The New Zealand electoral system tends to produce a flattening effect on the results. Most electorate results don’t really make a difference to the national outcome so the results analysis tends to just look at the national figures. But there is value in delving down and looking at the geography.
Overall the swings away from National and towards Labour were remarkably consistent. In this post I’ll include some maps and tables showing the swings in different parts of New Zealand, and look at how the swing on the candidate vote tended to be bigger in National seats.
I have compiled an estimate of the 2017 results based on the 2020 redistribution. I don’t believe anyone else has done this in New Zealand. This means it is possible to calculate swings from this data. You can view these 2017 estimates here. Elections NZ did not make it easy to grab all the booth data in one go – the data was in 142 separate CSVs in a messy format, there are no unique booth IDs and the macrons used for many electorate names and booth names were treated inconsistently by my software. But it’s finally finished and will be used in this post.
Firstly, let’s look at the electorate seats. I’m basing this analysis on data I downloaded on Tuesday evening.
Prior to the election, National held 42 seats (including one newly created notional seat), Labour held 29 and ACT held one.
Following this election National has been reduced to just 26 seats, with Labour on 43 seats, with ACT, Green and the Māori Party each holding one seat. Interestingly this means that every party in Parliament holds at least one electorate seat.
Labour gained fifteen seats off National, Green gained one seat off National, and the Māori Party regained Waiariki from Labour. Most of these changes were in regional seats. National lost four seats in Auckland and one each in Wellington and Christchurch, with the other ten in regional areas.
This map shows the winner of each general seat, with the seats that changed in brighter colours.
You can click through to see the candidate vote and swing for the four biggest parties, as well as the change in margin, for each electorate.
But the votes that really matter are for the party vote, which decides the share of seats for each party in parliament.
The next map shows Labour’s party vote swing per general electorate, and can be toggled to show the swing for National and Green. It’s zoomed in on Auckland but you can scroll around or zoom out.
Labour tended to get smaller swings in the core of the Auckland area, and did much better on the outskirts of Auckland and in regional areas. You can see a similar trend in the seats surrounding Christchurch.
Another way to look at this is to break the country up into regions. I broke up seats into Auckland, Christchurch, Wellington and the remainders of the North and South islands, as well as the Maori seats. This table shows the swings to or from the six parties with representation in the new or old parliaments in each region.
| Region | ACT | Green | Labour | Māori | National | NZF |
| Auckland | 6.5 | 1.9 | 11.7 | -0.2 | -17.9 | -3.2 |
| Christchurch | 6.6 | 2.7 | 12.7 | -0.1 | -19.6 | -3.0 |
| Maori | 1.3 | 1.9 | 3.1 | 0.2 | -3.7 | -4.0 |
| North Island | 8.9 | 0.3 | 14.3 | -0.1 | -19.0 | -6.4 |
| South Island | 9.5 | 0.5 | 13.3 | -0.1 | -19.7 | -4.7 |
| Wellington | 5.1 | 4.0 | 11.5 | -0.1 | -17.9 | -2.5 |
| New Zealand | 7.5 | 1.3 | 12.3 | -0.2 | -17.7 | -4.5 |
The race in the Maori electorates is very different to the rest of the country, with a very high Labour vote and a very low National vote, so the swings were much less dramatic.
The boost in the Green vote mostly took place in the big cities, while Labour’s biggest gains took place in the regional seats.
Another way to break down the results is to look at how much the candidate vote swung in Labour and National electorates. The swing to Labour in seats they retained averaged 11.5%, while the swing to Labour in seats they gained averaged 23.8%. The swing against National in seats they retained averaged 21.5%. This suggests that the biggest movement, at least on the candidate vote, took place in National electorates.
Of course there is a difference between how people vote on the party vote and on the candidate vote. There was a slightly higher Labour vote on the party vote (49.1% as against 47.5%) while the National vote was much higher on the candidate vote (35.7% as against 26.8%).
There is a history of these sorts of gaps appearing in past elections, which often result in one party winning almost every seat on the party vote while their opposition performs more respectively on the candidate vote, but that’s a story for another day.
Thanks for this Ben – quite interesting. It’s odd how little deep dive electoral analysis there seems to be in NZ. Perhaps it’s a consequence of the MMP electoral system – when it comes to how many MPs each party gets, the local electorate votes are almost an afterthought because they have almost no influence in the final outcome.
None the less I’m curious to know whether you see clear evidence of some tactical voting in the seat the Chloe Swarbrick won for the Greens. As you know, there were some earlier polls which suggested the Greens might not make the 5% threshold and thus end up with no MPs, unless they won a local electorate.
Andrew, Swarbrick managed 34.2% while the Greens only got 19.4% party vote in Central Auckland. Labour’s candidate got 32.4% while Labour itself got 45.1%. It’s very clear many Labour voters were intentionally voting to maximise the chance the Greens were not kicked out of parliament.
This is similarly the case in the neighbouring seat of Epsom where ACT wins electorate vote and National the party vote. This time David Seymour got 50.4% of the electorate vote while the party he leads got only 11.1% party vote in the seat! The National electorate candidate meanwhile got 14.8% to National’s 39.9% party vote!
I believe this is the case in the Maori seat as well (Labour party voters voting Maori for the electorate).
Elections NZ publishes the overlaps between party and candidate in each seat. The two ballots are on a single piece of paper so you can tell, unlike in Australia. But they haven’t published this data for 2020 yet.
I published a post about this exact data in Epsom back in 2014, if you go looking.
https://www.tallyroom.com.au/19967
While no doubt part of the Auckland Central result (and associated ticket-splitting) can be explained by tactical voting, it seems likely that it was also driven by Chlöe Swarbrick’s remarkable profile and popularity for a first-term list MP.
Hi Ben,
I hope you are well. My understanding is that there are still approximately 480,000 special votes which are yet to be counted & added to the national total. I’m looking at a few close electorate races in Auckland Central (Green lead 492 votes), Waiariki (Maori lead 415) & Whangarei (National lead 162). I would have expected all of these to be in doubt, given the close margins, yet everyone seems to be treating these outcomes as final. Am I missing something here?
I haven’t tried to delve into the close races. The NZ data is hard enough to analyse without trying to keep on top of those races. Auckland Central and Whangarei won’t matter to the national seat count anyway.
“I’m looking at a few close electorate races in Auckland Central (Green lead 492 votes), Waiariki (Maori lead 415) & Whangarei (National lead 162). I would have expected all of these to be in doubt, given the close margins, yet everyone seems to be treating these outcomes as final. Am I missing something here?”
Historically the special votes have been very good for the Greens and good for Labour and the Maori Party, and very bad for National and bad for New Zealand First. National always end up losing a list seat to the Greens after the specials are counted. In New Zealand you can enrol to vote at the booth on the day, and my understanding is that it’s largely young people and Maori who take this up-their votes get counted as special votes. My understanding also is that overseas Kiwis tend to be left leaning.
Because of this, I think National are in huge danger of losing Whangarei. Also, Maungakiekie could be marginally at risk for National. Swarbrick is more likely to increase her majority than decrease it, and the Maori Party likely have the upper edge (a lead of 400 in a Maori seat is the equivalent of a 700 vote lead in a general seat)
@Shazzadude – Thanks for the clarification. We’ll just need to wait another week or two & see what happens. It’ll be interesting to see whether anything changes after the special votes are included.
Thanks for this. Really useful. But maybe it’s just me but the Whangarei electorate doesn’t appear to work on the maps?
What I’ve found interesting is in 2017 the South Island, Urban centres and provincial towns all swung to Labour but Auckland (outside of the inner city) either had no swing or even a swing to National. This time the swing was more consistent even across Auckland.
It’s not just you, but it’s been fixed.
“ What I’ve found interesting is in 2017 the South Island, Urban centres and provincial towns all swung to Labour but Auckland (outside of the inner city) either had no swing or even a swing to National. This time the swing was more consistent even across Auckland.”
A lot of that was attributed to Labour taking a stronger stance on immigration and foreign home ownership, and the high Asian vote in Auckland taking exception
Comments are closed.