Fri 24 Apr 2015 2:14AM

Usual residence one year ago

SD Sophie Davies Public Seen by 300

Usual residence one year ago was last included in the 1981 Census, and is proposed for inclusion in the 2018 Census.

New Zealand has high external (international) and internal migration rates by international standards. Therefore, accurate estimates of internal migration are essential for producing accurate population estimates.

The census is currently the only official data source on internal migration, as captured through ‘usual residence five years ago’ and ‘years at usual residence’. However, to calculate annual subnational population estimates we require an estimate of movements over the last year. This information is currently not available from the census or other official sources.

Our current recommendations relating to usual residence one year ago

  • We recommend that usual residence one year ago is included in the 2018 Census.

See our preliminary view of 2018 Census content (page 19) for a more detailed discussion on usual residence one year ago information.


Bronwen (Facilitator) Thu 30 Apr 2015 9:52PM

Hello & welcome to our discussion of “Usual residence one year ago”.
I’m Bronwen, from Statistics New Zealand. I look forward to open and inclusive discussion over the next six weeks to understand your “Usual residence one year ago” needs. Looking forward to hearing from you


Charles Sullivan Fri 22 May 2015 2:40AM

One extra use for this may be to improve analysis from the IDI (Integrated Data Infrastructure), assuming that 2018 Census data will be linked to IDI. Address at time of Census enables us to fill an IDI data gap in terms of knowing who is in same household, but is limited to that single point in time. Having address a year earlier could allow us to greatly broaden the period of analysis for such household data. Confession: quite how important/useful this will be I'm not sure---won't know until if and when we've been able to explore usefulness of 2013 address/household data from Census in IDI.


Bronwen (Facilitator) Sun 24 May 2015 7:27PM

Thanks for joining into the Location discussions. Your thoughts & ideas are appreciated @charlessullivan - Bronwen


Kim Dunstan Mon 1 Jun 2015 11:15PM

Great to see 'usual residence 1 year ago' (UR1YA) proposed for inclusion for the first time since 1981. As a demographer, I can say we currently have no other data source to rival census for an authoritative and comprehensive measure of internal migration. It’s the geographical detail we get from census, combined with detailed personal characteristics (eg age, ethnicity), which makes this information source so valuable.

The inclusion of UR1YA comes at a time when we are trying to progress our methods of population estimation. UR1YA is most suitable for modelling the single-year of age distribution of migrants as it better identifies the age at move. UR1YA will give the spatial patterns for the most recent year (ie 2017-18). By comparison, 'usual residence 5 years ago' (UR5YA) is valuable as it gives a robust spatial pattern over a longer period - there are more moves and the data is less affected by a single event in the most recent year. The question generally ties in with the date of the previous census, so all the intercensal components of population change can be estimated.

In time we might be able to get similar internal migration information from administrative data. If and when we do, having these address questions in census will be an important calibration/validation of that administrative data.


Derek Phyn Tue 9 Jun 2015 2:16AM

Hi Bronwen,

Usual residence one year ago would be a very valuable statistic for Civil Defence and Emergency Management (CDEM) planning. CDEM invests heavily in public education and encourages locals to be resilient and prepared for emergencies. Locals who have lived in an area for longer are more likely to have good local knowledge, be more socially networked, know the local hazards and be better prepared for an emergency. Knowing how many usual residents have lived in a location for less than a year, right down to meshblock scale, helps us identify locations with higher proportions of "new residents" that are likely to be less resilient. Knowing this information thereby helps us identify where we should prioritise our efforts in public education and community resilience development.

Given CDEM practices and hazard risks vary from district to district I would ask that the census question be posed in a way such that we can identify those that lived overseas one year ago and those that lived in another TA (district or city) one year ago.

Something for a formal submission perhaps?



Sophie Davies Fri 12 Jun 2015 1:28AM

Hi @derekphyn I'm the other facilitator for this topic. That's a great comment, really gives us an idea of how it would be of use for emergency management as we currently don't have many specific examples of how it would be useful for your area. Would be great if you there's anyone else in the emergency management sector that could also provide an insight as to how this topic would be of use.
Our current recommendation is to include usual residence 1 year ago, usual residence 5 years ago however we are wanting feedback on whether we still need to collect information on years at usual residence. Would this also be of use for you or would usual residence 1 year ago and 5 years ago fulfil your need?
We would encourage you to put in a formal submission as well through www.stats.govt.nz before the 30th of June however it would be great to get some more discussion on this topic first


Derek Phyn Fri 12 Jun 2015 3:59AM

No just usual residence 1 year ago and 5 years ago would fulfil our needs.