Last month I noted that Māori voter growth in Auckland (Tāmaki Makaurau) was slower than in all the other Māori electoral districts, and that this almost certainly reflected very low Māori population growth in Auckland. I also argued that Pakeha as well as Māori were leaving (or not arriving in) Auckland in larger numbers than in the past. And I said that the census data would show that Auckland’s population (excluding the former Rodney and Franklin Districts) would be less than 30 percent of the national total.

This month’s chart comes the day after the census data was released. Because the infill process – filling in for census non‑participation – was more comprehensive than for past censuses, the population growth data overall may be slightly inflated. But, on the whole, these growth data are probably as reliable as in the past.

The chart shows that, in Auckland’s inner isthmus suburbs (from Mt Albert and Point Chevalier through to St Heliers, and including the central business district), population growth in 2013‑18 was comparable to Palmerston North and Dunedin. This is in contrast with 2006‑13 when central Auckland grew by ten percent. We may note that residential property prices in these years grew much faster in central Auckland than Dunedin, suggesting that population pressure was not the driving force behind these rising prices. Rather, people responded to inflated property prices by leaving inner Auckland (or not coming).

The rest of the isthmus grew faster in 2013‑18, compared to inner Auckland and compared to 2006‑13. But the growth was not remarkable; it was comparable with Napier and Nelson. Likewise the ‘outer core’ – which includes much of North Shore and Manukau had population growth after 2013 that was well below the New Zealand average of 10.8 percent.

Auckland’s ‘less unaffordable’ outer suburbs grew faster. The ‘outer’ axis includes areas like Henderson and Massey in the northwest, and Manurewa, Otara and Howick in the southeast. In those areas, population growth was comparable with Hamilton.

In the super‑’city’, the fastest growing areas were the fringe – Papakura, Upper Harbour, Hibiscus and East Coast Bays, and Waitakere Ranges – and Rodney and Franklin. Growth there was comparable with Northland (which includs Whangarei) and Bay of Plenty (which includes Tauranga).

The census data confirms the election data, showing a substantial demographic shift from core Auckland to (and beyond) Auckland’s fringe. Although Auckland City is growing relatively slowly, the historic Auckland Province now holds 54.6 percent of New Zealand’s population, up from 53.8 percent.

Auckland excluding rural Rodney and Franklin had 30.4 percent of New Zealand’s resident population. This includes the ‘Hibiscus’ area of Orewa which was in Rodney District before the super‑city was formed in 2010. So it looks like I am correct, the combined population of the former Auckland, North Shore, Waitakere, Manukau and Papakura – what most people have thought of as Auckland – is below 30 percent of the national total. The rest of the country continues to have much life and soul.

Auckland’s 2012‑16 real estate bubble was just that, a bubble. It had the substantial consequence of revitalising much of the rest of the country, though creating many of the same social problems as Auckland (especially relating to unaffordable rental housing) in many other cities.

As a final note, Auckland’s unoccupied dwelling count increased by over 6,000 to 39,393. 97 percent of the national increase in empty dwellings was in Auckland. And Auckland province registered a 9,363 increase in empty properties, 151 percent of the New Zealand net increase of 6,201 unoccupied homes.

by Dr Bryce Edwards

New Zealanders should have less faith in many of the vital decisions of government agencies, due to serious problems with the way last year’s census was carried out. It’s a multi-billion-dollar problem. And the Government and officials appear to be doing little to assuage concerns. In fact, the recent decisions of the Minister of Statistics and Statistics New Zealand are giving the public more reason to be suspicious.

I covered this a month ago, on the anniversary of the 2018 census, pointing out the many problems – see: The absolute debacle of the 2018 Census. The “digital first” experiment had gone horribly wrong, leading to a suspicion at the time that about one-in-ten citizens hadn’t completed the census – something that would lead to all sorts of problems for government decision-making. Since then, the full extent of the debacle has become clearer, but accountability has not.

Most significantly, despite Statistics NZ’s reluctance to let the public know, it’s now clear that there was also a problem with those who did fill out the census forms – about 240,000 never finished filling in the forms. This is on top of the 480,000 who didn’t even start filling them in.

Yesterday’s Otago Daily Times editorial quite rightly labels the most recent census a “a failure of epic proportions”, saying that “Long delays in the release of information rang alarm bells, but that one in seven New Zealanders have not had their information recorded is truly disheartening” – see: When it all goes horribly wrong.

The editorial explains why it’s such a problem: “This matters because it means the census is going to deliver skewed and incomplete results. The census, held every five years, provides vital information – a snapshot of the nation – that is used primarily for allocating funding for core services, for future planning, for the drawing of electoral boundaries, for so many things. It must be done, and it must be done right.”

There are billions of dollars at stake, because so much of what the government spends is affected by the detailed knowledge of New Zealand society that the census produces. Financial journalist David Hargreaves explains this in his column, Maybe we should scrap the 2018 census and start again.

Here are his examples: “In the case of migration, obviously if thousands less people have come into particularly Auckland this has big ramifications for housing and infrastructure. Having pristine statistics is not some ‘nice to have’ fanciful thing. It’s actually absolutely vital and therefore should not be subjected to Government penny-pinching”.

Although Statistics NZ is currently attempting to “backfill” the gaps in the census data with a “patching” process that uses other government agency statistics, there doesn’t seem to be much confidence amongst commentators that this will be adequate.

Hargreaves discusses this: “however well they do in patching this up, there will, always be a suspicion over the information and whether it is of the quality on which big spending Government decisions on things like infrastructure can be confidently committed. Where we want information as gold, we might end up with copper. Which ain’t good enough. Anything where information has to be extrapolated, say from a sample size, is always subject to some question marks.”

Similarly, Massey University’s pro vice-chancellor Paul Spoonley is unconvinced that the patching process will work, saying “I’m still very, very sceptical” – see 1News’ Massey University official calls on Stats NZ head to ‘front’ on ‘failed’ 2018 Census.

According to this report, Spoonley believes it still “leaves the serious question of whether New Zealand can have confidence that the replacement data will be correct”. Spoonley says, in terms of government decision-making, “a whole lot of systems are actually breaking down because we don’t have Census data available”.

For the best discussion of the quality of the data, see David Williams’ Census ‘disastrous’, but not useless. Of particular interest in this article, is the evaluation of Richard Arnold, a former Statistics NZ analyst and now a statistics lecturer at Victoria University of Wellington. He says the latest census is “absolutely not up to spec”.

Arnold explains the importance of a full national census over simple sample surveys, saying a census “gets in everywhere” in terms of the smaller demographic groups. And because only about six out of seven citizens have been properly recorded in the census, all sorts of minority groups will have a reduced profile in the snapshot of the nation.

He says: “It’s interesting to see how they are growing and changing over time. There’ll be people from minority religions who want to see how many they are and how they’re changing. There are little groups everywhere. Even regional areas, where people want to know how the rural population is changing.”

There will now be significant errors and bias in the information the government is relying upon. Whereas very small amounts of missing information can be patched up, in this case Arnold suspects “when it gets out to this kind of level, it’s a really big problem” and “You just don’t know because you haven’t measured”.

Similarly, economist Brian Easton lacks confidence in the ability of Statistics NZ to patch up the problems, especially given what he’s learnt about the process: “We have been told that the patching will not all cover all questions which researchers use, while the patching is likely to invalidate many of the issues which researchers want to explore” – see Thomas Manch’s One in seven failed to complete Census 2018, a back down from Govt Statistician reveals.

According to this article, “Easton said the gap in census response was now so large the data may be useless for research.” National’s state services spokesman Nick Smith says the problems with the census data “will create problems for years in allocating tens of billions of dollars in funding”. While New Zealand’s chief statistician Liz MacPherson is reported complaining that “misunderstandings have resulted in unfounded comments regarding the integrity of the official statistics system and Stats NZ.”

MacPherson has been in the firing line for her role overseeing the whole post-census debacle. And she’s come under even more scrutiny recently due to her extreme attempts to withhold information from the public and politicians about the unusually high number of people who only partially completed their census forms. This is best covered in Thomas Manch’s article, Chief Statistician ordered by MPs to produce information on Census 2018.

As head of Stats NZ, MacPherson has had to appear twice this year at a select committee, where both times she has declined to provide details on the census completion rates, stating last week “Without the appropriate context, these individual numbers would be open to misinterpretation”. Commenting on this, constitutional expert Andrew Geddis says “I can’t remember a time a public servant has refused after being told they must answer”.

MacPherson eventually yielded, after the select committee MPs unanimously decided to threaten her with contempt of Parliament. But the episode was enough for blogger No Right Turn to call for her to be sacked, saying her disobedience was an alarming defiance of the public’s democratic rights and of accountability – see: This is not how our public service should work.

According to Friday’s Press editorial, the whole episode needs more investigation: “Statistics Minister James Shaw’s claims that the gaps in the data are not a concern, and a chief statistician who was almost dragged kicking and screaming by politicians to present a fuller picture of a shambolic process and a poor result, are both worthy of greater scrutiny” – see: NZ census: A black hole of big data.

The editorial joins a host of commentators and experts who are calling on the Government to fix the problems of last year’s census debacle by bringing forward a new census for 2021 – something that James Shaw will not countenance.

And the blame game will continue – with arguments about whether the census debacle was due to initial decisions made by the last National Government, or mismanagement under the Labour-led administration. But regardless, Oscar Kightley wrote on Sunday, it’s Shaw’s “job to fix it, and he could start by admitting that this is an embarrassing disaster that could have serious consequences for the country” – see: Why the census debacle is so serious.

There will also be consequences for some of the 700,000+ who failed to start or finish completing the census last year. Jono Galuszka reports that, so far, “60 court cases were being lodged in relation to people not completing the census” – see: Stats NZ starts taking people to court over non-completed census forms. But not everyone is being prosecuted: “Criteria for prosecution included actively refusing to do the census, being strongly negative, or being abusive towards census staff”.

Finally, although many are calling for a return to a paper-based census process, is it actually time to further embrace digital data? In the age of “big data” there are a huge number of other possibilities for innovation in terms of measuring and understanding what’s going on in New Zealand. With the use of greater technology, government policy making could actually become much more reliable and effective according to Pattrick Smellie, who asks: In a world full of big data, is it time to rethink the Census altogether?

Exactly a year ago, on 6 March 2018, the government census was carried out. It was a “digital-first” census, with citizens expected to primarily use the internet to answer the compulsory questions about their lives. And it was soon apparent that huge numbers of New Zealanders had been missed out by the exercise – likely to be at least one in ten. 

But we are only now finding out how much of a debacle that event was, along with increasing knowledge of the seriously negative implications for our society and democracy. And to make matters worse, there’s a distinct lack of political or bureaucratic accountability for what has happened.

A number of newspaper stories have been published today about the anniversary of the 2018 census debacle. The most important is Thomas Manch’s 365 days and still counting: Census 2018 results nowhere to be seen. In this, he explains that “The 2018 census data release has been delayed three times due to low response rate” and “Statistics New Zealand remains tight-lipped about when the long-delayed results will be available”.

The article explains that the census operation resourcing was inadequate, especially in terms of the number of field staff hired to help people get their census filled out: “Newly released information shows Stats NZ employed 1800 enumerators, or field staff, to knock on doors and uncover those who failed to complete Census 2018. This was a substantial drop from the 7000 boots on the ground during Census 2013”.

A number of economists and statisticians explain the growing awareness of the severity of the “shambolic” census operations. For example, University of Auckland statistician Andrew Sporle is quoted saying: “It’s a bit of a disaster, we don’t know how bad, but we know it’s a disaster.”

Sporle explains that the data, once it’s eventually released, “will provide increased uncertainty in matters from the Government’s wellbeing targets to the number of Māori electorates.”

Manch looks at the Māori electorates in another article today, Māori electorate seat at risk due to Census 2018 debacle, and says that, due to the way Statistics NZ have run the census, there are likely to be much fewer respondents reporting that they are of have Māori descent, which could lead to one of the Māori seats disappearing.

One specialist is cited as saying that although the 2018 census participation rate is estimated to be about 90 percent for the general population, “the response rate of Māori may be as low as 80 or 70 per cent in some corners of New Zealand”. In addition, “more than 20 new iwi won’t be properly counted” in the census, and this is a problem because “census data was particularly important for smaller iwi trying to do good with fewer resources”.

Societal consequences of the census debacle

There are plenty of other serious implications if the census 2018 data is unreliable, as is increasingly expected. For example, planning and funding for health and education is highly reliant on this population data, and some hospitals and schools might receive inadequate resourcing as a result. Much of this is explained in Cate Broughton’s article today, Health boards, schools may lose funding as Ministries forced to use 2013 census data.

In this, University of Otago’s health systems expert Robin Gauld explains how hospitals might be negatively impacted: “It could have fairly profound implications for a DHB off a $1 billion dollar budget – if you’re a per cent or two off, that’s $10 or $20 million dollars, the difference between surplus and deficit.”

The article points out that this is likely to have a particularly adverse impact on poorer communities, with Child Poverty Action Group’s Alan Johnson quoted: “What you will get in places like South Auckland is there might be 10,000 – 15,000 people missing from the count – well the DHB won’t be getting funded for them so them and everyone else in that area will struggle with less funding.”

Currently the problems with the census data means that government departments are having to rely on 2013 census data. In terms of schools, the president of NZEI, Lynda Stuart, says: “If they are looking at using 2013 census data then yes potentially schools’ [funding levels] could stay the same and yet communities do change so it’s obviously highly problematic”.

So will the census debacle have a negative impact on social wellbeing? Certainly, the Government is making much of putting “wellbeing” at the centre of Grant Robertson’s upcoming Budget, but there must be questions as to the credibility of an approach that emphasises the need to measure social indicators when measures are so inadequate.

Furthermore, the Government has admitted that Robertson’s Budget won’t make use of the latest data: “Crucial funding decisions in Budget 2019 will be made without data from Census 2018” – see Thomas Manch’s Census data won’t make Budget 2019. National’s finance spokesperson, Amy Adams, says that this is “highly unusual”, but the Government has claimed that they never intended to use the new census data in the Budget.

Adams has also asked whether hospitals and schools which receive inadequate funding based on old data will eventually have their correct funding restored and backdated once the new data arrives.

Electoral consequences of the census debacle

The other major census headache is the upcoming general election, which is constitutionally-mandated to be run on the basis of electorate boundaries being redrawn in light of the census. The exercise of the boundary re-drawing is carried out by the Representation Commission (which effectively involves the Electoral Commission and representatives of the Labour and National parties). But if they don’t have access to the latest census data, or if the census data is deemed unreliable, then the whole exercise could collapse or lead to messy court proceedings.

This is best covered by Henry Cooke in his article, Statistics New Zealand confident census data will be ready for 2020 election boundaries. In this, National’s Electoral Reform spokesman Nick Smith challenges the robustness of the likely census data for the election, and says: “It is totally unsatisfactory to be determining electorate boundaries that can effectively determine who will be the next Government on the basis of guesswork”.

If National doesn’t have confidence in the census data being used to re-draw the electorates then the party could take legal action or simply pull out of involvement in the Representation Commission, which would collapse the whole process. This article says that National want the 2013 census data to be used instead, and “National was not ruling out ‘alternative action’ if this did not happen”.

A big issue is the matter of when Statistics NZ manages to get the data to the Electoral Commission, which is currently unknown. National’s David Farrar explains the problems with the timetable: “This would normally have been September 2018, so we’d have new boundaries by March 2019 – well before the election. If Stats NZ can’t release census data before say December 2019, then we’d have final boundaries in June 2020. That would be a disaster. Selections would have occurred by then. You can’t change boundaries just three months before an election. Parliament would probably have to legislate to delay the new boundaries until after the 2020 election. Ideally new boundaries should be finalised before election year. That means Stats NZ really needs to get the data out by June 2019. Any later than that and it will create a real headache for the boundaries review” – see: The census disaster gets worse.

Blame game and unaccountability

Given the anniversary of the 2018 census debacle, there is now a renewed interest in working out exactly what went so wrong a year ago and who is to blame. Journalists have had great trouble getting answers on any of this, because both the Government and Statistics New Zealand have been uncooperative and uncommunicative about what has happened.

But with the help of the Official Information Act, Newsroom’s David Williams has accessed 189 pages of information from Statistics New Zealand, which help illustrate the process during which the disaster unfolded – see his must-read account of “a digital-first experiment gone wrong”: Bungled, costly census to produce less ().

Williams has also been seeking comment from officials and politicians. He says “Stats NZ said no one was available for an interview.” And the Minister of Statistics James Shaw is asked “about his confidence in the integrity of the census data”, to which he rather weasily replies: “I am confident Stats NZ is making every effort and applying as many options as possible to deliver robust Census data.”

And a Statistics New Zealand official is interviewed, who says: “There is no accountability or responsibility being taken internally for what is turning out to be the worst census in over 50 years.”

For a very in-depth and interesting account of the census debacle, see the Otago Daily Times feature story, And then there were nine, by Bruce Munro, which was published in the weekend. This is, so far, the ultimate exploration of what went wrong.

Kathy Connolly, Stats NZ’s census general manager, answers questions put to her about whether the debacle was a result of the previous National Government running down the public service and replies “no comment”. Was the debacle due to Stats NZ being asked to “Go do a cheap census”? Again: “No further comment”.

There is also great debate in the article between Labour and National politicians about underfunding and oversight. They both blame each other, of course.

Munro also reports: “What exactly happened is hotly debated, but not loudly. Most of those close to the action will not talk on the record. At an operational level, when it came to rolling out New Zealand’s first largely online census, several wheels fell off, they say. There wasn’t enough publicity. Statistics New Zealand relied on the diminished postal system to get initial information to people. There were not enough forms. Fewer people were employed to follow up on those who had not filled out their form.”

In the end, the results of the census debacle are likely to impact negatively on the marginalised of society. In this regard, the views of the University of Otago’s professor of public health, Peter Crampton, are reported: “those for whom good social policy is of the highest priority” will be undercounted. The result will be a worsening of their position: “If these groups become invisible in the census then policy-making becomes doubly difficult and some of the least advantaged communities are at risk of becoming further marginalised.”

The decision by Statistics New Zealand to slash the number of census field workers is “beyond comprehension” according to economist Brian Easton – see his opinion piece today: Census mess can be resolved with a new one in 2021. But Easton argues that it’s quite feasible for the Government to get a robust new census up and running for the early date of 2021.

Finally, have Statistics NZ staff been spending too long “checking their privilege” instead of checking the pulse of the nation? Just before Christmas it was discovered that the government agency had their staff carry out a game to determine who was ‘white’, ‘Christian’, ‘able-bodied’, ‘male’, ‘heterosexual’ and had ‘no speech impediment’ – see Dan Satherley’s Stats NZ defends hosting ‘Check Your Privilege Bingo’ game.