Much has already been said and written about Australia’s 2016 National Census, and there will be much more to come. Privacy concerns, data protection issues, conspiracy theories about the motivation of government in using personal data, poor change management, public apathy and ignorance, and the final ignominy of the online system crashing on Census night have made the 2016 Census one to remember. The momentum for digital government in Australia will take a significant hit.
The intention of this blog is not to dissect what went wrong in the 2016 Australian Census; a procession of inquiries and reports over the coming months will do that. What I would like to address is one of the most contentious issues of the 2016 Census: the linking of personal information over datasets collected at different time intervals and circumstances. This issue does require more public attention, and hopefully it does not get buried in the aftermath of this digital disaster.
Data linkage is an example of digital’s potential to radically change the game for public policy development. Digital can enable longitudinal data sets to be created from census and other datasets to provide deep insights leading to innovation in public policy. A census reminds us that in this digital age, the fundamentals of government remain constant. What is changing is how government goes about its role to protect, provide, and ensure the nation state prospers.
How many of us think about how a government goes about developing new policies? Some people might think that new policies are just thought bubbles of politicians, and there may be cases where this is true. But in the main, new policy comes through a rigorous process that includes an analysis of information and data—exactly the sort of information that comes from a census and other statistical survey instruments.
Governments have been collecting data about their population since biblical times. Was this the beginning of a great conspiracy to control populations, or simply the gathering of information to facilitate the development of new public policy? Indeed, there are instances in history where governments have misused information against their own citizens, but in general, governments collect information about people to create public value through new policy proposals. As part of the social contract, we as citizens “give to get” – giving up some of our personal information in return for evidence-based public policy.
A defining feature of a national census is the “snapshot” of the nation at a single point in time. These snapshots provide invaluable insight and evidence for public policy-making. There is, however, a deficiency with the snapshot model – no direct link between one snapshot and the next. For example, there is no link between a person who identified as unemployed in one census and employed in the next – these are discrete elements within the snapshot of data.
From a policy perspective, understanding some of the characteristics of the person who transitioned from unemployed to employed could provide insight leading to new policy initiatives to help the unemployed find stable employment. The same applies to someone who identified as homeless in one census to living in stable housing by the next. How and why did that happen, and were there any patterns of characteristics identified which contributed to success across a cohort of unemployed or homeless people? This is valuable information that enables governments to understand when, where, and which programs and interventions work, and which do not.
Public policy makers understand this concept as a longitudinal data set – data that tracks the changes in circumstances of a person, family, or even a community over time. One only has to think of the highly successful documentary series Seven Up, which traced the fortunes of a group of British children from a variety of backgrounds and different areas of the UK, returning at seven-year intervals to take snapshots of their lives. It features the same people every seven years and their life path is followed and analysed – albeit in a very public forum.
Establishing a longitudinal study takes time and is not easy to set up: By definition, it takes time to see results from such a study. Digital technology provides the capability to link data between census snapshots, thereby creating a longitudinal data source. Unlike the highly public nature of the Seven Up series, this can be done by anonymising the data and using a linkage key to keep track of changes over time. For some, this is a bridge too far in terms of government surveillance of their lives, and this was a major factor in the negative public discourse over the Australian Census of 2016.
However, we need to stand back for a moment and see data linkage for what it is in terms of intent: a valuable research asset for public policy rather than an invasion of privacy. To move on, one needs to come back to the fundamentals of government and the level of trust we have in our government institutions. Do we have legal recourse when data is misused? Are governments transparent and accountable in how data is used? If we believe our governments have good intent (which is perhaps not always matched with capability to realise the intent), then perhaps there is not so much to fear as policy makers access more and better quality data in a longitudinal manner. The pieces of data you give up over several years may be the missing link to solving one or more of the many social and economic issues our societies face.
Clearly, many things went wrong and were not well managed with respect to the Australian Census of 2016. As the dust settles and the inquiries start, we should not lose sight of the value the digital age is providing with new capabilities to enable the business of government. These new capabilities have the potential for value creation, but like most things in life, this value does not come without risk. The public debate that has to be had should focus on the public risk/reward equation while recognising that condidence and trust need to be established in terms of the perennial question, “Is my personal data safe and secure?”
Widespread public and government backlash against the collection of personal information will be a consequence of Australia’s Census of 2016. The extent of the fallout will be determined by how quickly the public debate can be re-focused on the issues that matter. We should be debating how our collective personal data, de-identified and linked to provide a longitudinal perspective, contributes to solving significant societal problems such as long-term unemployment, indigenous disadvantage, youth social exclusion, rising health costs, child abuse, violent extremism, and family violence. Once we can agree these are the problems that matter, we can focus on how digital technology contributes to solution options and desired outcomes. At the same time, we need to assess the risk appetite within society to achieve these desired social and economic outcomes. Most importantly, trust and confidence needs to be established to assure the public that their information is safeguarded while under the custodianship of policy makers as they find solutions.