Analysis by Keith Rankin.
We are learning that, in the absence of high levels of natural immunity to Covid19, New Zealand will need extremely high vaccination rates before it can repudiate its growing ‘hermit kingdom’ status.
(It is important to note that the expression ‘hermit kingdom’ dates to a book title from 1882 – ‘The Hermit Nation’, by William Elliot Griffis – referring to Korea under the Joseon dynasty, and in particular during the centuries [17th to 19th] of the isolationist Tokugawa Shogunate in Japan. The expression ‘hermit kingdom’ is more of a Buddhist reference than a reference to cold war communism, and has been applied in more recent times to Bhutan of ‘happiness’ fame. The best example of a substantial self-isolating hermit nation today – Hillary Clinton’s reference notwithstanding – is not North Korea; it is Turkmenistan. The most complete contemporary example, though, would be Abkhazia.)
To achieve vaccination rates in excess of 90 percent, we need clearly defined population (and sub-population) denominators. Ninety percent of what? As a matter of definition, we have wavered from 90% of people aged over 16, aged over 12, aged over 5, and aged over 0. Further, while we have a reasonable idea of what our nationwide populations are for each age and ethnicity cohort, we have much less idea of where they are.
The five-yearly census – run since the 1860s in Aotearoa New Zealand – is meant to be a means of calibrating our population data, which otherwise are based on estimates. The 2018 census was an extreme case of us losing sight of this principal reason for holding a census; instead, it became a large-sample survey, not very helpful because the accuracy of survey data depends more on a lack of sample bias than it does on sample size.
We know that in the 2013 census Māori communities in Northland were substantially undercounted. The lack of trust in government agencies indeed goes back several decades. The methodological problems with the 2018 census were already well-established, and almost certainly date back as far as 1991 when, at census time (at least from the 1930s), New Zealand first had highly marginalised population sub-groups. (The 1931 census, during the Depression, was cancelled as an economy measure; indeed, an unfortunate false economy measure.) Inequality as well as unemployment increased dramatically between the 1986 and 1991 censuses; income inequality kept increasing through the 1990s. Further, the 1991 census was politicised, through its radical redefinition of the way unemployment was counted, making its labour force data unable to be properly compared with earlier census data, and substantially understating the actual extent of labour market failure.
There are almost certainly significantly fewer people living in Auckland today than most policymakers believe to be the case. This is even more so for Māori, and is an important part of why published Covid19 vaccination rates for Māori in Auckland Tāmaki Makaurau are so low.
This should not be a surprise, because the big story in town from 2012 to 2017 was the real estate bubble. (This bubble followed a similar event from 2004 to 2008, despite these being years of falling net immigration and rising interest rates in New Zealand.) The major demographic impact of the dramatic 2010s’ decade house price inflation was that of people selling houses – realising capital gains – and moving to the hinterland (including a few heading for the South Island hinterland). Not only did many people leave Auckland, but most of us knew they were leaving Auckland. There were more than just anecdotes, there were television programmes highlighting the Auckland exodus. And, no, international immigration was not nearly as focussed on Auckland as most people, especially the mayors of Auckland, imagined.
The biggest gap in our demographic data is our lack of accurate sub-national population estimates. Instead of correcting our increasingly unreliable locality population estimates, in order the fill-out the 2018 census, we resorted to much of these uncalibrated estimates. In principle we know how many people are in New Zealand, and their ages and their preferred ethnicities; but even that was doubtful due to the very sloppy way Statistics New Zealand used to collect immigration data (as a casual add-on to tourism data, based on how people completed exit and entry cards). But we know very little about net migration within New Zealand; that includes people having left New Zealand for a while and later resettling in a different part of New Zealand.
Electoral data can be a useful way to check on population patterns. In 2014 and 2017, the same electoral boundaries were used. And more than half of these boundaries were retained for the 2020 and 2023 elections. New Zealand has both general electorates (aka constituencies) and Māori indigenous population electorates.
Electorate boundaries are redrawn every five years or so, based on census data. Using a slightly bizarre format which fixes the number of South Island general electorates. A population quota is then calculated for all electorates, based on the population in the South Island who do not identify as Māori. The quota is then applied to the North Island ‘general’ population. As a result, one more North Island electorate was established in 2020, and a certain degree of population reshuffling was required; for 2020, this reshuffling mainly affected the Waikato region, but also affected Northland. Both shuffles impacted on Auckland; indeed the new electorate was inserted into South Auckland.
Nevertheless, more than half of Auckland’s electorates did not experience boundary changes. Further, the quota system allows for an electorate’s estimated population to be up to five percent below or above quota. The Auckland electorates which did not experience boundary changes found themselves below the quota; in some cases, just within the allowable five percent margin. These unchanged electorates are: East Coast Bays, North Shore, Northcote, Te Atatu, Kelston, Mt Albert, Auckland Central, Epsom, Tāmaki, Pakuranga, and Mangere. Maungakiekie had its boundaries unambiguously extended in 2020, as did the Māori electorate Tāmaki Makaurau. New Lynn and Mt Roskill did sideways shifts to accommodate increased population in Northland and the North Auckland rural (and semi-rural) fringe. A new electorate was created in Takanini, South Auckland, allowing the bloated electorates further south to become a bit smaller.
To the constant surprise and frustration of Aotearoa New Zealand’s Māori elite, not enough Māori opted to go on the Māori electoral role. This meant that the number of Māori electorates stayed at seven. The target population size for Māori electorates is the same as for general electorates, though the average population base for the seven Māori electorates may necessarily diverge by more than five percent compared to the general electorates.
To effectively use electorates as locational demographic indicators, we need to used votes cast as a percentage of nationwide votes. In New Zealand, we can do the exercise separately for Māori and general electorates.
In 2008, 14.6% of Māori votes were cast in Tāmaki Makaurau (Auckland); over one-seventh of all Māori votes. In 2011, that fell to 14.4%. The ‘winner’ in 2011 was Te Tai Tokerau (the north) and the loser was Te Tai Tonga (the south, essentially Wellington), reflecting a loss of government employment in Wellington following the election of a National-led government in 2008.
In 2014, Tāmaki Makaurau was reduced in size, and its share of the Māori vote accordingly fell to 13.8%. Then, with the same boundaries as 2014, Tāmaki Makaurau fell to 12.6% of the Māori vote. Te Tai Tokerau stayed top, and Te Tai Tonga increased its share, reflecting a degree of restoration of Māori employment in the Wellington bureaucracy.
In 2020, Tāmaki Makaurau was enlarged in size (reclaiming many South Auckland voters from Hauraki-Waikato), and got 14.1%, while Hauraki-Waikato slumped to 13.2% of the Māori vote. While the combined Auckland-Waikato Māori vote did show a small increase after 2017, evidence from the general electorates indicates that this increase was in Waikato, not in Auckland.
In 2020, Te Tai Tonga increased again, to 15.8% of the Māori vote; reflecting renewed opportunities for Māori employment in Wellington’s government nexus. Te Tai Tokerau, with unchanged boundaries, fell from 15.2% to 14.7%. This last shift would appear to be a reversal of the 2011 effect, and also a shift of Māori out of West Auckland and into the far north iwi homelands (and Whangarei) where houses are less expensive.
As a recently retired lecturer in Unitec’s School of Applied Business, I can affirm that, while Pasifika clearly represented an increasing proportion of student enrolments, Māori enrolments showed a steady decline in the 2010s. This was despite excellent, indeed improved, Māori student support on campus. The most obvious, and almost certainly correct, explanation was that the Māori population within Auckland has diminished; and from an already low percentage base. About 25% of all non-Māori New Zealand residents live within the borders of the Tāmaki Makurau Māori electorate; yet even in its restored and enlarged form, Tāmaki Makaurau only registered 14.1% of Māori votes in 2020, down from 14.6% in 2008.
The total Māori population in Auckland may be ninety percent (or even less) of what the Ministry of Health believes it to be. Thus, even an official 75% vaccination rate of Māori in Auckland might in reality be 90%. To get the nationwide Māori vaccination rate up to ninety percent, Auckland will not be the place to go looking for the unvaccinated.
The General Population
From an electoral point of view, nearly half of New Zealand residents with Māori ancestry have chosen to identify as ‘general’ rather than as Māori. While ten percent of electorates are Māori, 17 percent of the population is Māori.
The 11 Auckland electorates (listed above) with unchanged boundaries, cast 14.6% of all general electorate votes in 2020, while representing 15.3% of general electorates. In 2014, these electorates cast on average 15.44% of votes, while representing 15.49% of all electorates. In 2017, 15.03% of votes were cast in these places, which represented 15.49% of all electorates. This suggests that the relative decline in Auckland’s general population persisted in the 2017 to 2020 period.
Of particular note was the Maungakiekie electorate (not one of the 11 previously mentioned), which cast 1.54% of all general electorate votes in 2014, 1.47% in 2017, and 1.38% in 2020 despite having its boundaries enlarged. This is a part of Auckland which once had a highish Māori population, but has now been substantially gentrified, and emptied.
Maungakiekie is one of the electorates which represent a north and west block of eight which have had boundary shuffles. While together, their boundaries are unchanged, Maungakiekie has unambiguously enlarged, whereas Upper Harbour and Whangarei have unambiguously shrunk. Upper Harbour’s share of the total vote has been consistent at 1.50% of the total, or just slightly under. Whangarei scored 1.63% in 2014, 1.71% in 2017, and 1.63% with reduced boundaries in 2020. Whangarei has clearly grown in its population share, just as Auckland has fallen in its share.
As well as Northland, the other area that has clearly grown much faster than the national average is Auckland’s ‘southern hinterland’ outside of the urban electorates of Hamilton East, Rotorua and Tauranga. Indeed this southern hinterland – a mix of South Auckland, Waikato, Bay of Plenty – contains the one new electorate arising from the 2020 boundary reshuffle. This zone of population growth particularly represents the outskirts of Hamilton and Tauranga, Coromandel, and Auckland’s southern fringe. Mangere and Pakuranga, in the south of Auckland, have unchanged boundaries. Their voter statistics represent the Auckland pattern, not the Waikato pattern. So it seems safe to conclude that, within this ‘southern hinterland’, the population growth has been greater to the south of core South Auckland, and not in core South Auckland.
Where to Vaccinate
While this demographic analysis of Auckland and its surrounds is important for a number of reasons, my main conclusion here is that, if we want to get our Covid19 statistics up towards a rate of ninety percent of people aged over five – and especially Māori statistics – then we should be looking for people in Northland and greater Waikato, and not looking for people in Auckland who are just not there.
And, to the mainstream media, whenever you report percentages, don’t forget to specify the denominators. Ninety percent of what? And ask that the denominators suggested by officials be quantified. (And remember that, when commentators refer to ‘Hermit Kingdom’, they may not be referencing modern North Korea.)
Keith Rankin (keith at rankin dot nz), trained as an economic historian, is a retired lecturer in Economics and Statistics. He lives in Auckland, New Zealand.