Statistics NZ have released the very first results from the Census. I think they show that estimating the population size is a complex task.
The 2013 Census night usually resident population was 4,242,048. This is just 214,101 more than the 2006 Census, and 228,752 fewer than Statistics NZ estimated in June 2013.
Statistics NZ base their estimates on available census data, and births, deaths, and migration since the previous census This post is not a criticism of Statistics NZ – it highlights the importance of the Census for providing the Government, government agencies, and others with accurate population data. UPDATE: Statistics NZ say:
“The population number is lower than New Zealand’s estimated resident population, as it doesn’t include New Zealand residents overseas at the time of the census or the estimated number of people who didn’t complete the census. This shows the value of checking our population periodically with the census.”
I expect the reasons for the discrepancy in actual vs estimated counts will become clearer as more Census 2013 data are released.
This discrepancy also highlights, for me, that survey researchers need to think carefully before using projections for survey sample design and post-stratification weighting.
UPDATE 2: I just asked Thomas Lumley over at StatsChat for his view on the difference. The estimate appears to include residents who are temporarily overseas, whereas the Census does not include these people. Thomas estimates that this may account for around half of the shortfall.
My understanding is that people ordinarily resident in NZ who are temporarily overseas are still counted in the estimate, but are not counted in the census. According to Stats NZ there were 205000 trips overseas by NZ residents in August 2013. According to MED, the average duration of a trip is 20 days, so that would account for about half the shortfall.
NZ citizens who are resident overseas aren’t counted in either number (there’s a lot more than 228000 of them).
125,000 or so not completing the census doesn’t sound entirely unreasonable. When I get time I’ll look to see if Stats NZ publish an estimate for this.