Who goes where: Internal migration in the Australian census

How often do Australians move house? At what age are people most or least likely to move? Where are people moving from and to? Lots of trivia from the 2016 census.

This article is viewable both on this site and on Mappage, my webapp for looking at statistics. If you’re reading this on a large screen, it’s best to read this article in Mappage, where I have set it up with maps and charts. If you read this version, you will see references to the charts you can’t see here.

This article is viewable both on this page and in plain form on the Mappage WordPress site. The one-letter links here will show maps and charts. There are lots of these links, so you need only click the ones that interest you. Most of what I thought important is in the text here. This is not so much an article with a Google Map embedded; it is my statistics & mapping webapp (called Mappage) which I have set up to accommodate articles like this. Clicking the links just changes the settings on the right. You’re not restricted to looking at what the links provide; you can look around for yourself at any other stats.

The census asks where people lived one L wjieul0b and five L 1vdtuefs years earlier. This lets us look at who is moving, where and when.

1 year ago5 years ago
Same place17,658,98475.5% 11,503,71949.2%
Elsewhere in Australia3,136,66713.4%7,442,45431.8%
Overseas398,6501.7% 1,378,6715.9%
Not stated1,930,3508.2%1,611,200 6.9%
Not applicable276,115 1.2% 1,464,722 6.3%

“Not applicable” refers to those who weren’t born 1 or 5 years earlier. “Not stated” includes those who didn’t do the census at all, as well as those not answering the question1Curiously this is much higher for 1y than 5y; as though 1% of respondents ticked same as 5y ago, assuming that implied same as 1y ago as well..

Excluding the last three categories leaves us with a movement rate of 15.1% in one year (moving once per 7 years, with multiple moves in a year counting as just one2or none, if they return to their original place) and 39.3% in five years. This obviously means many people are moving more than once in the five years (in fact those in the 39.3% average nearly two moves). There were 122,000 returners3Returners: Of those living the same place 5 years ago, there were 122,000 living elsewhere in Australia 1 year ago. Many of these returners are in rural areas. There were also 20,000 living overseas a year ago but in the same place as 5 years ago. These mostly lived in rich SA4s, with the highest concentration in Canberra..

Of the 3,137,000 people who moved within one year4excluding those whose 1y ago address was not stated:
301,000 (9.8%) of those moved interstate.
902,000 (29.2%) moved to a different SA4 in the same STE.
477,000 (15.4%) moved to a different SA3 in the same SA4.
695,000 (22.5%) moved to a different SA2 in the same SA3.
711,000 (23.1%) moved within the same SA2.
So although 15.1% moved in a year, only 8.1% moved to a different SA3.

For indigenous people the movement rate is 20.2%, with 14.9% for non-indegenous.

T m8nd0gvb Looking at migration in and out of states, WA had interstate migration of 33,000 out and 22,800 in. NSW, SA and NT also recorded net migration out while there was net interstate migration into Victoria (14,400), Queensland (10,600) and ACT. T 21hzc7rv We can also look at how many uni graduates are among them.

Usually when any “1 year ago” variable is used, there will be a “1y→5y” link that changes all such variables between the 1 and 5 year variables (data isn’t always available).

C v4pdwt2q Looking at the 1y movement rate by age, we see 21.3% of 1-year olds had moved, reducing to 11.5% at age 16, increasing to a peak of 33.2% at 25, descreasing quickly then slowly to a low of 5.1% at age 77. C b674tiy1 For females, the peak is 34.8% at age 24, but the average is still 15.1%5Young women move more than young men, while at other ages the rates are about the same. Women live longer and so have more low-movement years..

C 48boxjqy Looking at the 1y movement rate by state, Queensland stands out with 17.9%. This holds true across age groups (for age 60-69, Qld is 9% v 7% nationally). WA is also above average. Breaking it down to SA4 level, the highest are places people move for employment – WA Outback North, Darwin and the inner SA4s of Brisbane, Melbourne and Sydney. All SA4s of Queensland are above average.

M kbaudl4s Looking at South Australia at SA2 level, Roxby Downs has the most movers at 32.7%, the city at 29.2% and the next ones are North Adelaide, Norwood and Mawson Lakes. Clicking ur⇆pur1p changes the percentage from moving out to moving in,6One of the constant complications with migration stats is whether you count as a proportion of where they moved from or where they moved to. This article isn’t consistent in that regard. so growth areas like Munno Para West show up. Lowest were One Tree Hill, Gumeracha, Athelstone and Coromandel Valley (all 6-7%), followed by lots of rural areas. The 5-year movement map is much the same. You can click “Res 1y ago” (or whichever says “[SA2]”) and another state/territory to see other parts of Australia. SA2s with jobs (and universities) stand out.

L nvlhy9ow Here is a table that breaks down the movers by geographical category of move in each state/territory. The territories see a lot of interstate migration and the larger states relatively little – I’m suprised that Victoria has only 1%. Queensland’s higher frequency of moving is spread across all categories.

movers% of
diff S/T4/St3/42/3same

Comparing with the 2011 census, the 1-year movement rate was lower at 14.6%. The 5-year rate was 37.8%. Queensland was 17.2% and 43.3%. But when we go back to further to 2006, the 1-year rate is higher at 15.5% and the 5-year rate is 40.3%. Queensland was 19% and 46.5%7Using Residence 1/5 years ago = Queensland, so includes migration out rather than the greater migration in.. The 2006 census also adds the detail that of those who moved in 2005-06, 1,358,000 moved from somewhere else in 2001-05, 863,000 were in the same place and there were 142,000 returners8Again ignoring all combinations with overseas, not stated and not applicable. The 4-year movement rate is 35.7%..

We can also look at movement within a year for households. In 13.9% of households9excluding Not stated and Not applicable, all household members lived at a different address a year earlier (this now includes those coming from overseas). Adding households where some members lived elsewhere increases it to 19%. I’ll express this as 13.9-19% as we look at how this varies.

B mlajbq26 B d5jp3try Looking at tenure type, renting households have a movement rate of 28.9-36%, three times that of owners. Renting is usually preferable for those planning to live in a home short-term.10Also, those wishing to stay long-term but renting have more chance of being foreced to move. Owned with a mortgage was 9.7-14.7%. Owned outright was very low at 4.1-7.3%. Click 1y→5y to see the 5-year mobility level for these categories.

Queensland has a higher proprtion of renters (33.4% v 30.1% nationally), but still B zopq1lyf a higher movement rate in all three main tenure categories.

O 5nwevu08 The movement rate for renting households correlates strongly with the socioeconomic status, with the Burnside and Unley SA3s turning over at twice the rate of Elizabeth, the poorest SA2 in Adelaide. For homeowners this correlation is weaker. For homes owned with a mortgage, those in poorer areas are more likely to have some (not all) people having moved in.

We can also break down households into different family/household types. B 3107w9q4 B ad7ho12u (Non-family) group households are highest at 27.3-49.2%. Lone persons are 13%. One parent families are higher at 14.3-21.1%. Couples are 13.2-16.7% while couples with children are lowest at 10.9-15.7%.11I was surprised that one-parent families are more likely than two-parent ones (6.9% v 4.8%) to have had just some people move in, because the latter category picks up cases where a new partner moves in, meaning an adult child or unrelated person is more likely to move with one parent than two. But this will also include aged parents moving in with their adult child. Queensland is still ahead of average in every major household category and near every major household & tenure combination. The 5-year household mobility level is 40.3-45.8%.

The high-movement categories will often coincide: Most likely to have moved are aged 20s, group households, renting, inner city. All these categories correlate with each other, some combinations quite strongly.

P vhi98sl0 This chart shows the previous SA3 of all residents of Para Hills SA2 who lived elsewhere in SA 5 years earlier. But much better than this, M t8g6qbr7 M naeyufg6 these maps show the source of people moving into Para Hills SA2 in 5 years. Most popular is Para Hills – many people move to another house in the same SA2 – and most of the main source SA2s are nearby. When you mouse over, the percentage in bold is the share of movers-in who came from that SA2.12The two maps differ only slightly. The first is coloured by people moving into Para Hills / total population; the second is people moving into Para Hills / total people moving from there. I thought the first would show up high-movement places like the CBD, but proximity dominates. The percentage used for the colour is in brackets; the one in bold is as a proportion of all the movers to Para Hills.

You can click on a region then click “Use instead of Para Hills” to choose a different key area. You can click ur⇆pur1p to go the other way and see the new residences of those living in the key area. In most cases, short moves will be the norm. Do move the map to a different state/territory, click the state/territory name in either Residence or Res 1/5y ago, then click region on the map, then “use instead”.

One of my main questions was where people in new housing areas have come from. M fypxl6bz This map lets us look for areas of high population growth.13I’m not showing the actual population growth, because I haven’t got the right shape of that data at hand. Instead, showing the totals for Residence 5y ago. I thought they would have more people from further away, but still most people are coming from nearby. Some of the interesting ones I looked at: M ujbn5efs Wollert (Melbourne) draws from a corridor to its south; M 8livnd4z Cobbity-Leppington (Sydney) draws from suburbs not too far away. 14 I was also wondering if there would be much migration between distant but demographically similar areas. There isn’t.

M p1c4gk68 We can look at the difference between migration in different directions. Usually there is flow both ways with nearby regions and little flow further away. There is often some flow towards new housing. There is net flow to Para Hills from the south (inc Ingle Farm) and northward from Para Hills to Salisbury East and some to Playford. Most suburban areas show net migration flowing away from the city. The city has net flow out in all directions because many people move there from overseas before finding cheaper accommodation elsewhere.

Australians of Greek ancestry have remarkably low movement rates of 8.1% for 1 year and 23.7% for 5 years15Using Ancestry 1st reponse = Greek. A bit of this imbalance reflects the age profile, but there’s a big difference in all age groups..

I still haven’t worked out why Queensland has a higher moving rate, across regions, movement categories, age groups, household types and tenure types, going all the way back to 2001. Are more jobs short-term? Houses don’t last as long? Less transaction costs in moving? Fewer grandparents tying families to one part of Brisbane?

Update Feb 2022: I found that Queensland has a divorce rate 15-20% higher than the Australia overall. (WA is also above average). But this alone has a small effect on moving: Each year 0.7% of Queenslanders divorce or are children of divorcing parents, compared to below 0.6% of Australians.