Check the ABS website, and you’ll find detailed population counts for every local government area LGA in Australia, reported “as of” June 30 for the end of the financial year.
The fastest-growing LGA in the financial year 2024-25 was the Darwin Waterfront Precinct (growing an impressive 9.3% by adding just 30 people), while the biggest decline was in remote Bourke, NSW (down 2.4% from losing 56 people).
But wait—how did the ABS know that? It hasn’t conducted a census since 2021, and postal and housing data don’t tell you how many people live in a particular household. The ABS can get a pretty good picture of the total population of Australia by tracking vital statistics and overseas migration, but how can it tell where people are once they’re in the country?
The short answer is that they can’t. Australia conducts full population censuses every five years. This gives the country more up-to-date population figures than countries like the US and UK, which use ten-year cycles.
But a five-year census is still a five-year census, and the ABS cannot magically determine population statistics for intercensal years. Still, the data monster must be fed, and planning requires best estimates to work from, even if they may not be perfect.
Intercensal ‘data’ at the LGA level is based on a combination of modelling and guesswork. The ABS starts from the most recent census as a baseline, then estimates inter-regional migration using registered Medicare addresses, with ATO tax data as a backup.
So every time your primary care physician asks you “has your address changed since your last visit,” that’s feeding through to ABS regional population estimates.
It’s a good start, but many people fail to update their addresses with Medicare, and nonetheless stay frustratingly well (keeping them out of the doctors’ offices for long periods of time).
If the distribution of people who didn’t keep their records up to date was random, this wouldn’t be a problem. But it’s not. Young people, single people, and manual labourers are less likely to update their addresses—and more likely to be mobile. As a result, Medicare records produce a biased picture of intra-regional mobility.
Even worse, the number one driver of population change in Australia is … temporary migration. International students and working holidaymakers account for around 80% of Australia’s population change at the national level. For many individual LGAs, the proportion is likely be much higher.
And international students and working holidaymakers are not on Medicare. As a result, not only is the net overseas migration component of LGA population change essentially unknown, but the inter-regional migration of temporary migrants within the country is also unknown.
For example, working holidaymakers who want a second year on their visas have to spend 3 months working in regional Australia. Those who want a third year have to spend 6 months in the regions. None of this movement to and from the regions is captured in Medicare data, and it is difficult to track in tax data.
So what’s a data-starved demographer to do? The ABS has chosen to estimate regional overseas migration “by breaking down overseas migrant arrivals and departures at the state level to sub-state areas, using information from the most recent Census.” But the most recent census was conducted in August 2021, at a time when net overseas migration was essentially zero. So instead they turned to patterns from the 2016 census to fill in the gaps.
That’s right. Your 2024-25 regional population estimates are based (in part) on a 2016 census question asking people where they lived one year ago. How much in part? We don’t know.
This is not to criticise the ABS. Their statisticians must do the best they can with the data they have. After all, decisions have to be made based on some estimate of regional populations, and an error-prone estimate is better than no estimate at all.
But ‘expert’ commentators should not be afforded the same clemency. When demographers, economists, and statisticians provide colour commentary for press reports on ABS releases, they should be expected to have read (and understood) the fine print.
Regional population changes are not the only ‘data’ produced annually by the ABS that are not actually based on timely, meaningful, real-world observations. Recent releases on services exports, country of birth, underemployment, and dwelling commencements all require expert interpretation.
Unfortunately, many purported experts are too lazy (or too ignorant) to provide the necessary guidance. Too often what passes for expertise is really just the ability to repeat information from the ABS website.
The next time you hear an expert interpreting ABS data in the media, listen for the caveats. If you don’t hear them, turn off the program—and check the fine print for yourself.
