If your dashboards read ABS Labour Force occupation data, the OSCA occupation classification change will reach your pipeline this spring. From the September 2026 reference period, the Australian Bureau of Statistics moves all Labour Force Survey occupation data to OSCA, the Occupation Standard Classification for Australia. The headline release also returns to the third Thursday of the month, first on 15 October 2026. This post explains what changes and what to fix before that date.
What the OSCA occupation classification change involves
The ABS has set out four changes in one update. From the September 2026 reference period, every Labour Force Survey occupation figure is categorised using OSCA. The ABS has backcast OSCA estimates to 2014, so you can see the new categories running back over a decade. The Longitudinal Labour Force Survey restarts from the same reference period. And Labour Force, Australia returns to a fixed third-Thursday release, starting 15 October 2026. None of these is a small tweak if your work depends on the numbers, so it helps to take them one at a time.
- From the September 2026 reference period
- Occupation data moves to OSCA
- Available to 2014
- Backcast OSCA estimates
- 15 October 2026
- First third-Thursday headline release
- Restarts from the September 2026 reference period
- Longitudinal Labour Force Survey
Why this touches your pipeline, not just a report
A classification change is not only a headline story. If your pipeline joins tables on the old occupation codes, that join can break when the codes change. Worse, it can keep running while it quietly mismatches categories, so your numbers look fine and are wrong. Models trained on the old structure inherit the error. Dashboards built on those models pass it to the people who trust them, often without a warning that anything has shifted.
The risk here is silent, and silent problems are the ones that surface at the worst moment. A staffing lead asks why a demand line moved, a board paper cites a category that no longer maps cleanly, and the trail leads back to a join no one checked. It is far cheaper to test the change on your own schedule than to explain it after a stakeholder spots it first.
