3 November 2017, National Gallery of Australia, Gandel Hall, Canberra
Concluding remarks from David W Kalisch, Australian Statistician
I would like to begin by acknowledging the traditional owners of the land on which we are meeting today, the Ngunnawal people, and pay my respects to their elders, past and present. I would also like to acknowledge other Aboriginal and Torres Strait Islanders who are with us today.
ABS has been proud to co-host this event with IPAA ACT.
I would like to thank all our speakers and panellists today for their insights and their time.
It reinforces my strongly held view that we benefit from providing a voice for people with a diversity of experience and expertise, and that these types of multilateral conversations are most useful to consider the full range of perspectives.
I would also like to thank the Conference Chair, Alison Larkins, the IPAA Secretariat and my staff at the ABS for facilitating this event and helping us to have such constructive discussions.
So how good is this, to have hundreds of people here today talking about data, new technical and technological opportunities and effective use of this resource for public purposes.
At this event today, we have a full array of people with data, policy and regulatory roles and responsibilities across our data eco-system.
At the outset, it is important to recognise that public data is a public resource that taxpayers have paid for and should benefit from. We, in government, are stewards of this data, and we have a range of responsibilities to protect and use this data for the benefit of the entire community. There is a balance to be struck, between safe use of the personal and business information that is provided to us, and enabling effective use of that data.
Increasing attention to data use and data integration is taking place in a changing and perhaps more challenging environment:
Amongst many in this room, there is significant appreciation and understanding of the importance and utility of data to inform key decisions, but we do need to recognise that is not universally shared across the community, or possibly even some decision makers.
There is clear evidence of reduced public trust in governments, key institutions and the media (Edelman Trust Survey 2017). In the data world, this reduced public trust has been amplified by a number of significant data breaches over recent years, particularly in but not exclusively the private sector, that has compromised sensitive or key identity information.
Expanded capability and reduced cost of technology is enabling us to bring together large data sets and draw key conclusions, while there is a very real and growing risk around cyber attacks and other disclosure risks that need to be effectively mitigated by those who hold, curate and integrate public data.
Those who manage administrative data sets now have a greater responsibility to properly curate that data (quality data and metadata, reasonable and safe access) in the context of its increased utility for broader data insights.
Data integration provides us with lower cost opportunities to deliver important statistical insights in the context of ongoing fiscal constraint. While data integration and enabling data access are not costless exercises, they are cheaper than collecting additional bespoke information from households or businesses, and without the respondent burden dimensions.
In the practice of public policy, data integration can provide new understandings of why programs are successful or not successful. We need to not only understand the outcomes that are achieved, but what is contributing to those outcomes if we are looking at the influence and importance of particular policy levers.
Today’s discussion focussed on aspects of innovation and culture, particularly as they relate to managing and releasing data. While much of the discussion today focussed on them as separate dimensions, I prefer to see them as integrally linked, as well as both important in their own right.
Today has showcased a number of the data innovations that are taking place in Australia, and many are world class. On the other side, there were many comments about the culture of data custodians, and some reference to the culture of “the community” to use of their data. One challenge I see at the moment is that innovation is potentially moving ahead of culture, and certainly ahead of community awareness. As some of the technical innovations are providing us with new techniques to use personal and business data more effectively and safely, the challenge shifts to ensuring there is also an associated major community engagement.
In October 2017, I provided a speech to an event held by the Committee for Economic Development in Australia, on the topic of data transformation, which considered some of these aspects in greater detail. This is available on the ABS website.
The data integration journey
While data integration has received more attention over recent years, it is not new in Australia or overseas. For example, the ABS has been using data integration techniques since the 1966 Census, in the processing of Census data and production of population estimates. The Australian Institute of Health and Welfare has been undertaking data integration since 1990.
Over the past decade, we have seen a number of advances in the techniques and use of data integration by a number of agencies – Commonwealth government, state government and the research community (especially health researchers).
However, we are not as far advanced with the practice of data integration in Australia as we should be, especially when we appreciate the rich data we have in Australia, including the increasingly valuable longitudinal datasets such as HILDA, LSAC, and many others in the health and social welfare environment.
With the new Data Integration Partnership for Australia (DIPA) initiative, announced by the Australian Government in the 2017 Budget, we have the opportunity to build some critical data integration infrastructure and essential collaborations in Australia over coming years.
ABS is proud to have been at the forefront of data integration in Australia for policy purposes, investing in our data integration capability and some flagship data integration projects to demonstrate how this is already providing new policy insights.
The ABS has been advocating publicly and within government for expansion of data integration activity, pointing to the international experience, including New Zealand, but also beyond.
The important DIPA initiative is an opportunity to further progress data integration capability in Australia, and embed further reliance on evidence for policy decisions. It is not the end game, as there is more to do to join up information across governments, particularly in the key policy areas of health, education, social services and infrastructure where all levels of government play important roles. DIPA is nonetheless a good further step forward along this journey.
Some key themes from the Conference
With such a rich discussion today, it is a challenge to produce some summary remarks, and do justice to the many valuable points made by all of our speakers and the responses to the challenging questions posed by the audience.
There was an overwhelming understanding that Australia is not using its public data to the extent we can and should, and this extended to underutilisation of aggregated readily available data as well as microdata that has more complex release requirements. Other developed countries are demonstrating some good practices we could consider.
There is a clear appreciation within people in this room of the benefits to the community from more effective use of public data. However, we need to get much better at explaining and communicating these dimensions. We need to stop treating this as so self-evident that it does not require elaboration, and use whatever opportunities we have to make the case in public forums.
Another key theme that I drew out of the presentations and associated discussions is that strong and enduring collaborations are required to generate more public value from use of our public data. Examples were highlighted of public/private partnerships, Commonwealth/State government collaborations and government/research partnerships
I would suggest that effective and successful collaborations need to be positioned around a higher purpose such as public value or public interest, rather than organisational interest, and will need to be scoped around mutual benefit and mutual trust across the system of data curators, data integrators and data analysts. In this complex data ecosystem, we need to recognise the skills, expertise and innovations already developed, and utilise this across our system rather than wastefully replicate capability in siloed agencies.
While it was discussed, one aspect that received less attention than I expected was the role of the citizen, and how we need to effectively engage with the public, to ensure there is strong community understanding and support for how we are proposing to use their information. The recent 2016 Census experience was one where those persistently advocating for more use of public data were, with few exceptions, comparatively silent in the public discourse. Those agitating for more effective use of public data need to take more responsibility for making the case to the community whose data we are seeking to use.
A number of our panellists today highlighted the importance of research questions, considering what key policy questions need to be answered and appreciating what data is useful for what purposes. Some of these core policy analytical skills need to be developed and practiced more, so we produce more comprehensive and compelling narratives that make best use of the available data. This requires us to be better at anticipating policy questions likely to be posed in the future, rather than playing catch up when issues hit the media and the Parliament.
We need to be wary about perfection being the enemy of the good. For many policy issues, we have sufficient data currently available to be able to provide high quality analysis. Data developments are being progressed, to add to what is currently available, and that is a good thing, but much of our current data are not used to anywhere near its capability.
In the Australian context, there is a major opportunity to move beyond analysis of data to a stronger evaluation culture and evidence-based practice in public policy more generally. We can provide better evidence about the outcomes from a range of programs and policies and ensure governments and other decision makers are better informed when they regularly make policy choices.