International migration estimates extended back to 2001

Estimates of migrant arrivals, migrant departures, and net migration using the new measure of migration are now available back to January 2001, Stats NZ said today.

“Extending the time series further back, from 2014 to 2001, is good news for customers who want a long, consistent series for analysis and forecasting,” population indicators manager Tehseen Islam said.

Stats NZ introduced the new outcomes-based measure in 2017 with an experimental series, and in January 2019, the new measure became the official measure of migration. Adopting the outcomes-based migration measure allows us to measure actual passenger behaviour rather than their stated intentions when they are crossing the border. The new estimates supersede the published experimental series and intentions-based series from 2001.

The extended outcomes series highlights the ups and downs in New Zealand’s net migration balance. It also shows that, at times, actual net migration has been different from that suggested by measuring intentions. However, it also generally confirms the patterns and trends measured by the intentions-based series.

The early 2000s saw significant net migration gains, driven by relatively high migrant arrivals and low migrant departures (figure 1). The new extended migration series shows that these net migration gains were even higher than previously indicated by the intentions-based series. For example, annual net migration peaked at 60,700 in the year ended February 2003, rather than peaking at 42,500 in the year ended May 2003 as reported at the time.

Figure 1

There was lower net migration over the period 2004 to 2013, with lower migrant arrivals and higher migrant departures than indicated by the intentions-based series. Between December 2010 and May 2013, departures exceeded arrivals. The new extended migration series shows that these net migration losses were slightly greater than previously indicated by the intentions-based series. For example, annual net migration dipped to a loss of 16,800 in the year ended February 2012, rather than dipping to a loss of 4,100 in the year ended August 2012, as reported at the time.

Since 2013, migrant arrivals have been consistently high, resulting in a sustained period of high net migration. In the five years ended June 2019, net migration added about 280,000 people to New Zealand’s population. The new extended migration series shows that annual net migration peaked at 63,900 in the year ended July 2016, compared with a peak of 72,400 in the year ended July 2017 as indicated by the intentions-based series. The July 2016 peak of 63,900 was only slightly above the previous peak of 60,700 in February 2003.

Contrasting migration flows by citizenship

Migrant arrivals and departures include the flows of New Zealand citizens, as well as the flows of non-New Zealand citizens, as both affect the population living in New Zealand.

Migrant arrivals are people who have been living overseas, including New Zealand citizens, who cumulatively spend 12 of the next 16 months in New Zealand after arriving.

Migrant departures are people who have been living in New Zealand, including non-New Zealand citizens, who cumulatively spend 12 of the next 16 months out of New Zealand after departing.

Recent net migration gains have been driven by non-New Zealand citizens (figure 2). The much smaller net losses of New Zealand citizens have also contributed to the recent high levels of net migration (figure 3). In the five years ended June 2019, net migration added about 310,000 non-New Zealand citizens to, and subtracted about 30,000 New Zealand citizens from, New Zealand’s population.

Figure 2

Figure 3

Provisional data

As part of the extended back series, we have revised monthly and year-ended final migration estimates up to April 2018; however, we have not revised the provisional estimates. As a result, year-ended provisional migration series are currently inconsistent with the sum of the monthly data. We will resolve this inconsistency when we publish the next international migration release on 12 November 2019.

Impact on population estimates

National population estimates: At 30 June 2019 and Subnational population estimates: At 30 June 2019 (provisional) incorporate the new migration estimates available at the time of release. Revised population estimates incorporating 2018 Census results, estimates of census coverage (net census undercount) from the 2018 post-enumeration survey, and final birth, death, and migration estimates will be available on 30 March 2020. Population estimates up to June 2013 are unaffected because these were revised after each Census of Population and Dwellings.

How we create travel histories

We create travel histories by identifying unique individuals in the border-crossing data within a secure and confidential Stats NZ environment. We use passport numbers, dates of birth, and names of travellers to identify unique travellers. These variables allow us to uniquely identify travellers who have multiple passports or who renew their passports.

Before June 2013, the border-crossing dataset did not contain all the necessary variables, so we were unable to use it to produce travel histories. Recently we have worked with administrative data available through the Integrated Data Infrastructure (IDI) to create an encrypted identifier that allows us to identify and construct travel histories for unique travellers back to 2001.

We have been able to extend the national-level outcomes-based migration series back to 2001 using the 12/16-month rule. We have extended the subnational migration series, which require extra data from the IDI, back to 2013.

Where to find the extended time series

The new migration estimates are available via published Infoshare series (Tourism – International travel and migration). We have made some small revisions to final migration estimates for June 2014 to April 2018, reflecting the improved and longer travel histories now available.

We have also extended the seasonally adjusted series back to 2001, with small flow-on changes to more recent seasonally adjusted estimates.

We have extended subnational migration estimates back to 2013.