Housing demand report wrong: SNZ
Statistics New Zealand is challenging a Colliers International report about population data that raised questions about whether housing demand was less than thought.
Colliers’ report looked at the growth in the population numbers in Census 2018 from Census 2013, an increase of 457,700 people, about 10.8 per cent.
Colliers then looked at the Stats NZ June 2018 provisional population estimate and compared that with the Census 2013 population count, and said the predicted increase was 643,500, substantially higher than the 457,700 growth from the census counts.
It then questioned if housing demand was overstated. It suggested 69,000 fewer houses might be needed.
Statistics New Zealand said the census population counts are not directly comparable with the population estimates (estimated resident population) because the population estimates include people not counted by the census.
Senior demographer, Kim Dunstan, said Colliers erroneously compared the 2018 national population estimate with the 2013 census count to get the so-called ‘‘643,500 forecast’’.
‘‘This is an entirely inappropriate comparison.
‘‘If they compared the revised June 2018 and June 2013 population estimates they would have got a five-yearly increase of 398,900. This is actually less than the change implied by the census counts.’’
The growth in the population using 2013 and 2018 census counts was 457,700.
The population estimates were always higher than the census figures in the same year, because the estimates made allowance for people not counted in the census like people overseas temporarily. That was about 80,000 people on census night 2013, Dunstan said.
New Zealanders overseas temporarily on census night were not required to fill it in, because the census was a snapshot of residents and visitors on census night.
And, there was also usually some under-counting in the census for other reasons.
The population estimates were a better measure of the population than the census, which provided a wealth of other information about New Zealand.
As well, the population estimates were not forecasts, but estimates based on recorded data, such as births, deaths and external migration.
‘‘The key thing is the estimates should be compared with estimates, the census counts with census counts. Don’t conflate things by trying to compare two different measures because the estimates are always higher than the census count in the respective year,’’ Dunstan said.
Statistics NZ had discussed its views with Colliers on how Colliers had compared the different data. The population estimates made adjustments based on the Post Enumeration Survey (PES), which was run shortly after each census as a check on how well the census count did its job.
Population estimates based on the 2018 Census would be published in March next year, he said. ‘‘At that point we’ll have the best measure of population growth since 2013 because those estimates will incorporate our estimates of net census under-count from the
This is an entirely inappropriate comparison. If they compared the revised June 2018 and June 2013 population estimates they would have got a five-yearly increase of 398,900. Senior demographer, Kim Dunstan
2018 Census as well as our estimates of how many New Zealanders were temporarily overseas at the time of the 2018 Census.’’
Colliers director of research and communications, Chris Dibble, said: ‘‘We appreciate the Census 2018 information released so far doesn’t provide the complete picture of our population yet.
‘‘But in the absence of new estimates based on Census 2018, and further revisions to come, we have used the information available for the analysis to create further discussion and debate.
‘‘We also stated in the report not to ‘take the data at face value’ for these reasons and others.
‘‘Given future estimated resident population numbers will require net migration adjustments, which is a data set that is also receiving a lot of attention, there will no doubt be ongoing discussion and debate in this area,’’ Dibble said.
‘‘We look forward to receiving additional revisions to Census 2018 data and estimated resident population information later this year and next year.’’