Opinion poll trends for the 2026 New Zealand General Election
Dr Lindsay R Morris
Overview
This report summarises opinion polling trends for seven political parties contesting the 2026 New Zealand General Election: National, Labour, the Greens, ACT, New Zealand First, Te Pāti Māori and the Opportunity Party.
It presents modelled estimates of party support, parliamentary representation, preferred Prime Minister and government approval. The analysis covers polls published between 6 November 2023 and 7 November 2026.
The estimates describe underlying trends in published polling data. They are not forecasts of the final election result. Polls are subject to sampling error, differences in survey methods and changes in public opinion between surveys.
Party support
The table summarises estimated support for each party, its change since the 2023 election, and its modelled trend. The parenthetical uncertainty shown with each estimate is the model-reported uncertainty measure.
| Party | Estimated support | Change since 2023 | Modelled trend |
|---|---|---|---|
| National | 28.3% (± 0.54) | down 10 points | Decreasing |
| Labour | 27.9% (± 1.42) | up 1 points | Decreasing |
| The Greens | 15.2% (± 1.18) | up 3.6 points | Increasing |
| ACT | 10.2% (± 0.85) | up 1.6 points | Increasing |
| New Zealand First | 9.9% (± 0.76) | up 3.8 points | No clear directional trend |
| Opportunity | 6.9% (± 1.05) | up 4.7 points | No clear directional trend |
| Te Pāti Māori | 1.2% (± 0.3) | down 2 points | Decreasing |
The trend labels describe the direction supported by the model. No clear directional trend means that the model does not provide sufficient evidence for either an increasing or decreasing trend; it should not be interpreted as proof that support is unchanged.
Published opinion poll results and modelled trends in party support.
Text alternative: The figure plots published poll results and modelled trends in support for the seven parties between 6 November 2023 and 7 November 2026. The trend directions are: National: Decreasing; Labour: Decreasing; The Greens: Increasing; ACT: Increasing; New Zealand First: No clear directional trend; Opportunity: No clear directional trend; Te Pāti Māori: Decreasing.
Parliamentary representation
The table shows each party’s estimated seat allocation, the range reported by the model and the change from the 2023 General Election. Seat estimates are based on 50,000 simulations of modelled party support.
| Party | Estimated seats | Modelled seat range | Change from 2023 |
|---|---|---|---|
| National | 34 | 30–39 | down 14 seats |
| Labour | 34 | 29–38 | unchanged |
| The Greens | 18 | 15–22 | up 3 seats |
| ACT | 12 | 10–15 | up 1 seat |
| New Zealand First | 12 | 10–15 | up 4 seats |
| Opportunity | 8 | 0–12 | up 8 seats |
| Te Pāti Māori | 3 | 3–3 | down 3 seats |
The following graphic illustrates the modelled parliamentary composition by party.
Illustrative parliamentary composition based on the modelled seat estimates.
Text alternative: The graphic shows modelled seat estimates for each party: National: 34 seats; Labour: 34 seats; The Greens: 18 seats; ACT: 12 seats; New Zealand First: 12 seats; Opportunity: 8 seats; Te Pāti Māori: 3 seats. These are estimates derived from the polling model and seat-allocation assumptions, not a prediction of the final parliamentary composition.
In the simulations, National–ACT–New Zealand First collectively held a majority of seats in 21.4% of cases. Labour–The Greens–Opportunity–Te Pāti Māori collectively held a majority in 72.5% of cases. In the remaining 6.1%, neither of these two groupings held a majority.
These are analytical groupings used for the simulations. They should not be interpreted as predictions of post-election coalition agreements or commitments by the parties concerned.
Seat allocations assume that each party retains at least one electorate seat it won in 2023, allowing it to remain represented even if its party vote falls below 5%. Based on electorate polling, Te Pāti Māori seat allocations assume that the party wins three electorate seats. The ranges and party-level estimates are produced from simulations; rounded party estimates may not sum exactly to 120.
Preferred Prime Minister
The table summarises modelled support for the preferred Prime Minister candidates.
| Leader | Estimated support | Modelled trend |
|---|---|---|
| Christopher Luxon | 21.1% (± 1.53) | No clear directional trend |
| Chris Hipkins | 19.8% (± 1.53) | No clear directional trend |
| Winston Peters | 8.5% (± 1.24) | Decreasing |
| Chlöe Swarbrick | 7.3% (± 0.82) | No clear directional trend |
| David Seymour | 5.8% (± 0.54) | No clear directional trend |
Preferred Prime Minister results were analysed across five of the six polling organisations with relevant questions. The Post–Freshwater Strategy polls were excluded because respondents could choose only between Chris Hipkins and Christopher Luxon, making those results not directly comparable with polls offering a wider choice of candidates.
Published poll results and modelled trends in preferred-Prime-Minister support.
Text alternative: The figure plots published poll results and modelled trends in preferred-Prime-Minister support between 6 November 2023 and 7 November 2026. Current modelled support estimates are: Christopher Luxon: 21.1% (± 1.53); Chris Hipkins: 19.8% (± 1.53); Winston Peters: 8.5% (± 1.24); Chlöe Swarbrick: 7.3% (± 0.82); David Seymour: 5.8% (± 0.54).
Government approval
The latest modelled net government approval estimate is -17.7 (model-reported uncertainty: ±3.4). The modelled trend is significantly decreasing by approximately 0.3 points per month.
Modelled trend in net government approval.
The polling data
Seven polling organisations contributed data to the analysis. Their polling frequency differs, as summarised below.
| Polling organisation | Most recent poll | Number of polls | Median interval (days) |
|---|---|---|---|
| The Post–Freshwater Strategy | 3 November 2026 | 9 | 78.0 days |
| Taxpayers’ Union–Curia | 3 November 2026 | 36 | 30.0 days |
| 1 News–Verian | 2 November 2026 | 22 | 56.0 days |
| RNZ–Reid Research | 26 October 2026 | 9 | 60.5 days |
| Roy Morgan | 25 October 2026 | 36 | 28.0 days |
| Anacta | 10 October 2026 | 2 | 30.0 days |
| Talbot Mills | 11 August 2026 | 21 | 33.0 days |
The median interval is the median number of days between polls published by each organisation over the period covered by the source data. Polling frequency is one factor affecting how quickly new survey results become available; it does not, by itself, establish that one organisation’s results are more accurate than another’s.
Methodology and limitations
Trend estimation
Generalised additive models (GAMs) were used to estimate underlying trends in party percent support over time. GAMs allow trends to change smoothly rather than assuming a straight-line increase or decrease. Models were fitted using a beta distribution, which is appropriate for percentages because it constrains estimates to the 0–100% range.
Smooth functions of time were specified using thin plate regression splines, with the basis dimension (k) set to 0.8 times the number of polls. Smoothing parameters were estimated using restricted maximum likelihood (REML).
A trend is labelled increasing or decreasing when the modelled probability of that direction exceeds 95%. Otherwise, it is labelled no clear directional trend. This classification is not a statement that support is necessarily constant.
Seat estimates
Seat estimates are generated from simulations of modelled party support. They depend on the seat-allocation assumptions described in the parliamentary representation section, including assumptions about electorate seats. They do not explicitly model all possible electorate-level outcomes, voter turnout or post-election negotiations.
Pollster effects
The pollster-effects analysis compares reported party support across polling organisations while holding the other modelled factors constant. Talbot Mills is the reference organisation. An odds ratio above 1 indicates higher estimated odds of reporting support for a party relative to Talbot Mills; an odds ratio below 1 indicates lower estimated odds.
The table below includes only pollster effects marked as statistically distinguishable from the reference organisation by the existing model output. These are differences in reported support after adjustment for the modelled factors; they should not be interpreted as rankings of polling organisation quality or accuracy.
| Party | Polling organisation | Odds ratio | Direction relative to Talbot Mills |
|---|---|---|---|
| ACT | Roy Morgan | 1.25 | Higher estimated odds |
| ACT | Taxpayers’ Union–Curia | 1.09 | Higher estimated odds |
| Labour | 1 News–Verian | 0.90 | Lower estimated odds |
| Labour | Roy Morgan | 0.76 | Lower estimated odds |
| Labour | Taxpayers’ Union–Curia | 0.89 | Lower estimated odds |
| National | 1 News–Verian | 1.08 | Higher estimated odds |
| New Zealand First | 1 News–Verian | 0.86 | Lower estimated odds |
| Opportunity | Roy Morgan | 1.62 | Higher estimated odds |
| Te Pāti Māori | 1 News–Verian | 0.72 | Lower estimated odds |
| The Greens | 1 News–Verian | 1.11 | Higher estimated odds |
| The Greens | Roy Morgan | 1.30 | Higher estimated odds |
Overall limitations
Opinion polls are subject to sampling error, methodological differences and changes in public sentiment over time. Late campaign effects, voter turnout and electorate-level dynamics are not explicitly modelled. Estimates of party support, parliamentary seats, preferred Prime Minister and government approval are model-based summaries of available polling data. They may differ from the results on 7 November 2026 and should be interpreted as indicative estimates rather than definitive forecasts.
Further information
The interactive dashboard provides a snapshot of the data behind this report.
The author is independent and does not endorse or favour any political party or candidate. The analysis is based on publicly available opinion poll data and is intended for informational purposes only.
For further information or collaboration, visit lindsay.kiwi.