If you have ever pulled two population figures for the same city, local government area or state and found they disagree, you have probably not found an error. You have found two different questions. One asks who was physically in Australia on census night. The other asks who actually lives here.
This guide explains what each measure counts, why the Australian Bureau of Statistics (ABS) does not simply publish the census headcount as the national population, and how to work out which number answers the question you are actually asking.
The two questions behind the two numbers
Start with the definitions, because almost every mismatch traces back to them.
The census counts presence. The ABS's Understanding Population Measures describes the census as counting people "in Australia at a particular time regardless of where they usually live." That is a snapshot of everyone physically present on census night. A tourist staying in a hotel, a visiting worker, or a relative home for a few weeks all get counted. A resident who happened to be overseas that night does not.
The Estimated Resident Population counts usual residence. According to population.gov.au, ERP is Australia's official population concept. Its test is residence over time: people living in Australia for at least 12 of the last 16 months. National figures are published by the ABS in National, state and territory population.
Those two definitions overlap heavily, which is why the numbers are usually close for large, stable populations. They diverge sharply wherever a place has lots of short-stay visitors, fly-in fly-out workers, students, or residents who travel.
Why the ABS adjusts the census rather than publishing it
ERP is not a separate survey. The ABS states plainly that "the Census provides the basis for the production of Australia's official population estimate," and that ERP is measured "by taking Census data and adjusting it."
That single sentence explains the whole relationship. The census is the starting point, not the endpoint. The ABS then reworks it so the result matches the 12-of-16-months residence rule rather than the who-was-here-that-night rule.
Practical interpretation of what that adjustment has to achieve, reading the definition literally:
- People present on census night who do not meet the 12-of-16-months test should not be in the resident population.
- People usually resident in Australia who were temporarily away should be in it, even though the census missed them.
- The result is then carried forward over time, so ERP exists for periods between censuses and for dates long after the last one.
The ABS is responsible for the method and the published figures; if your work depends on the precise adjustment steps, read the current ABS documentation rather than inferring them from the definition alone.
A comparison you can actually use
| Question | Census count | Estimated Resident Population |
|---|---|---|
| What is it? | A count of people in Australia at a particular time, regardless of where they usually live | The official population concept, based on living in Australia for 12 of the last 16 months |
| Who is included? | Everyone physically present, including short-stay visitors | People whose usual residence is Australia |
| Who is missed? | Residents temporarily away on the night | Visitors and short-term arrivals |
| How is it produced? | Measured directly in the Census by the ABS | Produced by taking Census data and adjusting it |
| Where is it published? | ABS census outputs | ABS National, state and territory population |
| Best for | Detailed characteristics of a place at a point in time | Tracking how many people live in an area over time |
The practical rule: if your question is "how many people live here," you want ERP. If your question is "what was the population here on census night, and what were those people like," you want the census count. Mixing them in the same calculation is the most common source of confusion.
How to compare two figures before you trust either one
Run through this checklist whenever two sources disagree:
- Identify the measure. Does the source say "Estimated Resident Population," "ERP," or does it just say "population"? Many secondary sites do not label it, and that ambiguity alone can explain a gap.
- Check the reference date. ERP is produced for specific points in time. A census figure belongs to census night. Comparing a census night count with an ERP from a different period compares two different moments.
- Check the geography and its boundaries. Statistical and administrative boundaries change between releases. A figure for an area whose borders were redrawn is not comparable with an earlier figure for the "same" area.
- Check whether the figure is a count, an estimate or a projection. These are three different things, and only one of them is a measurement of people who were counted.
- Check the revision status. Population estimates can be revised after a census. An older ERP and a revised ERP for the same date may differ legitimately.
- Check the basis of any overseas comparison. Some administrative and international datasets count people based on citizenship or residency permits rather than actual residence. Those figures will not reconcile with ERP, whose test is time spent living in Australia, not legal status.
Population clocks and world figures are a different category again
The US Census Bureau's international population clock illustrates the point clearly, and it is worth understanding before you use any live counter.
That tool states that populations shown for the most populous countries and on the world map are projected to 1 July 2026. It also states that the US population total and population change have been adjusted to be consistent with the results of the 2020 Census, while the components of population change have not been adjusted — so inconsistencies will exist between population values derived directly from those components and the population displayed in the odometer and the Select a Date tool.
Two lessons carry across to any Australian work you are doing:
- A clock reading is a projection, not an official count. Do not reconcile it line by line against a census result and expect the arithmetic to close.
- When one part of a dataset is adjusted to a new census benchmark and another part is not, internal inconsistencies are expected and documented. That is a disclosure, not a defect to work around.
Treat population clocks as orientation tools. Treat ABS ERP and ABS census outputs as the reference points.
What this means for everyday research
Suppose you are comparing two suburbs for a report. Pull the census figure for one and an ERP figure for the other and you have built a comparison on two different definitions. Your conclusion will be wrong in a way that is invisible on the page.
The safer approach is to choose one measure and hold it constant across every area you compare, note the reference date and the release, and state which measure you used. If you need both, present them side by side and label each one rather than merging them.
For small areas especially, be wary. The presence-versus-residence gap is not spread evenly across the country. A tourism centre, a university precinct, a mining region or an inner-city area with high visitor turnover can show a visible difference between the two measures, while a stable suburban area may show almost none. The direction and size of the gap depends on local conditions, so check what the ABS says about the specific area rather than assuming a national pattern holds locally.
Your next step
Before you cite any population number, write down three things: the measure, the reference date, and the source publication. If you cannot fill in all three, you are not ready to compare it with anything.
Then go to the ABS National, state and territory population release for the official ERP figure and the relevant ABS census output for the census count, and read the explanatory notes attached to each. Those notes tell you the reference period, the geography in force, and whether the figures have been revised — which is exactly the information you need to decide whether two numbers are genuinely in conflict or simply answering different questions.
General information only
This article is general information about Australian public data. It is not legal, tax, migration, credit or financial advice, and it does not account for your circumstances. Population measures, census reference dates, geographic boundaries and published estimates change over time, so verify current figures with the ABS and population.gov.au before relying on them. Australian Today is an independent information publisher; it is not a government body, regulator, lender, broker or comparison service, and it does not produce official population statistics.