This privacy impact assessment (PIA) assesses the implications of acquiring Immigration New Zealand (MBIE) monthly visa data to potentially improve the accuracy of Stats NZ's international migration estimates.
Download the document below, or read the recommendations and summary online.
Recommendations from Stats NZ's privacy team
There are no recommendations for the statistical model acquiring MBIE visa applications or approvals unit record data. Stats NZ will need to continue to upgrade and update the security of the tools used and ensure only staff who require access to the data have access, as is the current process.
Summary
Stats NZ's migration estimates do not currently have access to Immigration New Zealand's (stated in this PIA as Ministry of Business, Innovation & Employment - MBIE) visa applications or approvals unit record data. Acquiring access to this data would potentially improve the accuracy of Stats NZ's international migration estimates.
Official measures of New Zealand's international migration are derived and published monthly by Stats NZ. Since 2019, these measures do not rely on traveller intentions as stated on passenger cards, but are calculated precisely from the actual duration of time spent in, and out of, New Zealand after crossing the border. For more information see Migration Data Transformation.
A 12/16-month rule is used to classify border crossings as short-term (non-migrant) or long-term (migrant) on the basis of whether travellers spend 12 months (or more) of the following 16 months in New Zealand.
A migrant arrival is an overseas resident who arrives in New Zealand and cumulatively spends 12 out of the next 16 months in New Zealand.
A migrant departure is a New Zealand resident who departs New Zealand and cumulatively spends 12 out of the next 16 months out of New Zealand.
These are statistical definitions of 'resident'. A New Zealand resident is someone living in New Zealand, irrespective of their citizenship or visa status. Overseas residents and New Zealand residents both comprise New Zealand citizens and non-New Zealand citizens.
After 16 months, the migrant status of all travellers can be definitively classified. However, to maintain the timeliness of New Zealand migration estimates, a statistical model is used by Stats NZ to classify border-crossings whose migrant status is uncertain. This enables provisional estimates of migration to be published about six (6) weeks after the reference month.
The model looks at 'features' in the historical border-crossing data:
- features from the border crossing, such as the direction of crossing, and date of the crossing
- features of the 12/16-month rule, such as the amount of time in/out of country, the amount of time that has passed since the border crossing, the count of border crossings after the crossing to be classified, whether a traveller is still in the country, and their resident status before the border crossing
- features from the passport data, such as the age, sex, and citizenship of traveller, and the broad type of visa they hold.
The statistical model does not currently have access to MBIE visa applications or approvals unit record data which is supplied into Stats NZ's Integrated Data Infrastructure (IDI). Acquiring access to this data would potentially improve the accuracy of Stats NZ's international migration estimates, especially the provisional estimates which require the most modelling, have the highest uncertainty, and are liable to the largest revisions.
For example, a significant number of migrants arrive on a visitor visa. Given the limitations of a visitor visa, most of these migrants are likely to have been issued another visa (for example, work, student) after arriving, to enable a stay in New Zealand of at least 12 months. Knowing that a traveller is no longer on a visitor visa, but on a visa with a longer entitlement of stay in the country, would therefore assist the migration model and benefit customers with more accurate provisional estimates.
The migration estimates are produced monthly, roughly four weeks after the end of each month. To support Stats NZ's statistical model, a supply of the MBIE data − inclusive of a historical back series − would be required on a monthly basis, roughly three weeks after the end of each month. Data for the IDI is currently supplied quarterly.
ISBN 978-1-991307-53-8
