Fair to say, the last few censuses were bungled.
The 2018 New Zealand census was severely undermined by an overreliance on a "digital-first" approach, leading to the lowest response rate in 50 years. Just 83 percent responded, well short of the 94 percent target. The fallout included delays in data release, a contempt threat in Parliament, and the resignation of Government Statistician Liz MacPherson after an independent inquiry found serious flaws in planning and execution.
The 2023 census, meant to be a one-night national headcount, dragged on for nearly four months. It was an embarrassing, drawn-out affair for what is supposed to be a credible statistical exercise. Cyclone Gabrielle was offered as the excuse, but it did not wash. Some of the worst response rates were in major cities that were completely untouched by the storm.
In the end, Statistics NZ resorted to handing out supermarket vouchers and wellness packs to coax people into participating. It was a degrading spectacle that further eroded the seriousness of the entire process. Adding to the farce were serious allegations that Te Pāti Māori misused census data for electoral advantage. The allegations are now the subject of an active and seemingly endless police investigation. The fact that such claims are even plausible tells you how far trust has already eroded.
The old manual census, while more painstaking and labour-intensive, somehow managed to be less expensive and far more effective. Enumerators delivered and collected paper forms. Communities were actually spoken to. The data was solid. Now we spend hundreds of millions on digital platforms, call centres and PR campaigns, only to end up bribing people with grocery vouchers and still falling short. We replaced diligence with gimmicks, and it showed.
So having failed to do something New Zealand managed successfully 33 times in the past, the government’s answer is not to demand higher standards or scrap the bloated IT boondoggles. No, the solution, apparently, is to scrap the census entirely.
There will be no 2028 census. And from 2030, the government will rely on stitched-together administrative data from various agencies and a few smaller annual surveys. These are the same data sources that the Government Statistician admits still lack the attributes needed to do the job.
IRD tracks tax. MSD tracks benefits. The health system tracks patients. None of them track who lives where, with whom, or how they identify. So we shoot one white elephant, the digital-first census, only to be gifted another: a sprawling government data integration project based on the hopeful fiction that the New Zealand state has suddenly become good at IT procurement.
And so it comes to pass that a once-reliable national institution is being binned, not because the concept no longer works, but because those responsible cannot be trusted to do it properly.
The government says this is about costs. But if that were really the concern, why not just move to a ten-year census instead of every five years? That is what most countries do. It would be cheaper, more accurate, and far less prone to political interference.
Because here is the thing: the census is not just a statistical exercise. A lot rides on it. This includes electorate boundaries, funding formulas, health and education planning, Māori representation, infrastructure decisions and more. It is one of the few remaining exercises in national coherence and democratic accountability.
Keep reading with a 7-day free trial
Subscribe to The Blue Review w/ Liam Hehir to keep reading this post and get 7 days of free access to the full post archives.