Understanding New Zealand 3: Who Voted ACT in 2020?

The ACT Party won a mere 13,075 votes in 2017, barely more than the joke parties. But in 2020 it won 219,031 votes. This 16-fold increase radically changed the composition of the ACT voting demographic.

Variable	Voting ACT 2020	Voting ACT 2017
European	0.74	0.16
Maori	-0.58	-0.51
Pacific Islander	-0.58	-0.23
Asian	-0.20	0.46

The main reason for the massive increase in ACT support from 2017 to 2020 was wealthy, old white people abandoning the National Party, but not abandoning the right wing. The correlation between being of European descent and voting ACT in 2017 was not significant, at 0.17. By 2020 this correlation had leapt to 0.74, which means that ACT now has the whitest supporters of any registered party, even whiter than New Conservatives.

Maoris were heavily disinclined to vote ACT in 2017 as well as 2020, and in 2020 the correlation between being a Pacific Islander and voting ACT was -0.51. This reflects the extent to which ACT policy disfavours the impoverished. Generally speaking, ACT appeals most to wealthy people who don’t want to be taxed.

Variable	Voting ACT 2020	Voting ACT 2017
Voting Labour same year	-0.12	-0.62
Voting National same year	0.92	0.61
Voting Greens same year	-0.08	0.17
Voting New Zealand First same year	0.18	-0.34
Voting Maori Party same year	-0.64	-0.42
Voting ALCP same year	-0.48	-0.52
Voting The Opportunities Party same year	0.18	0.03
Voting New Conservative/Conservative same year	0.68	0.04

That these new voters came predominantly from National is apparent when one looks at the strength of the correlation between voting National in 2020 and voting ACT in 2020: 0.92. This is much stronger than in 2017, when it was 0.61, or 2014, when it was 0.40. In 2020, ACT voters and National voters were from extremely similar demographics.

In fact, the correlation between voting ACT in 2020 and voting National in 2020 is so strong that the two voting demographics are close to identical. There were also strong correlations between voting ACT in 2020 and voting for the other parties whose demographics are wealthy, old and white (i.e. enfranchised), such as New Conservative (0.68) and Sustainable NZ (0.54).

Unsurprisingly, then, there were strong negative correlations between voting ACT and voting for the young, poor and brown parties, such as the Maori Party (-0.64) Vision NZ (-0.60) and ALCP (-0.48).

Variable	Voting ACT 2020	Voting ACT 2017
Aged 20-24	-0.51	0.18*
Aged 25-29	-0.44	0.18*
Aged 30-34	-0.34	0.36**
Aged 35-39	-0.22	0.36**
Aged 40-44	0.15	0.36**
Aged 45-49	0.58	0.36**
Aged 50-54	0.73	0.17***
Aged 55-59	0.78	0.17***
Aged 60-64	0.79	0.17***
Aged 65-69	0.77	0.11****
Aged 70-74	0.76	0.11****
Aged 75-79	0.74	0.11****
Aged 80-84	0.68	0.11****
Aged 85+	0.63	0.11****

The ACT Party also got much older in 2020. In 2014, the correlation between median age and voting ACT was 0.02. By 2017, it had increased to 0.26. By 2020, it had increased to 0.54 – stronger than the correlation between median age and voting National that year.

Most notably, the correlation between voting ACT in 2017 and being aged 65+ was 0.11, but the correlations between voting ACT in 2020 and belonging to any age bracket above 65 were all at least 0.63. The ACT demographic of today is much, much older than the demographic of even a few years back. Whether this reflects a permanent shift or just a temporary change in sentiments remains to be seen.

Variable	Voting ACT 2020	Voting ACT 2017
No qualifications	-0.12	-0.73
Level 1 certificate	0.20	-0.59
Level 2 certificate	0.05	-0.50
Level 3 certificate	-0.66	0.25*
Level 4 certificate	0.09	0.25*
Level 5 diploma	0.02	0.50**
Level 6 diploma	0.79	0.50**
Bachelor’s degree	0.03	0.70
Honours degree	0.16	0.58
Master’s degree	0.03	0.65
Doctorate	0.11	0.51

The easy assumption up until now was that the ACT Party appealed to a younger, more educated and more liberal demographic than National. This assumption used to be accurate, but by 2020 it no longer was. The ACT Party got so many votes from core National supporters that the two voting blocs are barely distinguishable when it comes to age, race, education or wealth.

By 2020, the average ACT voter was not significantly more likely than the average person to hold a university degree. This was a massive change from 2017. The correlation between holding a bachelor’s degree and voting ACT collapsed from 2017 (when it was 0.70) to 2020 (when it was 0.03). The correlations for other degrees fell by lesser, but still large, amounts.

In 2017 it was highly unlikely that a person with no NZQA qualifications would vote ACT – the correlation between the two was -0.73. But by 2020 that correlation had come in to -0.12. 2017’s ACT was heavily disproportionately supported by educated people. 2020’s ACT was much closer to broadly representative of the various education levels.

Variable	Voting ACT 2020	Voting ACT 2017
Living in an urban electorate	-0.23	0.37

One of the main reasons for the increase in ACT support from 2017 to 2020 was their support of firearms rights. Many of the new ACT voters were rural firearms enthusiasts. This is evident from the fact that the correlation between living in an urban electorate and voting ACT switched from a significantly positive correlation in 2017 (0.37) to a borderline significantly negative correlation of -0.23 in 2020.

Variable	Voting ACT 2020
Working in agriculture, forestry or fishing	0.43
Working in mining	0.20
Working in manufacturing	-0.15
Working in electricity, gas, water and waste services	-0.10
Working in construction	0.11
Working in wholesale trade	-0.05
Working in retail trade	0.03
Working in accommodation and food services	-0.08
Working in transport, postal and warehousing	-0.58
Working in information media and telecommunications	-0.21
Working in financial and insurance services	-0.04
Working in rental, hiring and real estate services	0.47
Working in professional, scientific and technical services	0.06
Working in administrative and support services	-0.65
Working in public administration and safety	-0.22
Working in education and training	-0.15
Working in healthcare and social assistance	-0.07
Working in arts and recreation services	0.01

Fitting with the high level of rural support for ACT are the significant positive correlations of 0.39 between voting ACT in 2020 and voting Outdoors NZ Party in 2020, and of 0.43 between voting ACT in 2020 and working in agriculture, forestry or fishing. There were also positive correlations, if not significant ones, between voting ACT in 2020 and working in mining or construction.

The most striking correlation here is the one of -0.65 between voting ACT in 2020 and working in administration and support services. This might surprise many, because ACT voters are so urbanised that one could expect heavy representation in industries that are typically urban, such as any office work.

The explanation is that ACT appeals mostly to those willing to take financial risks and to gamble, and so they tend to choose more entreprenurial industries. The choice of administration and support services is usually made by those who like to play it safe.

It is striking that such a strongly historically urban party as ACT might get more support from rural electorates in 2020 than urban ones. This speaks to the sense of betrayal that the right-leaning firearms community felt about National supporting restrictive firearms legislation. Almost all of these new, rural ACT voters will have been National voters in the previous election.

In several ways, the correlations between belonging to certain demographic categories and voting either ACT or National in 2020 are identical. Voting for either party had a correlation of 0.17 with casting a special vote for Yes in the euthanasia referendum, one of 0.58 with being aged 45-49 years old, one of -0.60 with voting for Vision NZ in 2020, and one of 0.68 with voting for the New Conservative Party in 2020.

Although the two parties have many shared sentiments, they have slight differences in some other ways.

Variable	Voting ACT 2020	Voting ACT 2017	Voting National 2020
Median income	0.20	0.61	0.23
Mean income	0.23	n/a	0.24
Median age	0.54	0.26	0.42
Mean age	0.53	n/a	0.39

Measured by income, the average ACT voter in 2020 was about as wealthy as the average National voter in 2020. This was a sharp movement towards the middle for the average ACT voter. In 2017, the correlation between voting ACT and median income was 0.61. By 2020 it had fallen so far that it was no longer significant.

The median age of an ACT voter, by contrast, increased sharply between 2017 and 2020. In 2017 the correlation between median age and voting ACT was barely significant, at 0.26. By 2020 this correlation was much stronger, at 0.54. In fact, this correlation was so strong by 2020 that it was stronger than the one between median age and voting National.

If National is an old person’s party, that’s now even more true of ACT. This marks a striking change from the young professional image that ACT usually projects.

Variable	Voting ACT 2020	Voting ACT 2017
New Zealand-born	-0.01	-0.56

In stark contrast to earlier years, when it was possible to write of ACT that they had the lowest proportion of New Zealand-born voters of any party, the correlation between voting ACT in 2020 and being New Zealand-born was -0.01. This is because the vast majority of their new voters were elderly and rural, and those demographics tend to be New Zealand-born.

These Kiwis did not come to favour ACT for nationalist reasons, however. Most of their sentiments were driven by anger over the incompetence of both Labour and National, who both supported heavy restrictions on firearms rights in the wake of the Christchurch mosque shooting. In the wake of these restrictions, ACT gained heavily from the perception of being a libertarian party.

Variable	Voting ACT 2020	Voting National 2020	Voting ACT 2017
Employed full-time	0.01	0.04	0.22
Employed part-time	0.50	0.27	0.07
Unemployed	-0.84	-0.82	-0.42

One notable difference between National and ACT voters is that the former are less likely to be employed part-time. The correlation between voting National in 2020 and being employed part-time was 0.27 – for voting ACT in 2020 it was 0.50. This speaks to the degree to which the ACT voters of 2020 value community engagement – in stark contrast to earlier years.

This also reflects the fact that ACT voters were much older in 2020 than in 2017, as older people frequently cut down from full-time to part-time work in their 50s and 60s rather than become unemployed or retired. So if a demographic has a strong positive correlation with working part-time, that demographic is probably full of the sort of person who volunteers for the Rotary Club or similar institutions.

However, the essential aspirationism of the average ACT voter is as strong as ever. The correlation between voting ACT in 2020 and being unemployed was -0.84, one of the strongest negative correlations in this study. ACT voters are definitely not the sit-on-the-couch type, even less than National voters are.

Variable	Voting ACT 2020	Voting National 2020	Voting ACT 2017
Work as manager	0.81	0.78	0.39
Work as professional	0.00	0.09	0.53
Work as technician or trades worker	0.08	0.05	-0.43
Work as community or personal service worker	-0.46	-0.61	-0.44
Work as clerical or administrative worker	-0.23	-0.01	0.05
Work as sales worker	-0.38	-0.18	0.06
Work as machinery operator or driver	-0.43	-0.43	-0.71
Work as labourer	-0.14	-0.29	-0.60

In 2017, ACT appealed roughly equally to managers and professionals. The correlation between working as either and voting ACT that year was significant but not particularly strong. By 2020, there was no longer any significant correlation between voting ACT and working as a professional. The correlation between voting ACT and working as a manager, on the other hand, become one of the most strongly positive correlations of any ACT supporter: 0.81.

The changes in ACT support from 2017 to 2020 reflect that much of their new support came from the sort of person who would be most interested in firearms rights. Occupations such as technician or trades worker, machinery operator or driver and labourer are typical among the firearms enthusiast community, and the correlations between all three and voting ACT became much less negative between 2017 and 2020.

Congruent with the perception that ACT voters tend to lack empathy, few ACT voters work as community or personal service workers. The correlation between working in this occupation and voting ACT in 2020 was -0.46. Votes from this demographic tend to go to the Labour Party.

Variable	Voting ACT 2020	Voting ACT 2017
No religion	0.37	0.07
Buddhism	-0.12	0.55
Christianity not further defined	-0.03	-0.12
Hinduism	-0.36	0.12
Islam	-0.39	0.10
Judaism	0.14	0.71
Maori religions	-0.49	-0.47
Spiritualism and New Age	0.05	-0.22

There was also a religious component to the pattern of ACT votes.

The correlation between having no religion and voting ACT in 2020 (0.37) was notably stronger than the correlation between having no religion and voting National in 2020 (0.16). ACT voters are less likely to be Hindus or Muslims by a similar margin. This speaks to how National has always pandered to Establishment religious sentiments whereas ACT has not.

In 2017, when ACT was a much smaller party, its voters were much more likely to be educated. This explains the strong correlations between voting ACT that year and being a Buddhist (0.55) or being a Jew (0.71), as those are two of the best-educated demographics in the country. By 2020, being a Buddhist or a Jew was much less likely to predict support for ACT.

Realistically, any party that moves from the fringe to significant electoral success will move towards the centre in most demographic measures. This is evident in many ways if one compares ACT voters in 2017 to ACT voters in 2020.

Variable	Voting ACT 2020	Voting National 2020	Voting Greens 2020
Being male	0.07	-0.02	-0.16

The correlations between voting ACT or Green in 2020 and being male are not particularly strong, at 0.07 and -0.16 respectively. But given the sizes of the demographics in question here, the differences are noticable. It’s not enough to say that ACT is for men and the Greens are for women, but future elections could easily entrench this division as the alternative continues to win votes from the Establishment.

In summary, from 2017 to 2020 ACT transformed. In 2017 they were a fringe party for high-income, low-empathy voters. By 2020 they had become a mainstream movement with the potential of challenging National as the de facto leader of the right wing. This was mostly due to a massive influx of old, white, rural voters.

The Four-Fold Rule that ACT appeals mostly to young men, in contrast to National’s appeal to old men and the Greens’ appeal to young women, mostly didn’t apply in 2020. The ACT Party muscled in on some of National’s traditional territory, winning many votes from older people.

The question for 2023 is whether ACT can achieve what parties in the alternative right dream of everywhere: to win enough young men to take power. It’s unlikely that ACT will win many voters from the non-voter demographic, as those tend to be poor and non-aspirational. But if ACT continues to win votes from National, ACT could become the larger of the two, and thereby the official default right-wing party.

*

This article is an excerpt from the upcoming 3rd Edition of Understanding New Zealand, by Dan McGlashan and published by VJM Publishing. Understanding New Zealand is the comprehensive guide to the demographics and voting patterns of the New Zealand people.

*

If you enjoyed reading this essay, you can get a compilation of the Best VJMP Essays and Articles of 2019 from Amazon for Kindle or Amazon for CreateSpace (for international readers), or TradeMe (for Kiwis). A compilation of the Best VJMP Essays and Articles of 2018 and the Best VJMP Essays and Articles of 2017 are also available.

*

If you would like to support our work in other ways, please consider subscribing to our SubscribeStar fund. Even better, buy any one of our books!