Method
We score the quality of each answer, not its direction
The rubric behind every score, the two weightings, the rules that keep scoring fair, how global risks are rated and the value behind that, how we checked our claims, and how confidence is rated.
Evidence cut-off: 19 October 2026. Every score and rating uses evidence to this date.
Overview
Every party is scored in the same 13 groupings, against the same eight tests, from the same kind of evidence, on the same date. Four core tests feed both scores; one test feeds only voter priorities; three feed only readiness.
How a score is built
8 tests, each 0 to 3 › Voter priorities: Right + Real + Now, out of 15 per grouping, weighted by voter concern › Readiness to 2050: Right + Real + Lasting, out of 21, weighted by consequence to 2050 › both scaled to 100.
Why Lasting carries most weight
Lasting holds three of the seven readiness tests (9 of 21 points) on purpose: it is the test a campaign is least likely to apply to itself. Voter priorities swaps them for one test: will people feel it before 2029? A separate confidence layer sits beside the scores and is never added to them.
Method
Eight tests, each scored 0 to 3, and a worked example
The readiness score for a grouping is the sum of seven tests, out of 21; the voter priorities score is the sum of the four core tests plus relief this term, out of 15. Silence in a grouping scores 0 on every criterion: it is a finding, not missing data. Deliberate inaction is scored on its merits.
| Criterion | 0 | 1 | 2 | 3 |
|---|---|---|---|---|
| Diagnosis | No position | Names the symptom | Names a driver | Names the driver, with evidence |
| Mechanism | None | Assertion only | Plausible, untested | Evidence and a measurable target |
| Funding | Uncosted | Cost stated, no source | Costed and funded in a plan | As 2, and independently reviewed |
| Delivery | No path | Intent only | Named vehicle and timeline | As 2, plus a track record |
| Relief this term (voter priorities only) | Nothing lands by 2029, or it makes the problem worse | Small, narrow or late | Noticeable relief for many, with a start date | Broad relief for most, with a start date and a measured effect |
| Future effect | Pushes large unfunded costs forward | Neutral | Modest build | Clearly strengthens the long run |
| Durability | Likely reversed | This party alone | Backed by a likely partner or institution | Cross-bloc support or a durable rule |
| Candour | Claims no losers | Vague | Names who pays | Names who pays, how much, and why |
Worked example: two opposite answers on NZ Super land 12 points apart
| Criterion | ACT: Super age to 67 by 2035 | Score | NZ First: no age rise, citizens only | Score |
|---|---|---|---|---|
| Diagnosis | Names ageing as the driver; cites Treasury | 3 | Frames the issue as eligibility, not cost | 1 |
| Mechanism | A known lever, Treasury modelling, a dated target | 3 | Eligibility change with no stated effect | 1 |
| Funding | Saving booked in a full plan; not reviewed | 2 | No saving or cost stated | 0 |
| Delivery | Phased; needs legislation; no partner support yet | 2 | "From 2029"; no vehicle | 1 |
| Future effect | Clearly strengthens the long run | 3 | Locks in rising costs with no way to pay | 0 |
| Durability | Opposed by its likely partner | 1 | A bottom line likely to hold in a right-bloc deal | 2 |
| Candour | Names who pays, how much; early access for physical jobs | 3 | Claims no losers | 0 |
| Relief this term (voter priorities only) | Makes it worse for those turning 65 | 0 | Keeps Super at 65 | 1 |
| Readiness /21 | 17 | 5 | ||
| Voter priorities /15 | 10 | 4 |
What the example shows
The rubric does not say raising the age is right. A well-evidenced plan to keep 65 and raise tax to pay for it could also score highly. NZ First scores low because it offers no way to pay. The two scores pull apart here: raising the age scores 17 of 21 for readiness but brings no relief this term.
Method
Two scores, two weightings, and why they differ
Voter priorities uses the core tests plus relief this term, weighted by how many people name each issue as a top concern. Readiness to 2050 uses the core tests plus the three long-run tests, weighted by each grouping's consequence for New Zealand to 2050, a judgement made openly and open to challenge.
| Grouping | Ipsos issues mapped | Voter share | Weight to 2050 |
|---|---|---|---|
| Cost of living and incomes | Cost of living 63, poverty and inequality 17, fuel 17, household debt 5 | 34% | 3% |
| Health | Healthcare 44, drugs and alcohol 7 | 17% | 6% |
| Economy, jobs and productivity | Economy 34, unemployment 21 | 18% | 16% |
| Housing | Housing 19 | 6% | 7% |
| Crime and justice | Crime 15 | 5% | 3% |
| Climate, energy and environment | Climate change 13, pollution and water 4 | 6% | 13% |
| Education and skills | Education 9 | 3% | 7% |
| Migration and population | Immigration 8, population 3 | 4% | 6% |
| Tax, public finances and retirement | Taxation 8 | 3% | 16% |
| Transport, infrastructure and local government | Transport and infrastructure 5 | 2% | 9% |
| Te Tiriti, Māori and the constitution | Race relations 4, issues facing Māori 3 | 2% | 3% |
| Defence, foreign affairs and trade | Defence and foreign affairs 2 | 1% | 9% |
| Community, inclusion and culture | Not listed (floor of 2) | 1% | 3% |
Voter shares: Ipsos NZ Issues Monitor, September 2026 (3 to 11 September, n=1,003), mapped to the groupings and normalised; groupings outside the Ipsos list get a floor of 2. [18] Weights to 2050: the author's judgement. Cost of living carries little weight to 2050 because its long-run drivers (productivity, housing, energy) are weighted in their own groupings; counting them twice would overstate it.
Sensitivity
- Equal weights: National 64, status quo 57, ACT 57, Labour 55, Opportunity 52, Greens 45, NZ First 35, Te Pāti Māori 35.
- Halving the defence weight lifts Opportunity to 58, because it is silent on defence and foreign affairs.
- Moving the climate or tax weights by a quarter changes no party by more than a point.
- National leads on readiness under every weighting we tested; the parties between 52 and 58 are too close to separate.
- Adding the relief test moved voter priorities scores by four points at most (Labour up 4, National up 3); Labour and Opportunity swapped places, and no other order changed.
Method
Global risks are rated on two counts, and rest on one stated value
The five global risks are rated, not scored, and the ratings are never added to the readiness score. Most parties say little on them, and adding silence into the score would squeeze every party towards zero and hide the differences that matter.
Two counts
- At home: what New Zealand controls: preparedness, regulation, capability and resilience.
- With others: what New Zealand does with other countries: treaties, multilateral institutions, standards bodies, alliances and Pacific partners.
The scale
| Rating | What it means |
|---|---|
| Silent | Nothing on record by 19 October 2026 |
| Acknowledges | Names the risk or its parts, no concrete measure |
| Plans | At least one concrete measure (a law, agency, programme, target or treaty position) with a stated vehicle |
| Funded plan | As Plans, and costed in the party's fiscal plan, or funded now for current settings |
The value behind it
The rest of this report scores quality, not direction. The global-risk ratings carry one stated value: these risks cannot be managed by any one country, so cooperation is part of the answer. We apply it evenly. A party offering a different form of cooperation (bilateral rather than multilateral, for example) can rate Plans. A party that leaves or rejects an agreement without proposing an alternative rates Acknowledges at most, whatever its reasons. Readers who do not share this value can read the at-home ratings on their own.
The horizon
The report looks to 2050. These risks grow beyond it, and some (advanced AI, engineered pandemics) could change quickly. The ratings describe what parties have said and planned by 19 October 2026, not how the risks will unfold.
Method
Continuity is rated, not scored, and rests on a second stated value
Each party gets four ratings for governing for the long term: its position on a four-year term, plans it offers to make with other parties, what it would do to long-term institutions, and whether it replaces what it scraps. Like the global-risk ratings, they are never added to the scores.
| Rating | What it means |
|---|---|
| Weakens | Opposes a four-year term; rejects working with others on long-term policy; would abolish or override an independent long-term body; or scraps a long-term commitment with no alternative |
| Silent | Nothing on record by 19 October 2026 |
| Acknowledges | Names the issue, or an alternative with no number or vehicle |
| Plans | At least one concrete measure with a stated vehicle; or keeps what it inherits, or replaces each thing it drops |
| Committed | As Plans, and already acted on with a current stated position, legislated, costed, or in force |
The value behind it
Governing for continuity matters: long-term plans need to survive a change of government. Continuity is not the same as never changing course. A party that changes direction and puts a stated alternative in place rates well; scrapping without an alternative rates low. Readers who do not share this value can ignore these ratings; the scores do not use them.
How we rated
Party documents and parliamentary votes to 19 October 2026. Four researchers rated two rows each; an independent check made the ratings consistent; ties went to the lower rating.
Method
Seven rules keep the scoring fair; one model and one confidence test apply to every party
The same rules apply to every party and grouping, and the 2050 debt model uses one set of assumptions for all.
Seven rules
- One cut-off date for all parties: 19 October 2026. Polls and the final leaders' debate are covered to 21 October; policy items announced after 19 October are listed as "Since 19 October (not yet scored)".
- Cite every score.
- Score blind where possible.
- A second reviewer checks the scores for consistency across groupings. In this edition 28 scores were adjusted by one point each on that review.
- Score grouping by grouping, all parties at once.
- Flag policies where likely partners differ sharply.
- Recompute with equal weights and report any change in order.
How the 2050 model works
- Start point: net core Crown debt of 39.5% of GDP in 2030/31, from Treasury's pre-election forecast.
- Path: Treasury's 2025 long-term projection (unchanged policy, 104% of GDP in 2050), rebased to that start point. PREFU 2026 forecasts lower debt than the forecasts Treasury's projection started from, so current settings reach about 96% in 2050, not 104%.
- Party effect: each party's change against the forecast in 2030/31 is held as a constant share of GDP and compounds through interest.
- Assumptions: interest 4.3% (Treasury's long-run bond rate); nominal growth 3.8% a year; GDP about $570b in 2030/31. Borrowing for assets is added as it is spent.
- Left out: growth claims and returns on infrastructure. This is cautious for parties that promise faster growth.
| Party | Ministerial experience | Stability | Partner record | Path to 61 | Confidence |
|---|---|---|---|---|---|
| National | Most of top 10 | Stable | Kept | Only with Opportunity, ruled out | Medium |
| Labour | Most of top 10 | Stable leadership | Kept | Short even with Opportunity (60 on the poll average) | Medium |
| Greens | Few | MP departures, 2024 | Kept | Only with Labour and Opportunity | Low |
| ACT | Under half | Stable | Kept | Only with Opportunity | Low |
| NZ First | Under half | Stable | Kept | Only with Opportunity | Low |
| Te Pāti Māori | None now | Expulsions; a breakaway party | Kept | Only via Labour and the Greens | Low |
| Opportunity | None | Stable | No record | It is the kingmaker | Low (by construction) |
Confidence that the plan happens
The confidence rating counts facts on record: ministerial experience in the top ten, stability since 2023, record with past partners, and whether stated partners reach 61 seats on the average of the two latest polls. High needs all four; Medium misses one; Low misses two or more. It never changes the scores.
Counts are our reading of each party's top ten at 19 October 2026.
Method
We checked our groupings and claims against Policy.nz
Policy.nz is the benchmark for comparing party policies, and every party claim in this report was checked against it.
| Grouping | Share of voter concern | Weight to 2050 | Policy.nz topics |
|---|---|---|---|
| Cost of living and incomes | 34% | 3% | Incomes and employment (prices, wages, benefits, tax and income support) |
| Health | 17% | 6% | Health |
| Economy, jobs and productivity | 18% | 16% | Economy (most subtopics) |
| Housing | 6% | 7% | Housing |
| Crime and justice | 5% | 3% | Law, justice and government (justice side) |
| Climate, energy and environment | 6% | 13% | Environment; energy |
| Education and skills | 3% | 7% | Education |
| Migration and population | 4% | 6% | Migration |
| Tax, public finances and retirement | 3% | 16% | Economy (government revenue and finance); superannuation; public sector |
| Transport, infrastructure and local government | 2% | 9% | Transport and infrastructure; local government |
| Te Tiriti, Māori and the constitution | 2% | 3% | Te ao Māori; parliament and democracy |
| Defence, foreign affairs and trade | 1% | 9% | Defence and foreign affairs; trade and exports |
| Community, inclusion and culture | 1% | 3% | Community and inclusion; media, culture and recreation |
- The groupings follow Policy.nz's topics. We compared our first 16 issues with Policy.nz's 13 topics and 74 subtopics, and its 659 policy summaries for the seven parties (21 October 2026). Its topics are easier for readers to find their way around, but they scatter the long-run drivers (Super sits under incomes, government finance under economy, and productivity has no topic). So this edition uses Policy.nz's topics, with tax, public finances and retirement drawn together, and adds the long game as pressures that cut across them.
- What the check added. Crime and justice, education, and community, inclusion and culture are now scored. Voters name crime (15%), education (9%) and poverty (17%) as top concerns, and the first edition had left them out.
- How conflicts were settled: from the parties' own documents. Policy.nz was not always right; where it differed from a party's own document, the party source was used.
- Coverage varies: Policy.nz holds 226 summaries for the Greens but 39 for NZ First and 45 for Te Pāti Māori, so it confirms less for the smaller parties.
Glossary
| Term | Meaning |
|---|---|
| GDP | Everything New Zealand produces in a year. Debt and spending are shown as a share of it. |
| Net core Crown debt | What the Government owes, less its liquid assets. |
| PREFU | Treasury's Pre-election Economic and Fiscal Update 2026, the forecast that costed plans are measured against. |
| Operating allowance | Money the forecast already sets aside for new spending each year. |
| Status quo | Current government settings. |
Method
How the quiz works: agreement and priorities, scored in your browser
The quiz has two parts. Neither is a recommendation: they show where you and the parties agree, and how well each party's plans answer the issues you care about.
Part A: positions match
You answer 30 of the report's statements, chosen so every grouping appears at least twice and framing is balanced. Each party's position comes from its own documents: agree, mixed (counted as neutral), disagree, or no stated position.
- Same answer as the party: 2 points. One step apart (agree and neutral, or neutral and disagree): 1 point. Opposite: 0.
- If the party has no stated position, it scores 0, and the results show how many of your statements each party is silent on, so a low match from silence is visible.
- Skipped statements count for nothing either way.
- "Matters more to me" counts a statement 2 times, in both points earned and points possible.
- Match = points earned divided by points possible, as a percentage, rounded.
Part B: your priorities score
You rate each of the 13 groupings from not important (0) to very important (3), then choose a balance between help now and readiness for 2050. Each party's score out of 100 is the weighted average, over the groupings you rated, of its Voter priorities score (out of 15) and its Readiness to 2050 score (out of 21), mixed by your balance. Current settings are scored the same way as a reference. If you rate every grouping as not important, the report's own weights are used, and the results say so.
With the report's voter weights and the balance fully on help now, the quiz reproduces each party's published Voter priorities score; with the weights to 2050 and the balance fully on 2050, it reproduces its Readiness to 2050 score. Automated tests check both, to within a point.
Privacy
Everything runs in your browser. Your answers are never sent or stored; your results exist only in the link the quiz gives you. Cookieless analytics count page views and quiz completions, nothing else. Take the quiz.