Māori miss out in Census 2018
Statistics NZ has been given almost $16 million in this year's Budget to fix up problems in the 2018 Census - and make sure they don’t happen again.
A shift to completing the Census online and a low response rate has hampered the 2018 effort, leading to calls from the National Party that patching holes in it would turn into “guesswork”.
Stats NZ fought back on Monday, saying Census data would be released from September 23, and the data would be robust enough for calculating the number of electorates needed for the next election, as well as where the boundaries for those electorates should be drawn.
The department says the complete Census records will use data collected from other government departments to fill in the gaps. It says 89 percent of the final record has come from the 2018 forms, with the remaining 11 percent coming from other government data.
Iwi data will not be released
While this is enough information to build a picture of the number of people in New Zealand, the quality of more granular data has been affected enough for Statistics NZ to block its release entirely.
Data relating to iwi affiliation, for example, will not be available for the 2018 Census.
A lack of iwi affiliation data could have an impact on Treaty settlements.
A Te Arawhiti spokesperson told Newsroom it used iwi affiliation to build understanding of the groups it was negotiating with and to create regional profiles and help the public sector with iwi information.
The data was also a “secondary” factor the Crown considered when developing its Treaty settlement offers.
The spokesperson said Te Arawhiti would work with Statistics NZ and iwi to gather the best usable data from 2018.
Data on Māori ethnicity had been collected accurately, and would be able to be used - just not at the level of iwi.
A step in the right direction
Users were also concerned with 2018’s poor showing, but agree the move to using more government data could be better in the long term.
Infometrics senior economist Brad Olsen told Newsroom that adding other government data to the Census would give economists “a much better read of what’s actually happening”.
But he said that the data would be less useful when compared to information gathered in the 2013 Census, which used the older, traditional methodology.
“We’re getting away from the apples and apples comparison,” he said.
Olsen said Statistics NZ could consider going back to the 2013 Census and updating it using the 2018 methodology to improve the comparison.
ANZ economist Miles Workman said the main data economists used appeared to have been untainted by the troubles at Statistics NZ.
“For us the big part is the population estimates, population growth, and migration because it’s an important part of the economic cycle,” he said.
Workman agreed that while the current system was experiencing teething issues, ultimately the data would be more robust.
An example of this was moving to an electronic measurement of migration, following the removal of arrival and departure cards. This would provide more accurate data in about 18 months, but the 18-month transition to the new, robust data could be fraught.
“It’s better data once it’s 18 months old, but if you’re an economist trying to get a pulse on economic activity, it’s difficult,” he said.
Kiwibank senior economist Jeremy Couchman said it would be important to know whether Stats NZ had an idea of which groups were most severely affected by the lack of data, and whether they belonged to a particular part of the country, or a particular ethnicity. This would build a more accurate picture of who was missing.
“Hopefully they can shed some light on that,” he said.
How it’s done
Adding government records to the 2018 Census without doubling up on records involves first creating a “backbone” of key, high quality data from other government records like Births, Deaths, and Marriages, as well as immigration data.
This is then fed into the main Census dataset.
Statistics NZ is still confident of the main population data, noting that its records for approximately 4.7 million people was within 1.2 percent of their best estimate on Census night 2018 - better than the 2.4 percent in 2013.
However Minister of Statistics James Shaw said he was reserving his judgment until the review of 2018 comes back in July.
He said he currently had confidence in chief statistician Liz MacPherson, however he would reserve full judgment until the final report.
“That will give us a politically neutral view of how the Census has been conducted over the last several years,” he said.