Experimental population estimates from linked administrative data
We have produced a series of experimental series of population estimates for New Zealand, and for a range of geographic areas, using the linked administrative data in the Integrated Data Infrastructure (IDI). They are part of Stats NZ’s Census Transformation programme exploring the possibilities of using administrative data to produce population statistics and other census-type information.
This 2017 release updates the methodology and extends the data series from a previous release in 2016. Two further experimental administrative data series of ethnic population estimates, and income estimates, will be based on the estimated populations derived here.
These experimental population estimates are NOT official statistics. Rather, they are published as output from research using a different methodology than is currently used to produce the official estimated resident population (ERP).
We derived these experimental population estimates using activity in administrative sources over a two year period to indicate an individual’s presence in New Zealand. We then remove individuals who have died before the reference date or were identified as being overseas residents.
No adjustment has been made to account for people who may have been missed, or people who may have been incorrectly included, using these activity-based rules.
Our comparison of the experimental population estimates produced directly from administrative data against the official figures is encouraging. Often there is close agreement with official figures, and we have seen results improving steadily over time. However there are still marked differences for some age groups and for some local areas, and we are looking for ways to improve the data and our methods.
We do not expect to fully meet accuracy requirements from administrative data alone, and are also developing statistical models to adjust for remaining errors.
Latest release: 2 October 2017
Our latest release includes estimates at 30 June 2007–16 for:
- New Zealand by single year of age and sex
- Territorial authority and Auckland local board areas by five-year age group and sex
- Area unit total populations
You can download the following:
Experimental population estimates from linked administrative data: 2017 release (PDF, 38 pages, 1.42MB) – a technical paper summarising the data, methods and results
Experimental population estimates: 2007–16 (zip file) – estimates by various geographic areas, age group and sex in both excel and CSV format.
What you'll see
You will see a series of population estimates at 30 June 2007 ̶ 16. The time series is available in both CSV and Excel format. An interactive tool will be added shortly, which will allow for comparisons of the experimental series against the official population estimates.
We'd like your feedback about
- the overall usefulness of the series and accompanying report
- the usefulness of the interactive tool included with this release
- any suggestions for possible improvements to the method described
- any other explanations for the observed trends
- how the quality of the estimates, and the observed differences would affect their usability
- anything else you would like to see in future releases.
To provide feedback, click on the 'Give feedback' button above.
This series has been produced using administrative data in the IDI from a range of agencies, including:
- Inland Revenue
- Ministry of Education
- Ministry of Health
- Department of Internal Affairs
- Ministry of Business, Innovation and Employment
- Accident Compensation Corporation.
This is an experimental series with known limitations. We are aware of potential sources of error leading to undercoverage (people missing from the population) or overcoverage (people incorrectly included in the population). See Experimental population estimates from linked administrative data: 2017 release for more details.
Previous release – 30 September 2016
A first series of experimental population estimates was released in September 2016. Improvements have since been made to the methods being used.
You can download the following:
- Experimental population estimates from linked administrative data: methods and results (PDF, 26 pages, 878kB) – a technical paper summarising the data, methods and results
- Experimental population estimates: 2007–14 (Excel, 4 sheets, 43kB) – estimates by various age groups and sex
- Experimental population estimates: 2007–14 (CSV file, 30kB) – estimates by single year of age and sex in machine-readable format.
The results in this paper are not official statistics. They were created for research purposes using the Integrated Data Infrastructure (IDI) managed by Stats NZ.
Access to the anonymised data used in this study was provided by Stats NZ in accordance with security and confidentiality provisions of the Statistics Act 1975. Only people authorised by the Statistics Act 1975 are allowed to see data about a particular person, household, business, or organisation. The results in this paper were confidentialised to protect these groups from identification.
Careful consideration has been given to the privacy, security, and confidentiality issues associated with using administrative and survey data in the IDI. See Privacy impact assessment for the Integrated Data Infrastructure for more information.
Updated 2 October 2017