AUCN's research articles combine historical check-record summaries with practical buying and selling guidance. This page explains what those observations support, where they stop and how to quote them without overstating the result.
The current articles use a check extract ending 28 September 2026 and a historical listing analysis prepared in September 2026. The check extract's starting date was not established for this publication. Updating the wording on 23 October 2026 does not make those figures October observations.
Data families and observation units
| Dataset | Observation used | Supported description | Unsupported interpretation |
|---|---|---|---|
| Check results | A recorded check outcome | Share of historical check records carrying a summary flag | Share of all Australian cars or future loss probability |
| State groups | State entered for the check | Returned flag rates within those search-input groups | Incident location or state road safety |
| Vehicle categories | Category returned in the source record | Unadjusted differences between checked categories | Like-for-like causal comparison |
| Market listings | Advertised vehicle records grouped by model year | Historical average asking-price index | Completed sale price or individual depreciation |
The checks were initiated by customers or connected services. They are not a random sample of the Australian fleet. Repeated searches can include the same vehicle; the published aggregates have not established vehicle-level deduplication.
Definitions and interpretation
A finance-related summary flag indicates that the returned data needs security-interest review. It is not a loan balance. Written-off and stolen-related flags describe returned records that need their dates, status and codes read individually; a summary flag alone does not resolve the current legal position.
Mileage evidence is separate from the PPSR search. The PPSR's published search scope distinguishes its security-interest and returned vehicle information from information such as odometer readings and ownership history. Source coverage also matters: NEVDIS information can be incomplete, and stolen information for Tasmania requires a separate state check.
For the remaining historical flag charts, the displayed percentage is flagged check records divided by records in that extract or segment, rounded for readability. These summaries describe returned flags. They should not be interpreted as a verified prevalence estimate after correcting for unavailable source data.
Missing data, negative results and withdrawn claims
These states must remain separate:
| Result state | Meaning | Appropriate interpretation |
|---|---|---|
| Flag returned | A summary or source entry requires review | Inspect the underlying record |
| Negative result in an available field | No flag was returned in that field at that time | A dated result within the source's coverage |
| Unknown or unavailable | The field or source information was missing | No conclusion about the vehicle's status |
Odometer fields became unavailable in parts of the 2026 extract. The earlier overall anomaly rate and combined “any risk” headline did not use a sufficiently consistent coverage basis, so they are withdrawn from the current articles.
Earlier age bands used a common reference year rather than each check's date. Those age rankings have also been removed. A future age comparison would need age at observation, usable build dates and stable coverage before being interpreted.
Why groups are not rankings of quality or safety
State, category and brand groups can differ in age, model, use, check date and repeat-search frequency. The historical charts are not adjusted for those factors. A difference between two rates does not establish its cause, the significance of the gap or the risk on a particular car.
Brand aliases were not fully consolidated, and formal publication thresholds and uncertainty intervals were not established for the earlier brand tables. The brand league tables are therefore removed rather than presented as reliable rankings. No claim of statistical significance accompanies the remaining descriptive charts.
Asking-price calculations
The historical model-year chart divides each group's average advertised price by the average for the selected 2025-model reference group and multiplies by 100. The unit is index points, with 2025 set to 100. That reference is not a new-car purchase price or a claim that 2025 was the newest available year.
The extract applied broad plausibility filters to price, model year and odometer fields. Those filters do not establish that all duplicate, misdescribed or anomalous listings were removed. The collection window and listing-level deduplication were not established for publication.
There is no control for generation, trim, condition, kilometres, original price or seller mix. Accordingly, brand retention claims, fuel-type depreciation rankings and forecasts derived from these averages are not supported.
Periods, privacy and corrections
Article update dates describe editorial changes. Each quantitative article separately states its historical data period. No automatic refresh schedule is represented, and a partial month must be labelled provisional.
Public articles exclude customer details, individual vehicle identifiers and AUCN transaction or segment volumes. Percentages and indexes do not expose business totals, but they still require the interpretation limits above.
To cite a chart, include the article title, AUCN, the stated data cutoff and a link to the specific page. Describe it as historical check-record or asking-price analysis, as applicable. For a correction, use the contact route on the AUCN website and identify the passage and supporting source without sending personal vehicle records publicly.
The editorial and corrections guide explains source handling. The experimental check-record index has additional weighting and partial-period limitations.
Frequently asked questions
Are the observations unique vehicles?
No. The historical analysis uses check records. Repeat searches can include the same vehicle, and vehicle-level deduplication was not established for these aggregates.
Does updating the article refresh the dataset?
No. An editorial update date and an underlying data cutoff are separate. These articles state the historic period; there is no automatic monthly refresh implied.
Why publish percentages rather than business volumes?
Percentages communicate the pattern without disclosing AUCN transaction volumes. They still require a clear denominator and do not remove selection or coverage limitations.
Can the asking-price index be used as a depreciation forecast?
No. It compares different advertised vehicles by model year, not the same vehicles through time, and does not establish completed sale prices.