The ABS regularly uses administrative data to support the collection of data in business and household surveys. The ABS also uses combined data assets, such as those supplied by the Australian Tax Office (ATO), to develop labour statistics to provide unique insights into the Australian labour market from both a jobholder and employer perspective:
- Jobs in Australia (produced from the Linked Employer Employee Dataset)
- Personal Income in Australia (produced from the Linked Employer Employee Dataset)
- Weekly Payroll Jobs and Wages in Australia (WPJW)
More granular demographic and employer characteristics are available in these statistics than in survey based outputs, providing detailed insights into jobs, jobholders, wages and income.
ATO administrative data assets used in the generation of these labour statistics include:
- Australian Business Register
- Business taxation information (BIT) for owner managers of un-incorporated enterprises
- Client Register (CR)
- PAYG payment summary
- Personal Income Tax return (PIT)
- Single Touch Payroll (STP)
Differences from traditional collection methods
Administrative datasets are not typically designed with statistical production in mind. The underlying concepts relate to administrative policy or procedures, rather than statistical constructs. This can result in coverage, definition and quality differences compared to survey based outputs.
Administrative data can cover large populations in more detail and therefore provide different levels of insight than traditional collection methods. However, administrative data can capture very specific populations or sub-populations, compared to surveys which collect information from a representative sample of the population.
While more detailed statistics are available, the estimates may present unique views of the population, particularly where adjustments are not applied to broaden the population represented.
In addition to more variable reporting quality, administrative datasets are significantly larger than those obtained from business and household surveys. The timeliness of outputs is weighed against the quality assurance of individual records. As such, the increased speed of statistical production may require macro level adjustments to address anomalous reporting (such as in WPJW). Where more time and information are available to resolve anomalous records, micro level adjustments may be in use.
Producing statistics from administrative data requires an alternative approach to processing and quality assurance than those used in survey based statistics. However, administrative datasets are an increasingly valuable source of new data, providing a rich variety of alternative insights into the labour market.
More information on the methodologies supporting the current suite of Labour statistics derived from administrative data can be found in the respective statistical releases: Jobs in Australia; Personal Income in Australia; and Weekly Payroll Jobs and Wages in Australia.