National

Save
Print

Census debate not enough: we need to rethink the Australian Bureau of Statistics

The bureau needs to be more than just another data collector.

The suggestion that Australia's national census change from a five to a 10-year cycle has sparked considerable discussion and reporting about the census's future. It's important to consider the ongoing nature and format of this major information-gathering exercise, but the debate should be much wider. The real question we face centres on the future of the Australian Bureau of Statistics itself.

The last census cost taxpayers $440 million and involved 29,000 collectors, who delivered 14.2 million forms to more than 9.8 million households. The result was a detailed and valuable snapshot of Australia and its population. The next census is due to be undertaken in 2016.

The current debate has centred on whether money should be saved by conducting a national census every 10 (rather than five) years. Some commentators point out that the savings could be used to upgrade the ABS's core IT systems and better prepare it for the future. However, we also need to discuss the most effective and efficient ways to collect information, and whether the data collected adds real value to our nation.

As part of this analysis, we must examine the ABS critically. As other government agencies increase their data-gathering activities and capabilities, it is important that the ABS's central function and role evolves to incorporate these changes.

For example, the ABS is increasingly hamstrung because of the legislated need to scrub collected data to ensure that individuals cannot be identified. While there is no question that individuals' privacy must be protected, "de-anomalising" data makes it very difficult to incorporate data from other agencies that has been collected and processed in a different way.

Advertisement

One option being canvassed is to amalgamate the ABS with another agency, such as the Australian Institute of Health and Welfare. If this happened, data from a range of sources could be combined and analysed more readily. Potential sources include Medicare, the Department of Immigration and Border Protection, and the Australian Taxation Office. Such a centralised store of multi-agency data would become an extremely valuable resource. Taking an open-data strategy such as this would also create new opportunities for individuals and businesses to gain fresh insights by using data that is much more up to date than that captured by a census.

The change would allow the ABS, in its new form, to provide valuable insights that previously would have been impossible. As a result, the need to conduct a traditional census every five years would be lessened if not removed altogether.

Consider this example. The National Centre for Vocational Education Research presently obtains data from the ABS to conduct comparisons with its own research. This might involve, for instance, the contrasting of training data with socio-economic data. Through tighter integration of such data sources, this type of analysis could be carried out seamlessly and allow much faster understanding of training trends across the country.

Similar improvements could be achieved by integrating data currently held by the ABS and Centrelink. By having these data sets in a single location, comparison would become much easier and more efficient, allowing social trends and changes to be more readily identified.

As these examples show, rather than relying on data collected every five years, making better use of sampling exercises and real-time data flows from a number of agencies can provide more accurate and valuable insights. By evolving its role in this way, and becoming an expanded peak body for information-gathering and analysis, the ABS can ensure it remains relevant and valuable in coming decades.

Steve Hulse is the chief executive of Space-Time Research, a business intelligence and data analytics company.