#censusfail - Adjournment speech

2016-08-31

Thank you, Senator Lines. I would like to add my congratulations to you for taking on the role and responsibility of Deputy President. I am sure you will serve this chamber well.
It is a delight to follow Senator Siewert in the adjournment slot on the first day that parliament sits. I would like to thank everyone who supported me in taking my place in here over the election campaign, particularly my family, Senator Siewert and the formidable Greens team who worked so hard so that both of us could be here tonight.

 
While parliament was in recess, a process that most Australians would normally find largely uncontroversial blew up spectacularly in the government's face, severely damaging confidence in the Commonwealth's lead statistical agency, potentially placing at risk the privacy of millions of people and likely compromising one of the most important datasets used daily by policymakers and researchers around the country. I am talking of course about the great census fail of 2016.
In 2006, after a relatively robust and open process, the Australian Bureau of Statistics shelved a proposal to create a permanent statistical linkage key which would have substantially altered what the census is and how it could be used, from a once-every-five-years anonymous snapshot, to the creation of a lifetime digital identifier that would follow you like a unique barcode for the rest of your life. The best arguments on all sides of the debate were presented, and I acknowledge at the outset that there are arguments on all sides of this debate. The best arguments were brought to bear; specialists and researchers in the field presented their views. They were examined and summarised, and the conclusion that was reached in 2006 was that the process would amount to a substantial and unacceptable risk to people's privacy. And so it was abandoned.
Perhaps mindful of the setbacks a decade prior, in 2015 advocates inside the ABS for this method elected to conduct a rather private internal assessment of very similar proposed changes to the 2016 census. No civil society groups across this country participated in this process; no independent experts were consulted either from the pro or the anti side as far as we know. The results of this assessment and the changes to the census that were eventually announced happened with zero fanfare a week or two before Christmas 2015, after this parliament had risen for the year.
Privacy advocates and civil society groups—and ourselves, I must admit—were forced to run something of rearguard action, expressing views that should really have been incorporated into the ABS's planning from the very beginning. Former Australian Statistician Bill McLennan called it 'the most significant invasion of privacy ever perpetrated on Australians by the ABS'. These are strong words for someone who formerly ran that same agency. Many concerned groups and individuals, exasperated with the bad faith displayed by the ABS and total silence, stonewalling and dismissal by the government, began advocating a boycott.
So it was that, at that time, about March of this year, our office launched a petition to save the census, because we saw that the legitimacy of the privacy concerns—and they are entirely legitimate—had given momentum and credibility to a boycott movement, so much so that the veracity and the integrity of the census data itself was at risk. At that time the Turnbull government, I think, should have intervened. They could have held over the creation of the statistical linkage keys—this unique identifier—until such time as a transparent and thorough assessment of the proposal could be completed. That was what we asked Treasurer Scott Morrison to do, and the government failed to do that.
I understand that such linkage keys are in relatively common use amongst researchers and statisticians, and I respect the value of having longitudinal identifiers that can follow particular cohorts of people if you are working in the health sciences, disability services or as many other fields as you would like to name. But I also believe that the creation of such a unified identifier for every single Australian—not opt in, not opt out, but mandatory—should have been subject to a substantial public debate and a measured weighing of the costs and benefits. Such a debate will inevitably take place in the context of high-profile data breaches here and around the world, indiscriminate surveillance and data storage by signals intelligence agencies, and the potential misuse of such data either by authorised agencies or by unauthorised actors.
This is a conversation that needs to be had. I am not sure that this is necessarily a Left or a Right thing either. This is something the Greens have advocated strongly from, but I would have thought the libertarian wing of the Liberal Party—or actual libertarians like Senator Leyonhjelm—would have some views on something like this. I do not think there is a clean Left or Right political divide here; this is a debate that absolutely needs to be had by the light of day.
In contemptuously dismissing those concerns, the government guaranteed that the census data would, at bare minimum, be compromised, and, as we found out on census night, it was actually a lot worse than that. It is hard to imagine a more complete failure—from the ABS social media feeds encouraging people to fill out the census long after admins had pulled the site offline to the technically incoherent press conferences in the days that followed as blame was passed around and deflected from hand to hand.
Those technical failures, the spectacular planning oversights and the ministerial ineptitude have succeeded in some way, I guess, in diverting attention away from what was at best a rather arrogant and insidious abuse of process. This sort of secrecy and misdirection would fail a research ethics assessment at any Australian university. Maybe people are happy for their governments to do this on their behalf, but the conversation has to happen in daylight.
It is too late for the government to save face, but it is not too late to restore some faith in the process. At the earliest opportunity, whichever minister is responsible for the census this week should make a statement to reassure anyone who withheld their name, as I did—or their address or any other identifying information—that the ABS will not be seeking to pursue fines or other punitive action against them. The minister can direct the ABS to continue their efforts to collect census forms—I think that happens normally; that happens every census night—but stop them from hectoring and frightening people who are either unable or unwilling to complete the census, thanks in part to the catastrophic mismanagement of the whole process.
Our phone lines have been running hot, and I am presuming that this is not just because we took a high-profile position. I suspect other MPs' offices will be taking these calls as well from people anxious to know whether or not they will be punished as a result of this debacle. Obviously, they should not be. That is something that the minister, whoever he happens to be, could fix tonight. The thousands of people and organisations that rely on census data to make policy and proposals—and that obviously includes us and probably everybody in this chamber; we declare an interest here—will need better than that. Nothing less than a full inquiry into the lead-up and conduct of the census fail will suffice. We need to find out what was lost. How accurate is the information that we have, and how possibly could it have come to this? How badly skewed is the data that was collected? Do we need to cut our losses and start again with a better process? Should the census be held again and the one that was conducted a couple of weeks back simply be scratched?
Every Australian should be given the opportunity to make an informed choice. We will be joining moves this week to conduct a thorough Senate inquiry into the census fail, and I hope that we have cross-party support to bring fresh proposals to the table to clean up the mess. That means we would get to hear from the researchers who use this information in their daily work. We would get to hear direct from the ABS about cuts to their budget from this government and the corners that they may have had to cut to put the census into the field. We could hear from the technical community, who can step us through what on earth went wrong on census night. We could hear from digital rights organisations like the Privacy Foundation, Digital Rights Watch and Electronic Frontiers Australia and from specialist researchers like Rosie Williams and Asher Wolf, who have led the debate online.
Census fail was clearly avoidable, but the best thing that can come out of this debacle is a measured and thoughtful debate on data sovereignty, the role of the census in setting policy, the place of big data in our society and what happens when governments try to cut corners on census night.