Chart of the month: Migration estimates like playing darts blindfolded
Tue 19 Mar 2019 by Brad Olsen in NewsletterMigration

It’s now been three months since Statistics NZ released new migration data based on traveller outcomes rather than intentions. And all we actually know for sure is that no one has any idea what is happening to migration any more. We’d do just as well to estimate next month’s migration figure by playing darts while blindfolded.

Migrant arrivals are now estimated and modelled according to the new 12/16-month outcomes-based rule, which means that the latest migration estimates can be revised each month for up to 16 months.

We’re at an absolute loss to say what’s happening to migration at present, as each revision in the last three months has brought an entirely new story of migration trends. Data for November 2018 showed migration in an accelerating decline, December 2018 figures showed migration plateauing, and January 2019 data (released last week) showed migration rising (see Graph 1).

Graph 1

The change in trend each month has seen net migration swing from an estimated inflow of 43,400 in the year to November 2018 up to 58,400 in the year to January 2019.

The revisions are necessary because of how many traveller movements into and out of New Zealand are ”unknown” each month, meaning they need to be modelled. For example, in December 2018, over 60% of all travel movements are currently uncertain, which equates to around 860,000 passenger movements (see Graph 2).

Graph 2

The new migration figures also make population estimates appear overenthusiastic. Our back-of-the-envelope calculations make it look like the population might be inflated by around 45,000 people.

Essentially, it’s increasingly difficult to make heads or tails of the new migration numbers, and even harder to trust them. With massive revisions in both the level of net migration and, more concerningly, the trend, we’re unsure just how useful they really are to track.

Of course, given the lack of any data from the 2018 census, all our population numbers at the moment are pretty much just guesses. Maybe the dartboard isn’t such a bad idea.

Related Articles