Voter turnout

Party vote performance in each age bracket

The most competitive market for votes in New Zealand is for those aged between 45-49, with National, Labour, the Greens, and the Māori Party all receiving, on average, similar shares of the party vote in electorates with above average shares of people in this age range.

Spurred on by my other recent work looking at voter turnout by age segments and how representative that made electorates, and how those age segments were more, or less likely to vote (indicatively at least), I thought I'd take the plunge and look at each party individually across all the Electoral Commission's age brackets that they collect data on, to see if it revealed any other insights.

What it's revealed is that while National and NZ First, and Labour, the Greens and the Māori Party might be poles apart in terms of their popularity with voters aged 18-34, such a big gap in voter preference based on age doesn't appear to manifest itself beyond 35 years old, starting to converge from its largest difference at 30-34-years-old, bar for the Green Party, Māori Party, and NZ First.

As I said in the introduction, it's interesting how at the 45-49 mark, National, Labour, the Greens and the Māori Party all converge, with all getting more or less similar party votes from electorates with an above average representation from this age bracket as they did on average across all the General Electorates. Though I'd caution here that as this analysis excludes the Māori Electorates (due to their big youth skew and low deviation among the seven electorates), that the Māori Party figures here should be taken with a grain of salt.

Also keep in mind that as you read through the following graphs, they do have different scales on the Y axis, so movements may be more pronounced in these than they are on the top comparitive graph.

National's worst performing age group appears to be the 30-34 bracket, though it under-performed on average in electorates with above average shares of voters under 44 to varying degrees, with it being weighted towards those electorates with above average shares of those aged 34 and under being the least likely to vote National. National also has the second most narrow deviation range in this analysis, with only The Opportunities Party doing better.  National's support is weighted towards those aged 45 and over, and peaks at those aged 65-69. In part, this larger likely support from older voters, who both enrol and vote at higher rates and in greater numbers than other age groups, no doubt contributed significantly to National's end result of 44.4%

As you'll see later, with NZ First doing so well with voters aged 55 and over, National can, in the short term at least, reinforce their vote by targeting NZ First's supporters. Longer term however, National will need to find a way to both preserve their strength in covers aged 50 and over, as well as doing a better job of appealing to younger demographics too.

Labour's graph in many ways is the mirror image of National's, albeit slightly more pronounced in its over and under-performance in the age brackets. The 30-34 age bracket is again interesting, as not only was this where National was most likely to perform worst, it's also where Labour performed best. Where National's deviation was relatively narrow, Labour's is much more pronounced, though it ranks in the middle of the six parties we're looking at in this. While Labour's support amongst youth voters is very strong, to offset the advantage that National gets from older voters, they would need much higher enrolment and turnout rates than they got even this election.

In news that will shock no one, New Zealand First's support overwhelmingly comes from those aged 55 and over, but especially those aged 60 and over. Because of this, NZ First has the largest deviation of any of the six parties in this analysis. NZ First and Winston Peters focus heavily on courting this demographic, so it's no surprise that they rely so heavily on support from them. It also begs the question that once NZ First loses its trump card for reaching them - Winston Peters himself - how are they going to manage going forward, as nobody else in the party seems able to capture that audience in the same way that Winston Peters does.

The Green Party have the second biggest deviation for their support after NZ First, and they're very much the opposite story to them too. Massive support across voters aged 39 and under, but this plummets to their average for those 40-44, before briefly rebounding for those aged 45-49, and not recovering beyond those aged 55 and over. Where the Green Party has an opportunity is to stop that leaching of support between people aged 35 years old and 44 years old, though in doing so they're likely to take voters from Labour.

As I wrote earlier, it's important to take these figures for the Māori Party with a grain of salt. This analysis is based off General Electorate votes, and with the Māori Party support coming from the Māori Electorates, which are much more heavily skewed towards younger voters than the General Electorates are, isn't representative of what's going on. That being said, given the skew in Māori Electorates towards younger voters, it probably suggests that this graph might be even more weighted to young voters. If there is one thing the Māori Party could take from this, and my earlier work, is that there's an opportunity with a big cohort of young Māori voters for them to win over between now and 2020.

The Opportunities Party is a bit of an interesting one in that because their vote was so heavily centralised around the Wellington region, it's likely the main influence on how this graph looks. TOP has done well with those aged between 18-29, and to a lesser extend those aged 30-34, and worst with those aged 40-49. National and TOP's support seems to switch at around age 39/40, while they bisect the rest of the parties between 49 and 54. Again, it's hard to read too much into these figures for TOP other than the fact that their deviation was the smallest among all parties.


As I said in looking at the possible influence of age on party voting preferences, it's very tempting to claim that these graphs show the likelihood of different age brackets voting for different parties - e.g. those aged 18-24 are 8% less likely to vote for National than the average New Zealander, and they're 11% more likely to vote for Labour. I think this data hints at that possibility, but without exit polling - which is illegal in New Zealand - it's impossible to know this for sure.

Where I think this data is very useful is using it to frame your thinking about where the parties position themselves in terms of their core support, and where they see the main battlegrounds are in terms of competing for votes from other parties. From around 30-years-old - where most of the parties graph lines start their journey towards converging on their national average - to the 50-54 bracket - where after that they diverge again, demonstrates I think that for the most part, the parties see voters within that 20 year age group - 30-years-old to 54-years-old - as the swing voters they need to target.

NZ First is the only really noticeable exception to this rule, but that's largely because Winston Peters has progressively clawed out those on NZ Super as his target voter base.

If you think generally about people in that 30 to 54 age range (and I'm talking very generally here) they're buying houses, getting married, having kids, they're likely to hit their career peak around between 40 to 49 (there's some US data around this, and sadly it has women's pay peaking at 40, and men's 49, highlighting again the gender pay gap). Retirement, while we're being constantly reminded about saving for it, is still a long way off, and the more immediate concerns are paying the mortgage or rent, affording school, doctors visits, dealing with health issues that become more and more likely to crop up, having a job, getting pay rises and getting ahead in life, and so on. Most have either finished up their travelling plans, or are about to, and are probably focused more on things like family or careers.

If you keep all this in mind, it starts to give you a bit of a picture of how and why political parties position themselves the way they do. In many respects they have to ensure their base votes for them, but they also have to reach out to that big segment of 30-to-54-year-olds to win their votes too.

The influence of age on party vote preferences

We all instinctively know that age has an influence on which parties people vote for. So I thought it'd be an interesting exercise to visualise the relationship between voter turnout by age and party votes in each electorate.

By using the age segments from my analysis of the youngest, oldest, most and least representative electorates, I've compared how electorates with above average shares of each of those segments voted relative to the overall party vote average (not the total party vote share) across the country.

The results are much as you'd expect, National is skewed towards voters aged 50 and over, Labour is skewed towards voters 49 and under. NZ First mimics National but with a much greater spread, and a similar situation presents itself for the Greens mimicking Labour. For the Māori Party I've only looked at General Electorate seats, largely because in the Māori Electorates there's not much of a difference between each of the electorates in terms of age segments. TOP I included only after I realised that their increased support in the 18-34 segment was interesting from the perspective that in Wellington Central and Ōhāriu, they would have been competing more with Labour and the Green Party for those votes, than they would have been with National or NZ First.

You could in theory argue that these numbers represent a likelihood of a certain age segment to vote a given way, but I think they're only indicative of that, as there are other factors at play, e.g. urban vs rural, affluent vs deprived, tertiary educated vs those without degrees, that will also influence these.

Even so, it helps contribute to our overall understanding of the electorate and how it voted this election.

In electorates with an above average share of Youth voters (those aged 18-34), National on average performed 11.01% worse (around 5.10 percentage points) than they did on average across the country. For Labour they performed on average 14.68% better in these electorates (5.25 percentage points) than nationally. The big differences though were NZ First who under performed here by 23.59%, and the Greens and Māori Party who over-performed by 20.87% and 22.18% respectively. As shown in the first graph, TOP's support, while having less of a spread than the other parties, over-performed in this segment.

Unsurprisingly, as our voters become older and we enter the Mortgage and Family segment, National still under-performs, but only by 2.88%, while Labour over-performs by 7.22%. NZ First and the Māori Party both still struggle in this segment (though I'd caution against assuming that these are representative of the Māori Party's support given the vast bulk of their support does come from the Māori Electorates). Interestingly, the Mortgage and Family was TOP's worst performing segment, coming in at 7.36% worse than there national average.

The Empty Nest segment - those aged 50-64 - is the first time we see National and NZ First over-performing their national average, by 7.07% and 14.73% respectively. Labour, the Greens and Māori Party all drop off from our prior segment as well. What I think is really interesting is how dramatic NZ First's support shifted between the Mortgage and Family segment and the Empty Nest Segment.

The Super segment - those aged 65 and over - sees NZ First get the lion's share of its votes. National, on the other hand, actually fairs slightly worse than it did in the Empty Nest segment, getting just 6.63% more of the party vote on average. What's also interesting here is that both National and Labour's support didn't move as radically as it did as we went through the other segments. National was obviously hurt by some support going to NZ First, but Labour doesn't seem to have suffered as you might expect they would looking at how the Green Party went. TOP's support, while still down on their national average, was better than their Empty Nest segment support though.

Hopefully I'll be able to do a bit more in-analysis of each of the age brackets that the Electoral Commission uses. But five graphs are easier to produce than twelve! 

Youngest, oldest, most & least representative electorates

Wellington and Otaki.png

Thanks to the Electoral Commission for their amazing work in producing election statistics, I'm able to present a few early tidbits as I breakdown the data that's now available around the age of people who actual voted. For ease of presenting the data for the youngest and oldest electorates I've broken down age groups into the following four segments:

  • Youth vote: 18-34-years-old
  • Mortgage and Family vote: 35-49-years-old
  • Empty Nest vote: 50-64-years-old
  • Super vote: 65+

For the most representative electorates - those that from an age perspective look the most like the average of all the electorates, I've looked at the average deviation is across all of the age groups specified in the Electoral Commission's data.

I've also separated out the General Electorates from Māori Electorates for the purposes of this analysis, largely because the Māori Electorates have quite significantly different age demographics than General Electorates as you'll see below.

Oldest General Electorates
(defined by having the largest percentage of voters aged 65 and over)

Ōtaki: 40.7% (my home electorate)

Coromandel: 40.0%

Northland: 34.8%

Tauranga: 32.8%

Rodney: 32.6%

Youngest General Electorates
(defined by having the largest percentage of voters aged 18-34)

Wellington Central: 43.9%

Dunedin North: 35.9%

Auckland Central: 34.2%

Māngere: 31.3%

Mt Albert: 30.9%

Empty Nest General Electorates
(defined as having the largest percentage of voters aged 50-64)

West Coast-Tasman: 33.7%

Northland: 33.2%

Helensville: 31.6%

Kaikōura: 31.4%

Waitaki: 30.9%

Mortgage and Family General Electorates
(defined as having the largest percentage of voters aged 35-49)

Mt Albert: 32.8%

Ōhāriu: 32.0%

Kelston: 30.4%

Northcote: 30.1%

Helensville: 29.8%

Least Representative Electorates
(defined by having the largest average deviation across all age groups)

Wellington Central: Ave dev of 3.91pp

Coromandel: 3.59pp

Ōtaki: 3.18pp

Northland: 2.96pp

Mt Albert: 2.83pp

Most Representative Electorates
(defined by having the smallest average deviation across all age groups)

Christchurch East: Ave dev of 0.42pp

Port Hills: 0.44pp

Rimutaka: 0.46pp

Tāmaki: 0.51pp

North Shore: 0.58pp

Māori Electorates

Youngest - Te Tai Tonga: 36.2%

Oldest - Waiariki: 13.8%

Mortgage and Family: Tāmaki Makaurau: 31.0%

Empty Nest: Waiariki: 27.8%

Most representative: Ikaroa-Rāwhiti: Ave dev of 0.42pp

Least representative: Te Tai Tonga: Ave dev of 0.82pp

Thoughts on all of this

It wouldn't surprise anyone, but out of the youngest electorates, all but Māngere appear in the top 10 electorates for Labour growing its party vote this election (remembering Māngere is already a pretty strong Labour seat anyway). With an average representation of Youth segment voters of 22.9% across the country, Wellington Central sits at nearly double that. Coromandel, Northland, Ōtaki, West Coast-Tasman and Kaikōura had the lowest levels in the Youth vote segment.

The same isn't necessarily true the other way though. For the oldest electorates, none of the top five are in electorates where National grew its share of the party vote (the Auckland mortgage beltway), however Rodney is one of National's most valuable electorates in terms of both the share of the party vote won and the high turnout in that electorate, and the other four all did deliver higher party vote shares than National achieved New Zealand-wide.

In a similar vein to before, Wellington Central, Mt Albert, Rongotai, Māngere, and Kelston had the lowest levels in the Super voter segment.

It's also worth noting the scale of  that both Ōtaki and Coromandel had 40.7% and 40.0% respectively of voters in the Super vote segment, with Northland coming in at 34.8%, this is against an average nationally of 24%.

However all five of the oldest electorates are in New Zealand First's top 20 electorates for party vote share. No surprise there given Winston Peters' relentless courting of this demographic.

In the Mortgage and Family segment, all bar Kelston (again having a strong Labour party vote anyway) experienced above average growth in Labour's party vote, with the Empty Nest segment imitates the Super vote segment, though to a lesser extent, with only West Coast-Tasman bucking the trend. There they delivering 5pp less in the share of party vote for National, which is split over slightly above what they achieved nationally for Labour, NZ First, and the Greens. The national average for Mortgage and Family was 25.4%, and for Empty Nest is was 27.8%, so neither of these segments produced as big a gaps as the Youth or Super segments.

On the flip side, the lowest Mortgage and Family segments were to be found in Coromandel, Ōtaki, Northland, East Coast, and Dunedin North. The five lowest Empty Nest segments were Wellington Central, Maungakiekie, Mt Albert, Northcote, and Dunedin North.

When you look at the Most Representative electorates, Port Hills and Christchurch East ranked 3rd and 9th for the biggest net swings fro the centre-right to centre-left. Rimutaka, Tāmaki, and North Shore all experienced above average swings to the left too.

In terms of the least representative electorates, Wellington Central and Mt Albert both saw Labour do better than they did New Zealand-wide, with their tilt in representation being skewed heavily towards the Youth and Mortgage and Family segments. Meanwhile Coromandel, Ōtaki, and Northland where the three highest electorates for the Super vote segment, with Northland and Coromandel both being over represented by the Empty Nest segment, and Ōtaki sitting around average. As you might expect, National and NZ First did better in these electorates while Labour and the Green Party under performed in them.

With the Māori Electorates one thing really caught my attention - how heavily skewed they were towards younger segments.

This is almost certainly because the gap between Māori and non-Māori life expectancy still sits at 7.3 years as of 2013. On average nearly a third of voters in the Māori electorates are in the Youth segment of 18-34, with another 29.7% in the Mortgage and Family segment, and 26.0% in the Empty Nest segment. The Super segment (65 years and over) made up, on average, only 12.2% of voters in Māori electorates. 

Māori Electorates are also notable for being very similar to each other in terms of the age demographics of their voters. Whereas for the General Electorates the average deviation was between 0.42pp and 3.91pp, for Māori Electorates the average deviation was between 0.42pp and 0.82pp,

Greying population and youth tremor sees middle age voter squeeze

It wasn't a youthquake, but a youth tremor and a greying population have seen New Zealanders aged 35 to 54 squeezed as a percentage of our voting population.

With youth enrolments down as a percentage of the overall potential voter population this election, a youthquake was always going to be difficult. However, better youth turnout numbers than in 2014 helped counteract the significant greying of voters in those three years.

The good news is that turnout was up across the board, ranging from 6.54 percentage points for those aged 18-24 down to a 0.6pp for those aged 70+. The bad news is that those aged 18-24 made up a lower percentage of overall enrolments than they did in 2014, meaning that their big increase in turnout didn't create a significant change from 2014.

Where the big change did occur was with voters aged 25-29, who increased their representation in the actual voter pool by 0.709pp versus 2014, the largest increase of any age group. Their increase of turnout of 5.45pp, combined with their larger slice of the enrolment pie,  and to a lesser extent those aged 30-34, delivered nearly all of the increase in youth voters this election.

The increase though was largely offset by an increase in voters aged 55 and over. With voters aged 70 and over representing 0.704pp more of voters than in 2014, and those aged 55-59 representing 0.240pp more, it largely balanced out the increases in the youth voter.

All of this means that those aged 35 to 54 saw their influence as voters decrease relative to 2014, which was partially driven by their contribution to the pool of enrolled voters also dropping relative to 2014. The biggest drop was in the 40-44 bracket where they contributed 1.026pp less to the final voter pool than they did in 2014.

You can find the Electoral Commission's numbers here.

Overall there was a marginal greying of voters, going from a median age of 47 in 2014, to 48 in 2017.

And in case you're curious, here's the raw numbers.

The Greyquake, and why NZ First, and maybe National, should be plucky

Projected Voter Turnout Rates.png

Greyquake, Superquake, Pensionquake, none of those names are as catchy as Youthquake, but unless there's a dramatic turnaround in voting behaviour by younger voters, this election is set to be dominated by those with, at the very least, a sprinkling of salt and pepper in their hair.

You can see in the above graph the difference between 2014's turnout and my projections for 2017's turnout. Assuming all things remain the same from the Electoral Commissions 12 September update of enrolment data, there's going to be a noticeable shift from the youngest half of voters to the oldest half of around 4.69 points.

That might not seem like much in the scheme of things, but in an MMP environment it's a massive shift. The difference between whether a party can feasibly form a government could be decided by as little as half a percent, so to have a shift of several times that towards an older demographic in the pool of electors will undoubtedly make a different.

As I wrote earlier today, the way that enrolments are tracking suggests that for all the efforts to encourage younger voters to participate in the election, nothing is changing. What is happening those is as New Zealand's population ages, the demographic of those enrolled to vote is sharply changing in favour of those aged 55 and over. The top half of enrolled voters goes down to around the 53 mark by my maths so you could stretch it a little further if you liked.

It's also why I suspect NZ First have historically done better on election day than they have in the polls. As far as I'm aware (and please correct me if I'm wrong!) both Colmar Brunton and Reid Research both weight their poll samples to reflect the general population. However the general population, or even the number of eligible voters isn't necessarily the best way to weight your polls.

Instead your polls should be weighted to reflect the voter turnout rates from the previous election (or even averaged off the past two or three just to hedge your bets a little). With older voters making up a far larger block of actual voters, their voices suddenly matter a whole lot more come election time than they do in the general population.

Projected changes in age cohorts.png

The above graph illustrates this change even better. You can see the massive shift away from youth voters towards those aged 55 and over (though as a mentioned earlier I suspect the tipping point may be around 53 or so). These are voters who probably own houses, probably still have mortgages they're paying off, their kids have probably left home or are about to, and they're planning for retirement.

What does this mean come Election Day? Assuming that voter turnout is similar to 2014 and enrolment rates don't change markedly, it means that NZ First is going to continue on its trend of performing better in the ballot than they do in polls, and a similar effect might help National as they've been focused on eating up some of NZ First's lunch too, especially in regional New Zealand.

It'll be interesting to compare these projects with the actual statistics from the election once they're released.