Learn · Suburb fundamentals
Population and household growth and property markets
Census people, ERP, households and occupied dwellings are different objects. There is no official annual household series.
Districts Research · ·

People vs households
Dwellings are occupied by households. Population can rise while houses stay empty of new demand.
- HouseholdsThe demand unit for dwellings
- PopulationA different series
Swipe the panels.
The arithmetic
Does population growth in a suburb mean dwelling demand is growing?
Dwellings are occupied by households, not by headcount. Population can rise while household numbers stall, and household numbers can rise while population is flat. The official usual-resident series between Censuses is ERP. It does not come with an official annual household count.
Read people, households, occupied dwellings and stock additions as separate dated lines. Do not interpolate households between 2016 and 2021 and call the result a source. Districts shows Census population where loaded. It does not publish a household forecast.
Four objects that get called “how many people”
Buyers quote “the suburb grew 8 per cent” without saying which object grew. People, official usual residents, occupied dwellings and empty dwellings move for different reasons. Name the object, then do the arithmetic.
| Object | What it counts | Update |
|---|---|---|
| Census usual residents | People who usually live in the area, as counted and later used as the ERP base. | Five-yearly Census. |
| ERP | Official usual-resident estimate after undercount and residents overseas, then births, deaths and migration. | Quarterly for states; less often for small areas. |
| Households / occupied private dwellings | Groups occupying dwellings on Census night (QuickStats household composition sits on occupied dwellings). | Census. No official annual suburb household series. |
| Unoccupied private dwellings | Empty on Census night for any reason. | Census. Not rental vacancy. |
Worked arithmetic: Brunswick (Vic.) 2021, no yearly household invented
QuickStats SAL20361, Census night 2021: 24,896 people; 11,040 occupied private dwellings; average 2.1 people per household; 5,712 family households, 3,774 lone-person, 1,554 group. 5,712 + 3,774 + 1,554 = 11,040 occupied dwellings. That identity is a Census snapshot. It is not 2026 ERP.
If a later ERP for a matching geography is higher, that is more usual residents on the ERP method. It is not 11,040 plus a made-up household increment. Dividing new ERP by 2.1 (the 2021 average household size) invents households the ABS did not publish. To talk about household change you need two Census household or occupied-dwelling counts on the same geography, or an ABS household estimate that states its method. Between Censuses, write that the household figure is the last Census figure.
Read household change against dwellings added, same period
If two Censuses on the same geography show 400 more occupied dwellings, and the register shows 650 dwellings completed in those years, supply ran ahead of occupied-dwelling change by about 250. That describes those years. It does not describe the next five. Completions are not always occupied. See housing supply.
Traps that treat headcount as dwellings
These mistakes turn a people count into a dwellings story. Name the series before you quote the number.
- Quoting population growth as housing demand.
- Building an annual household series by dividing ERP by the last Census average household size.
- Reading a planning projection as a count.
- Comparing growth across unmatched boundaries.
- Treating unoccupied dwellings as vacant rentals. Holiday houses and dwellings between occupants sit in that Census count.
What a population count does not establish
It does not establish tenure, income or that growth will continue. Districts cannot say what has happened to household numbers since the last Census. Only population has an official intercensal estimate. Past growth is not future demand. A planning-department household projection is a model with stated assumptions. Read the assumptions. Do not file it next to ERP as if both were counts.
How Districts shows people
Suburb profiles show Census population with its date when loaded. Household statistics appear only where the source publishes them for that geography. Districts does not publish projections and does not describe suburbs as high-growth.
How to confirm which series you opened
On the ABS page, read whether the table is Census QuickStats, a Community Profile, ERP, or a household projection. Note the geography code and the date. Do not join two of those into a homemade annual household line. Household and family projections published by the ABS are models that apply living-arrangement propensities to ERP. They are useful as scenarios. They are not a substitute for a second Census count on the suburb you are buying in.
How Districts derives it
Explore using Districts
Sources
Official material this page used. Dates are when Districts checked the page, not the life of the instrument.
Australian Bureau of Statistics
2021 Census All persons QuickStats, Brunswick (Vic.) SAL20361
People, occupied dwellings, household composition and average household size for the worked arithmetic.
Checked
Australian Bureau of Statistics
National, state and territory population methodology
ERP construction from Census usual residents, undercount, residents overseas, then births, deaths and migration.
Checked
Australian Bureau of Statistics
Household and Family Projections, Australia methodology, 2021-2046
ABS household projections are models on ERP and Census propensities. They are not an annual suburb household count.
Checked
Related guides
Suburb fundamentals
How to research a suburb before buying
A dated workflow: Census versus ERP, then Districts medians, rents, pipeline and the lot. Do not mix periods on one line.
Supply and liquidity
How housing supply affects property markets
Stock versus flow: existing dwellings, planning approvals, building approvals, commencements, completions. ABS for macro, local registers for the lot.
Rental and holding risk
How to research rental demand in an area
Read renter share, household mix, anchors and letting times for one class. Census is a dated snapshot. Not a rent forecast.
Data literacy
Correlation vs causation in property research
Two series moving together is a prompt. Rail, school and population cases. Not a cause.
Research purposes only. Not personal financial advice, a valuation, or a planning certificate. Always speak to a licensed financial adviser before you act.