, and thirty other seats where tradies who gave their party vote to ACT will still walk in and tick the National candidate on election day.
The 2020 precedent: Blue-rosette country and the sympathy vote
The analytical core of the 2026 overhang prediction has an established New Zealand precedent, and two of the MPs most explicitly associated with it now sit in Cabinet.
In the 2020 general election, Labour won 49.1 per cent of the party vote — the highest result any party has achieved under MMP. In many electorates, Labour's party vote exceeded National's. Yet National's candidates still won the electorate vote seat after seat, returning to parliament even as their party was being routed nationally.
Chris Penk won Helensville — now Kaipara ki Mahurangi — in 2020. Mark Mitchell won Rodney. Understanding why requires understanding what these electorates actually are — and it has considerably less to do with individual campaigning skill than with the political geography of Auckland's northern fringe.
Rodney and Kaipara ki Mahurangi are National heartland in the most literal sense. Gordon Coates — New Zealand's first National Prime Minister — represented the Kaipara area from 1911. Murray McCully held East Coast Bays for 27 years. These electorates are where the tradespeople, small business owners, rural contractors, and urban-fringe homeowners who form the backbone of the National Party have voted blue for three and four generations. The political culture is entrenched. Anyone with a blue rosette on the right day can hold these seats. Keep the tradies happy. Keep the motorway extensions on the agenda. Keep the party vote conversation quiet at the barbecue. That is the model.
There is also a phenomenon the 2020 result illustrated clearly: social pressure as an electoral distorter. In an election dominated by Jacinda Ardern's personal popularity and Labour's pandemic management, voting Labour became the social norm in ways that made it difficult for many voters in these communities to publicly admit an alternative position. But the electorate vote offered a private escape. In the ballot booth, a voter who had — reluctantly, and with some embarrassment — given their party vote to Labour could still give the electorate vote to the familiar local National candidate and maintain their social standing in a community where the blue rosette is the expected uniform. The result: Labour wins the party vote in suburbs from Kumeu to Warkworth. National's candidate wins the electorate. The overhang begins.
In 2026, the same dynamic operates in reverse. Right-of-centre voters genuinely angry at the government — about the fuel crisis, electricity costs, the cost of living — are directing their party vote to ACT or NZF as a protest signal. But they are not abandoning their local National candidate, whom they have voted for for years. The party vote goes one way. The electorate vote goes another. The gap between them is the overhang.
In 2020, the wasted vote amplified the effect. New Zealand First fell below 5 per cent with 2.6 per cent of votes distributed to no seats, effectively boosting Labour's proportional share and delivering its majority. In 2026, the wasted votes come from the small parties — The Opportunities Party at 1–2 per cent and others at 2–3 per cent combined — modestly inflating the effective shares of qualifying parties, including NZF, and making Peters more powerful than the raw polling numbers suggest.
Right-of-centre voters who divert their party vote to ACT or NZF keep those parties above the 5 per cent threshold while simultaneously inflating National's electorate count beyond its proportional entitlement. The result is an overhang, an expanded parliament, and a coalition formation problem structurally worse for the right than the headline numbers imply.
The polling picture: The Curia timing question
Last night's 1 News-Verian poll arrived, in one commentator's phrase, "like a whack to the solar plexus." National is at 29.7 per cent — rounded to 30, but a fraction less in the actual numbers. It is the party's lowest result since Luxon became leader at the end of 2021. Labour is up five points. The left bloc can form a government without New Zealand First for the first time in months.
Luxon's preferred prime minister rating is 16 per cent. One in six New Zealanders picks him as their choice. His likely successor, Housing Minister Chris Bishop, is at 2 per cent — up one point, and the only upward movement in the National caucus outside Winston Peters, who is on 12 per cent and rising.
The table below covers recent polling with two columns not previously assembled in New Zealand media coverage: the dates on which the Taxpayers' Union Curia poll entered the field, set against the dates on which 1 News-Verian and RNZ-Reid were simultaneously conducting their own surveys.
I have raised concerns previously — including in "The Bullshit Economy" — about coordination patterns in New Zealand's polling ecosystem. The pattern visible in the table is consistent with what I have documented: the Curia poll, run by David Farrar in his capacity as National's own pollster under a Taxpayers' Union contract, is repeatedly released into the media environment while mainstream polling firms are actively in the field or about to enter it.
The implication is straightforward. If a Curia result showing National performing 2–3 points stronger than reality enters the news cycle while 1 News or RNZ are actively polling, it shapes the information environment in which respondents answer questions about voting intention. A voter who sees National polling at 33 per cent in a Curia release while they are being polled by RNZ — who will find 31 per cent — may be more likely to report National support. The spiral of silence and bandwagon effects are well-documented in the political polling literature.
Note also the direction of the Curia house effect. The poll consistently shows National 2–3 points higher, ACT 1–2 points lower, and NZF 1–2 points lower than other firms. This is precisely the distribution that serves the overhang outcome: a higher National party vote reading dampens enthusiasm for the split-ticket strategy among right-bloc voters, while lower ACT and NZF readings understate those parties' actual standing. The Curia result, in short, is more useful to the property-preservation faction of the National Party than to anyone else.
TABLE 1: OPINION POLLS — PARTY VOTE (%) WITH CURIA FIELD-DATE TIMING
Pollster | Field dates | Published | NAT | LAB | GRN | ACT | NZF | Lead | Timing note |
1 News–Verian | 19 Apr 2026 | 20 Apr 2026 | 29.7 | up5 | — | — | — | LAB majority | ⚡ Latest — left bloc can govern without NZF |
Talbot Mills | Apr 2026 | Apr 2026 | 29 | 36 | 7 | 8 | 15 | LAB +7 |
|
The Post/Freshwater | 6–12 Feb 2026 | 13 Feb 2026 | 30 | 37 | 10 | 6 | 11 | LAB +7 |
|
1 News–Verian | 7–11 Feb 2026 | 12 Feb 2026 | 34 | 32 | 11 | 9 | 10 | NAT +2 |
|
TU–Curia (Farrar) | 1–3 Feb 2026 | 4 Feb 2026 | 31.3 | 34.1 | 10.3 | 6.7 | 10.5 | LAB +2.8 | ⚠ Result published 3 days before 1News entered field |
RNZ–Reid Research | 15–22 Jan 2026 | 23 Jan 2026 | 31.9 | 35 | 9.6 | 7.6 | 9.8 | LAB +3.1 |
|
TU–Curia (Farrar) | 14–18 Jan 2026 | 19 Jan 2026 | 31.5 | 34.4 | 7.7 | 7 | 11.9 | LAB +2.9 | ⚠ Field dates overlap RNZ–Reid (15–22 Jan) — simultaneously in field |
1 News–Verian | 29 Nov–3 Dec 2025 | 4 Dec 2025 | 36 | 35 | 7 | 10 | 9 | NAT +1 |
|
TU–Curia (Farrar) | 3–7 Dec 2025 | 8 Dec 2025 | 30 | 31.6 | 10.8 | 8.9 | 8.1 | LAB +1.6 | ⚠ Released while 1News Dec result was live in media cycle |
RNZ–Reid Research | 4–12 Sep 2025 | 13 Sep 2025 | 32.5 | 34.3 | 10.9 | 7.2 | 8.7 | LAB +1.8 |
|
TU–Curia (Farrar) | 31 Aug–2 Sep 2025 | 3 Sep 2025 | 33.1 | 33.8 | 10.7 | 6.7 | 8.1 | LAB +0.7 | ⚠ Published day before RNZ entered field (4 Sep) |
1 News–Verian | 2–6 Aug 2025 | 7 Aug 2025 | 34 | 33 | 10 | 8 | 9 | NAT +1 |
|
TU–Curia (Farrar) | 3–5 Aug 2025 | 6 Aug 2025 | 31.8 | 33.6 | 9.8 | 8.6 | 7.8 | LAB +1.8 | ⚠ In field simultaneously with 1News (2–6 Aug) |
1 News–Verian | 29 Mar–2 Apr 2025 | 3 Apr 2025 | 36 | 32 | 10 | 9 | 7 | NAT +4 |
|
TU–Curia (Farrar) | 29 Mar–1 Apr 2025 | 2 Apr 2025 | 33.5 | 29.8 | 11 | 10 | 7.4 | NAT +3.7 | ⚠ Identical field dates to 1News — both in field 29 Mar–1 Apr |
RNZ–Reid Research | 21–27 Mar 2025 | 28 Mar 2025 | 32.9 | 32.3 | 10 | 9.4 | 7.2 | NAT +0.6 |
|
5-poll average | Jan–Feb 2026 | — | 31.4 | 35.1 | 10.3 | 7.6 | 10.5 | LAB +3.7 |
|
2023 election result | 14 Oct 2023 | 14 Oct 2023 | 38.08 | 26.92 | 11.61 | 8.64 | 6.09 | NAT +11.2 |
|
⚠ Yellow rows = TU–Curia polls. Note overlapping field dates or pre-release timing relative to mainstream firms. Curia consistently shows NAT 2–3 pts higher than other firms. Sources: Wikipedia polling page; individual pollster releases.
The seat arithmetic: Sainte-Laguë and the overhang
New Zealand allocates its 120 parliamentary seats using the Sainte-Laguë method — dividing each qualifying party's vote by successive odd numbers (1, 3, 5, 7...) and allocating seats in order of the highest resulting quotients. A party must clear 5 per cent of the party vote or win a general electorate. Te Pati Maori, holding Maori electorates, coat-tails in regardless of party vote, as demonstrated in 2023 when it held six Maori seats on 3.08 per cent.
National at 30 per cent is entitled to roughly 36 seats from the base 120. If National retains 42–44 electorates on the back of the split-ticket dynamic documented above, the overhang is 6–8 seats. Parliament expands seat-for-seat. A 128-seat parliament requires 65 seats for a majority. National and ACT combined at current polling hold approximately 54 seats. They cannot govern without New Zealand First.
The overhang that Maori voters delivered organically for Te Pati Maori in 2023 is precisely what right-of-centre voters could deliver for National in 2026 through split-ticket voting. ACT and NZF supporters giving their electorate vote to the National candidate while directing their party vote elsewhere. The theory has been discussed since MMP's introduction. 2026 may be the first election where it is tested at scale on the right.
TABLE 2: SAINTE-LAGUË SEAT PROJECTION — BASE CASE (APRIL 2026 POLLING)
Party | Party vote | Proportional seats | Electorates won | Overhang | Total seats | List seats | Bloc |
National | 30% | 36 | 44 | +8 OH | 44 | 0 | Right |
Labour | 36% | 43 | 9 | — | 43 | 34 | Left |
NZ First | 14% | 17 | 0 | — | 17 | 17 | Kingmaker |
Greens | 8% | 10 | 0–1 | — | 10 | 9–10 | Left |
ACT | 8% | 10 | 2 | — | 10 | 8 | Right |
Te Pati Maori | 3% | 4 | 6 | +2 OH | 6 | 0 | Left |
Parliament total | — | 120 | 62 | +10 | 130 | 68 | — |
Majority required | — | — | — | — | 66 | — | — |
Sainte-Laguë method per Electoral Act 1993. OH = overhang. Parliament total = 130 seats. Majority = 66. Right bloc (NAT+ACT+NZF) = ~71 seats — majority, but only with Peters. Source: Electoral Act 1993.
The final month: Why the polls will lie
The seat projection above uses April 2026 polling averages. The result on election night will look different — and the differences are structural and predictable, not random. Three recurring patterns in New Zealand's MMP era materially affect the final result relative to what the polls will show in the weeks before polling day.
ACT's polling implodes in the final month of every serious campaign. It is a pattern so consistent it should be treated as a structural feature of the party rather than a sequence of unfortunate coincidences.
The causes are well-documented to anyone who has watched ACT campaigns closely. The party's recent campaigns have relied on Australia-based corporate campaign professionals who produce technically proficient but generically conservative messaging — slick, robotic, and stripped of the sharp libertarian edge that built ACT's base in the first place. As the campaign enters its formal period and the party's professional handlers tighten control, the distinctiveness that brought voters to ACT disappears. The result is a campaign that says nothing the National Party is not also saying, only with less credibility and fewer MPs.
The party also has a documented history of failing to manage its own internal culture. Volunteers and lower-ranked candidates who are not inside the professional campaign's inner circle find themselves frozen out, under-resourced, or quietly alienated. Grass-roots networks that should be mobilising voters in the final weeks dissipate through attrition and resentment. And scandals — the product of a candidate selection process that prioritises ideological alignment over basic due diligence — tend to surface during the regulated campaign period, when they receive maximum media attention and the party has minimum time to contain the damage.
The implication for the seat projection is significant. If ACT's party vote slides from its current 8 per cent polling to 5–6 per cent on election day — a compression entirely consistent with its historical pattern — it arrives in parliament with seven or eight seats rather than ten. If it falls to or below 5 per cent without winning the Tamaki electorate, all ACT party votes are wasted entirely, and the right bloc loses approximately eight seats from the projection above. This scenario — low probability but non-trivial — would change the election from a question of who Peters chooses to a question of whether the left bloc can form a government outright.
Underneath all of this sits a deeper structural question that ACT has not answered: what does the party look like after David Seymour? Seymour holds the Epsom electorate — the historic coat-tail seat that has kept ACT in parliament since Don Brash brokered the arrangement with Rodney Hide in 2008. He has rebuilt ACT from 0.7 per cent and a single MP in 2011 to a party polling at 8–10 per cent and holding two electorates in 2023. He is the party's debate performer, its media brand, its philosophical anchor, and its only nationally recognisable face. Without him, ACT is a list of policies in search of a personality.
Seymour has publicly indicated that he would retire and settle down at some stage, recently announcing being engaged to a high-end property purchaser. His departure from ACT's leadership — however it occurs and whenever it comes — would represent an existential transition. Van Velden's exit removes the only other ACT figure with anything approaching a national profile. The departure of senior staffers alongside her hollows out the institutional knowledge. James Christmas, contesting Tamaki, has the credentials but not the profile. ACT, in 2026, is a one-person party being held together by the force of Seymour's personal vote in Epsom and his performance in the campaign debates.
New Zealand First, by contrast, has a documented and consistent history of spiking closer to the election and achieving results that exceed most polls. Winston Peters performs in the formal campaign period — his debate skills, his command of the press conference, his ability to dominate the political news cycle with precisely targeted salvos — in ways that drive late-deciding voters toward NZF. The shy NZF voter is the structural complement to the shy Tory effect: voters reluctant to admit to pollsters that they are backing Peters nevertheless tick his party on election day.
In 2023, NZF polled at around 4–6 per cent in the final weeks and delivered 6.09 per cent. In elections where NZF is polling at 12–15 per cent, the combination of his late-campaign surge, the shy-voter effect, and the formal campaign period's television debate schedule — where Peters excels — suggests the actual result could be 15–18 per cent. At 18 per cent, NZF holds approximately 22 seats in a 130-seat parliament. Peters does not just hold the balance of power. He is the largest single coalition voice in the building.
The third pattern is the most analytically underweighted in current coverage. Labour may substantially reduce Te Pati Maori's parliamentary representation — and the mechanism that has protected TPM in recent elections is showing signs of breaking down.
The Maori electoral roll's split-ticket architecture has delivered TPM enormous structural leverage under MMP. Voters on the Maori roll who give their candidate vote to TPM and their party vote to Labour have effectively doubled their electoral influence: TPM coat-tails into parliament with six or seven seats on 3 per cent of the party vote, while Labour also benefits from those party votes in the proportional allocation. The result has been a left bloc disproportionately stronger than the party vote alone suggests, because TPM's overhang seats come at the cost of no proportional entitlement — they are genuinely additional seats that expand parliament and require no list allocation.
But in a struggling economy — with a fuel crisis, food processing plants closing, and a Press Gallery framing the election as a choice between a competent Labour-led government and a chaotic National-led one — the calculus for Maori voters changes. Labour is contesting all seven Maori electorates in 2026. Te Pati Maori has spent the past 18 months in internal chaos: Mariameno Kapa-Kingi and Takuta Ferris were expelled, a by-election was held in Tamaki Makaurau in September 2025, and the party's co-leadership is contested. The cultural and political momentum that delivered six electorate seats in 2023 on the back of the Treaty Principles Bill backlash is not what TPM is fielding in 2026.
If voters on the Maori roll decide that Labour is the safer pair of hands in an economic crisis — that the party with the track record and governing credibility is more likely to protect Maori interests in a difficult term — they may give both their party vote and their candidate vote to Labour. Each Maori electorate Labour wins from TPM is a seat the left bloc holds without the overhang mechanics. The disproportional advantage the Maori roll split-ticket has delivered to the left bloc in recent elections does not disappear — it transfers from TPM to Labour, where it sits more securely and without the three-party negotiation overhead.
There is a fourth dynamic that sits above all three of these patterns and that could become the defining analytical story of this election if the overhang argument gains public traction. Once the overhang strategy is widely understood — once it becomes common knowledge that right-bloc voters giving their party vote to ACT and NZF while voting for the National candidate produces more seats for the right than any other tactical combination — the question becomes whether the left can respond in kind. And the answer is that the demographic most likely to do so has already demonstrated, more clearly than any other group in New Zealand's MMP history, that it understands exactly how to play this game.
Maori voters are more sophisticated strategic split-ticket voters than any other demographic in New Zealand. They have been doing this deliberately and effectively for two decades. The Maori roll's ability to deliver TPM coat-tail seats on small party votes while simultaneously boosting Labour's proportional allocation is not an accident. It is a community that has read the MMP rulebook carefully and acted on it collectively. If those same voters understand that National is generating a right-bloc overhang — and that the left needs its own structural counter-response — they are the most likely voters in the country to act on that understanding, at scale, in a coordinated way.
The counter-strategy is straightforward. Give the candidate vote to TPM in electorates where TPM holds or challenges, generating Maori electorate overhang on the left. Give the party vote to Labour, maximising Labour's proportional allocation. The result expands parliament in the opposite direction to National's overhang, adding seats to the left bloc that the right cannot match through the proportional calculation. The two overhangs, stacking on top of each other, could deliver a parliament of 132 to 136 seats — and a left bloc with a majority despite a party vote that looks, on the raw numbers, insufficient.
This is not hypothetical. It is the logical extension of a strategic culture that Maori voters have already demonstrated. The question is whether the overhang argument reaches them, and reaches them in time. A political journalist or commentator who publishes the structural analysis — explaining, clearly, that National's split-ticket advantage can be countered by an equivalent Maori roll split-ticket — before the regulated campaign period begins could materially change the arithmetic of this election. The 2026 election may therefore be remembered as the most consciously, deliberately strategic vote-splitting contest in a generation: right-bloc voters generating one overhang, Maori roll voters generating the counter-overhang, with the New Zealand First kingmaker holding the balance between two expanded blocs that have both played the system to its logical limits.
The combined effect of ACT's late implosion, NZF's late surge, Labour's potential gains in the Maori electorates, and the possibility of a conscious counter-overhang strategy from the Maori roll is a projected election night result that looks substantially different from April polling. ACT smaller. NZF larger. The parliament larger than any Sainte-Laugué projection from current polling would predict. Peters holding more leverage, not less, than the headline numbers suggest. And the New Zealand voter — once described as too small a country to have complicated politics — demonstrating that it has, in fact, mastered the most sophisticated form of proportional representation strategy in the democratic world.
The redrawn map: What the 2025 boundary review means
The 2025 Representation Commission review delivered the most substantial Wellington reshuffle in the MMP era. Three electorates — Ohariu, Mana, and Otaki — were abolished and replaced by Kenepuru and Kapiti. Wellington Central became Wellington North, shifting north to absorb Khandallah, Ngaio, and Wadestown. Rongotai became Wellington Bays, extending into Brooklyn and Mt Cook. Ohariu, held by Labour's Greg O'Connor since 2017, who beat Nicola Willis — the National candidate — by 1,260 votes in 2023, ceased to exist entirely. O'Connor is seeking a Labour list position with an eye on the Speakership.
National holds no Wellington general electorates. This was true before the boundary changes. The new map confirms it structurally. Wellington North (Green), Wellington Bays (Green or Labour), and Kenepuru (safely Labour from its Porirua base) offer National no path whatsoever at current polling.
TABLE 3: 2025 BOUNDARY CHANGES — ELECTORAL IMPACT
Electorate | Boundary change | Direction | Impact for National |
Hutt South | Gains Newlands (Lab 35%, Nat 31%); loses Epuni/Avalon to Remutaka | Shifts left | Bishop's 1,332 majority eroded. Most at-risk National seat. |
Maungakiekie | Gains 9,600 from Panmure (Lab 42%, Nat 29%); loses Greenlane to Epsom | Shifts left | Fleming's 4,607 majority substantially reduced. Genuine seat-loss risk. |
Wellington North | Loses Mt Cook/Brooklyn; gains Khandallah, Ngaio (Nat 40%) | Slightly right | Green MP Paul likely holds. No National path. Willis may contest. |
Mount Albert | Loses 15,300 left-wing; gains Balmoral, Sandringham (Nat 42%) | Shifts right | Labour held by 18 votes in 2023. Genuine National pickup opportunity. |
Wigram | Gains Prebbleton (Nat 55%); loses Addington/Spreydon to Chch Central | Shifts right | Open seat — Megan Woods list-only. Strong National prospect. |
Palmerston North | Gains 10,000 from Rangitikei (65% coalition voters) | Shifts right | Utikere faces harder race. Possible National pickup. |
Tukituki | Loses Pakowhai, Waiohiki, Fernhill north to Napier | Slightly left | Loses conservative rural base. Compounds Wattie's/McCain crisis. |
Mt Roskill | Gains Blockhouse Bay (Nat 48%) from New Lynn | Shifts right | Helps Cheung hold a 2023 Labour-wave gain. |
The electorates at risk: National's vulnerability matrix
Hutt South is the most watched electorate in the country. Chris Bishop won it in 2023 by 1,332 votes — reclaiming a seat he had held from 2017 to 2020, lost to Labour's Ginny Andersen in the 2020 wave, and recovered in the 2023 counter-swing. The seat swung 5,000 votes across two elections. It is one of the most responsive electorates in the country to national sentiment. Newlands, absorbed from the abolished Ohariu, voted Labour 35 per cent and National 31 per cent in 2023. Adding net 4,700 left-leaning residents to a seat held by 1,332 votes is arithmetically straightforward.
Bishop was demoted in the April 2 reshuffle — losing the Leader of the House role (to Louise Upston) and the campaign chair role (to Simeon Brown). He was reported to have attempted to gather support to challenge Luxon in November 2025. He appeared on Q+A on Sunday displaying, as the Spinoff put it, "an authority and coherence, a passion and sense of humour that stood as an indictment of a prime minister who struggles with all of that." It is not the performance of a man whose electorate seat is secure.
TABLE 4: NATIONAL ELECTORATES — VULNERABILITY ASSESSMENT
Electorate | National MP | 2023 Majority | Boundary effect | Econ stress | Risk |
Hutt South | Chris Bishop | 1,332 | Adverse (left shift) | Very high | CRITICAL |
Tukituki | Catherine Wedd | ~10,000 | Adverse (loses north) | Very high | HIGH — 3-way contest |
Maungakiekie | Greg Fleming | 4,607 | Adverse (left shift) | High | HIGH |
Napier | Katie Nimon | 2,800 | Gains rural buffer | Very high | HIGH |
Northland | Grant McCallum | ~3,200 | Unchanged | High | HIGH |
Whangarei | Shane Reti | ~4,200 | Unchanged | High | MEDIUM — retiring |
Mt Roskill | Carlos Cheung | ~1,100 | Favourable (right shift) | Medium | MEDIUM — boundary helps |
Pakuranga | Simeon Brown | 7,800 | Unchanged | Medium | LOW |
East Coast Bays | Erica Stanford | ~8,000 | Unchanged | Low | LOW |
Botany | Christopher Luxon | 9,200 | Unchanged | Medium | LOW — PM's seat |
Sources: NZ Herald electorate reporting; The Spinoff.
The list that vanishes: Who does not come back
National brought five list MPs into parliament at the 2023 election on 38 per cent of the party vote and 43 electorate seats: Nicola Willis, Paul Goldsmith, Melissa Lee, Gerry Brownlee, and Nancy Lu — in that order. Willis, as Finance Minister and Deputy Leader, is a list MP. She lost the Ohariu contest to Labour's Greg O'Connor in 2023 and has no electorate anchor. Her position at the top of the list is her only protection — and at zero list seats, that protection is worth nothing.
James Christmas, a National barrister who worked under John Key, Bill English, and Chris Finlayson, missed parliament in 2023 by three list places. Having watched National's party vote collapse from 38 per cent to around 30 per cent since then, he will have run the list arithmetic himself: at 30 per cent with 43 electorates, there are no list seats whatsoever. His switch to ACT is not ideological drift. It is rational parliamentary career planning.
At 30 per cent with 43–44 electorate seats, National's proportional entitlement is approximately 36 seats. If they hold 43 electorates, they have zero list seats. Every MP not holding an electorate is gone. At 32 per cent with 44 electorates, there is perhaps one list seat — Willis. The list does not contract gradually. It falls off a cliff.
The lower the list position, the higher the exposure. At 30 per cent with 43 seats, the line is drawn above the entire list.
Gerry Brownlee, the Speaker, is retiring at the 2026 election. His departure removes a board-nominated list slot but does not free up any room below him.
TABLE 5: NATIONAL MPs MOST AT RISK — LIST COMPRESSION AND ELECTORATE VULNERABILITY
MP | Portfolio | List rank / seat | Exposure | Notes |
Nicola Willis | Finance (Deputy Leader) | List rank ~2 | HIGH RISK | List MP — lost Ohariu contest to Greg O'Connor 2023. High list rank is her only protection; at zero list seats she is also gone. |
Maureen Pugh | Backbench | Lower list | CERTAIN LOSS | First casualty of list compression. No electorate. TU criticism. |
Suze Redmayne | Backbench | Lower list | CERTAIN LOSS | Below any rescue at current polling. |
Nancy Lu | Junior minister | Mid-low list (~20s) | VERY HIGH RISK | TU criticism. No electorate. List entitlement at 30% does not reach her rank. |
Melissa Lee | Broadcasting/ICT | Mid list | HIGH RISK | Long-serving, no electorate. List compression at 30% threatens position. |
Paul Goldsmith | Revenue/Trade | Mid list (~8–10) | HIGH RISK | List-only. "Hold your nerve." TU criticism. At zero list seats, gone regardless. |
Chris Bishop | Housing/Infra/AG | Hutt South (1,332 maj) | HIGH RISK | Adverse boundary + national swing. Demoted in reshuffle. No list backup if seat lost. |
Greg Fleming | Backbench | Maungakiekie (4,607 maj) | HIGH RISK | Adverse boundary. 4,607 majority under real threat. No list backstop at 30%. |
Gerry Brownlee | Speaker | List rank 4 (2023) | RETIRING | Retiring 2026. Removes a board-nominee list slot — does not free space below. |
Catherine Wedd | Tukituki MP | Electorate | HIGH RISK | Wattie's/McCain closures. Randell (NZF) splitting right vote. Boundary adverse. |
The caucus war: Factions, the seventeen, and a leadership in free fall
Political commentator Liam Hehir — speaking on Breakfast — described the disgruntled group as non-ministers whose careers have stalled, some disgraced, none current ministers. There is "no sort of leader to it." National has 49 MPs. Approximately 20 hold ministerial portfolios. That leaves roughly 29 non-ministerial MPs. Seventeen non-ministers would represent a majority of the non-ministerial caucus. Add four or five frustrated ministers, and you have a majority of the full caucus.
The chief whip Stuart Smith attempted to brief Luxon on flagging caucus support before Easter and was, by multiple accounts in the Herald, evaded by his own Prime Minister. Plotters want a voluntary resignation. An anonymous MP told Tova O'Brien: "Nobody wants blood to spill. Anything other than him stepping down would be a nightmare and he knows that."
The following table of caucus factions are based on media commentary. Contrast it with my analysis.
TABLE 6: NATIONAL CAUCUS FACTIONS — STRUCTURAL ANALYSIS
Faction | Members (public record) | Structural interests | Position on leadership crisis |
The Luxon Loyalists | Todd McClay, Mark Mitchell, Simeon Brown, Louise Upston, Shane Reti, Nicola Willis (list) | Safe seats or high list rankings. Ministerial portfolios dependent on Luxon survival. Northern Auckland growth corridor. | "1000% confidence" (McClay). "Rock solid" (Mitchell). The reshuffle's beneficiaries. |
The Bishop Wing | Chris Bishop, Erica Stanford | Housing reform — prices must fall. Confronting the property-preservation funding base. Publicly denied any challenge; body language contradicts words. | Bishop demoted, then ubiquitous in media. Stanford (0.9% preferred PM) also mentioned. Both implicitly repositioning. |
The Disgruntled Backbench | Greg Fleming, Maureen Pugh, Nancy Lu (list), Melissa Lee (list), Suze Redmayne (list), multiple 2023 Auckland intake MPs | Imminent unemployment. Marginal electorates or very low list positions. Won Auckland Labour seats in 2023 on a tide that will not recur. | "Nobody wants blood to spill. Anything other than him stepping down would be a nightmare" — anonymous MP to Tova O'Brien. |
The Nervous Old Guard | Paul Goldsmith (list), Stuart Smith (whip), Gerry Brownlee (retiring, list) | Institutional memory of 2020 chaos. Publicly loyal, privately conflicted. Smith tried to brief Luxon on flagging support — was evaded for days. | "Hold your nerve, knuckle down." — Goldsmith. The institutional alarm bell is ringing. The PM is not answering it. |
The property-industrial complex: Who funds the party and what they want
The National Party's internal war is a conflict between two incompatible visions of the New Zealand economy — a conflict I analysed in detail in "The Bullshit Economy". The factions reflect who funds the party, who holds the growth-corridor electorates, and whose constituents benefit from the status quo of rising land values.
Sir John Key and Paula Bennett represent the party's financial spine. Key, who served as ANZ chairman after leaving the Beehive, predicted as recently as 2024 that house prices could double over the next decade. His formulation — that homeowners might complain their kids can't afford a home while their nest egg increases in value — is an unusually frank acknowledgement of the duplicity at the heart of New Zealand's property debate. The grievance and the beneficiary are the same person. The party's donor base has no interest in a government that delivers what Bishop has explicitly promised: cheaper land.
Bennett's role is more direct. Since leaving parliament in 2020, she has functioned as National's primary fundraiser. In a single three-week drive in 2022, she raised $1.8 million from Graeme Hart, Murray Bolton, Nick Mowbray, and Warren Lewis, whose $500,000 contribution is reportedly the largest single political donation in New Zealand history. She attempted to meet every person on the NBR Rich List. Between 2021 and 2023, National declared $8.2 million in large donations, much of it attributed to Bennett's network.
Bennett's former electorate — Upper Harbour — is now held by Cameron Brewer, who positioned himself for that specific seat for more than a decade. He declined Tamaki in 2011, declined ACT candidacy, and declined Helensville in 2017, before joining the Rodney Local Board in the Kumeu subdivision and winning Upper Harbour in 2023 with an 11,192-vote majority. Brewer was promoted in the April reshuffle to Commerce and Consumer Affairs Minister.
Brewer's career history maps precisely onto the Auckland property and business network. He was press secretary to Jenny Shipley and then to Rodney Hide — the ACT leader who designed the Auckland supercity structure. He was CEO of the Newmarket Business Association. He served on the Rodney Local Board in the Kumeu subdivision — the heart of Auckland's northern greenfields growth zone. Upper Harbour covers Hobsonville, Scott Point, and Albany. This is succession planning by a network.
Mark Mitchell (Rodney), Chris Penk (Kaipara ki Mahurangi), Erica Stanford (East Coast Bays), and Simon Watts (North Shore) represent the same corridor — an unbroken arc from Auckland's inner north to the Northland border, encompassing every major greenfields development zone: Kumeu-Huapai, Warkworth, Hobsonville, Scott Point, the Albany basin. These are the electorates where rezoning decisions, motorway infrastructure (SH16, Penlink), and water and wastewater investment directly govern land values. The April reshuffle promoted Penk into Cabinet, gave Brewer his first ministerial role, and moved Watts to the Auckland portfolio. It rewarded this network.
Against this stands Chris Bishop, from Wellington, explicitly arguing that house prices must fall — that "declining rates of homeownership is the greatest threat to the centre-right worldwide." He was attacking the economic model on which the party's donor base depends. Luxon contradicted him publicly, saying he wanted "steady, consistent increases in property values." Willis told Auckland buyers they should "absolutely" buy a house. The conflict is not subterranean. It is in the Hansard. I have written at length on the structural drivers of this conflict in "The Bullshit Economy" and in the electricity gentailer series at klaut.media.
Tamaki: Van Velden goes, Christmas arrives, and it is not a gift seat
Brooke van Velden won the Tamaki electorate from National's Simon O'Connor in 2023 — the first time the seat had changed hands from National since 1957. O'Connor is no longer in politics. Van Velden announced last month she would not stand in the November election, citing a desire for private sector work. She is 33.
James Christmas, a barrister who worked under Key, English, and Finlayson, and co-authored a book on Treaty settlements with Finlayson, has switched from National to ACT. Christmas missed parliament in 2023 by three list places — National won 43 electorates and five list seats, and Christmas was next but not close enough. Having watched National's party vote collapse from 38 per cent to around 30 per cent since then, he will have run the list arithmetic himself: there are no list seats whatsoever at that level with that electorate count. His switch to ACT is not ideological drift. It is rational parliamentary career planning. National has selected Mahesh Muralidhar to contest Tamaki. Labour has announced Max Harris. Three credible parties are competing for a seat that has only ever been held by National or ACT since 1957.
The Tukituki crucible: Where everything is on the line
Tukituki is the most analytically rich electorate in the 2026 campaign. It sits at the intersection of the government's most conspicuous failures — Cyclone Gabrielle recovery, food processing industry collapse, electricity costs, wine glut, rural land use conflict — and is now the site of a three-way contest in which NZF's most high-profile candidate carries ideological baggage that is, to put it charitably, difficult to explain to his own party's base.
The seat is held by Catherine Wedd (National), who won it in 2023 from Labour's Anna Lorck with a majority of approximately 10,000 votes. Wedd is a first-term electorate MP noted by commentators as having rising influence within National — yet not promoted in the April reshuffle.
In the same week in late March 2026, Heinz Wattie's — founded in Hastings in 1934 — confirmed closure of its frozen packing lines in Hastings, while McCain Foods announced closure of its Hastings vegetable processing plant. The Omahu Road plant processed over 50,000 tonnes of vegetables annually. The closures affect more than 100 growers — 21 per cent of the processed vegetable sector — with an estimated $18 million annual revenue loss. Central Hawke's Bay Mayor Will Foley: "hundreds and hundreds of people impacted... we'll be getting into the thousands." Wattie's had already cut Hawke's Bay peach contracts in September 2025 following Chinese dumping.
Cyclone Gabrielle devastated the region in February 2023. The rebuild has been chronically slow and underfunded. Companies cited energy costs and post-cyclone inflation as reasons not to rebuild. The government's inadequate recovery response created the conditions in which these commercial decisions became rational. The Heretaunga Plains are among the most productive horticultural soils in the Southern Hemisphere — yet two structural threats press simultaneously: urban sprawl onto fertile land, and exotic forestry for ETS carbon credits displacing productive agriculture. Add the wine glut compressing Hawke's Bay viticulture, electricity costs surging through horticulture inputs, high household debt in suburbs such as Flaxmere, and the boundary change losing conservative rural communities north of the Ngaruroro River to Napier — and Tukituki becomes an electorate where every crisis of this government arrives simultaneously.
In 2019, Randell brokered a deal with the Hawke's Bay Regional Council to lend 100,000 carbon units — worth $2.5 million — to Kahutia to fund tree planting on erosion-prone land. The deal involves exactly the nexus — ETS carbon credits, iwi asset management, preferential Council arrangements — that New Zealand First has built its rural conservative base on opposing. Shane Jones has been scathing about "zombie forests." NZF's opposition to co-governance and Treaty-based preferential arrangements is the bedrock of its appeal to rural conservative voters.
Randell had prior contact with Labour and the Maori Party before choosing NZF. Hipkins' comment at the candidacy announcement — "Taine Randell, former All Black captain, he's soon going to discover there's only one captain in New Zealand First" — registers both the political threat and the cultural reality. Peters runs a tight ship.
The question for Tukituki is not whether Randell can win outright. National's 10,000-vote majority is a substantial wall. But that analysis predated the Wattie's and McCain closures, the April Verian poll at 29.7 per cent, and Randell's candidacy. If Randell takes 10–15 per cent of the electorate vote from National's right flank, Wedd's effective majority against Labour contracts sharply. A national swing of 8 per cent from National to Labour reduces a 10,000-vote margin by several thousand. With a strong Labour candidate running the factory closures as the campaign centrepiece, a Labour win in Tukituki is not the base case. It is a realistic scenario. And for NZF, regardless of the electorate result, Randell will almost certainly hold a winnable list position — making Tukituki a proving ground for the post-Peters future.
Winston Peters: The perennial kingmaker
In 2017, New Zealand First polled 7.2 per cent. Peters chose Jacinda Ardern's Labour over a National Party that had won the most seats and the highest party vote. The result was a Labour-led government. The precedent is not obscure. It is the established operating manual of New Zealand First in a hung parliament.
In 2026, NZF is polling at 12–15 per cent. Peters is the third-highest preferred prime minister at 12 per cent — the only upward trajectory in the entire political landscape outside Labour. His party's policies — opposition to mass immigration, energy reform, scepticism of the ETS, rural economic advocacy — are all experiencing a surge of popular relevance as the fuel crisis, food processing closures, and economic stagnation make his diagnosis of New Zealand's problems look prescient.
In any scenario where the right bloc does not hold a clear majority — and the base case shows National plus ACT with approximately 54 seats in a 128–130 seat parliament against a majority requirement of 65–66 — Peters holds the balance of power. He has done this before. He will do it again. National's internal chaos, its leadership uncertainty, and its fundamental conflict over housing policy make it a less attractive coalition partner than its seat count might suggest. In 2017, he chose the party with the better policy offer. The same calculus applies in 2026.
