May 02, 2019
Sam Green

Census asking questions we already knew?

The 700,000 people missed by the 2018 census have been filled in by admin data held by StatsNZ. If this was already available, do we really need a census?

The 2018 census has been a highly publicised debacle. With 700,000 people missed by the census, StatsNZ have been filling in the gaps with data collected from other sources. StatsNZ assured the country this week that the 2018 census is actually “better than 2013” in estimating the size and key demographics of the New Zealand population.

Living in New Zealand, most of our life is recorded. When we are born or arrive in New Zealand, file a tax return, visit a doctor, get a student loan, graduate, receive a benefit or register a vehicle. All of the activities are stored in some way by StatsNZ as part of their administration data. Following the low response rate from the 2018 census, this information has been put to use in filling the gaps. So if the census was able to be updated with “real data about real people” why did we bother with the census in the first place?

StatsNZ has also asked this question, and they have indicated that we may not see a traditional census like 2013 again. The methodology used to patch up the 2018 census was already in development for use in 2023. The end goal of this project is a utopia where administration data is available for the full census, and it could be held annually without asking for any additional information from the public.

The reason we aren’t doing that already is the administration data for many of the key census areas does not exist, or is currently too limited to rely on for a snapshot of New Zealand. The StatsNZ leadership team provided no information yet as to how many variables they have been able to source from admin data, though I expect it to be limited to broad demographic information (age, sex and ethnicity) and income.

The census is for much more than just counting how many people live in New Zealand, but even that can be challenging. Migrating to New Zealand involves a long paper trail, but even providing a reliable estimate of migration numbers has been a challenge.

The census asks questions that do not usually appear on Government forms. This includes some personal information (religion, iwi affiliation, and occupation), and household information. The 2018 census had additional first time questions about households including presence of dampness and mould, things that will have limited usefulness with high undercount in vulnerable communities. One of the other critical pieces of information is the household structure information from the census dwelling form, which is particularly important with the ongoing issues with housing across the country including overcrowding.

We’re still holding out for a full summary of the data that will be available. StatsNZ has promised a first public release of the results on 23 September, more than 18 months after census night.