Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Joint Standing Committee on Electoral Matters
15/02/2012

CLARKE, Dr Roger, Chair, Australian Privacy Foundation

Committee met at 09:38

ACTING CHAIR ( Mr Somlyay ): I declare open this public hearing of the Joint Standing Committee on Electoral Matters for its inquiry into the Electoral and Referendum Amendment (Maintaining Address) Bill 2011. This bill proposes that the Australian Electoral Commission, AEC, be able to directly update an elector's enrolled address following the receipt and analysis of reliable and current data from third-party sources. Today the committee will hear from the Australian Privacy Foundation, which has expressed some concerns about the bill in its submission to the inquiry. The evidence given today will be recorded by Hansard and will be covered by parliamentary privilege. I remind members of the media who may be present at this hearing of the need to fairly and accurately report the proceedings of the committee. I welcome Dr Roger Clarke. Do you have any comments to make on the capacity in which you appear?

Dr Clarke : I am an electronic business consultant of long standing. I am also a visiting professor at the ANU, in computer science, and at the University of New South Wales, in the faculty of law. I am also a member of Civil Liberties Australia—which is said for the benefit of an absent member of the committee. I am appearing before the committee in my role as Chair of the Australian Privacy Foundation.

ACTING CHAIR: You should be aware that these hearings are legal proceedings of the parliament and therefore have the same standing as proceedings of the respective houses. The committee has received a submission from the Privacy Foundation. Thank you for agreeing to meet with the committee. We have today received a submission from you, which I am sure members of the committee have not had a chance to read, so I will ask you to make some opening remarks and then run us through your supplementary submission.

Dr Clarke : My apologies for the lateness of the submission, but it was researched and written yesterday, reviewed by the board last night and finished between five and 6.30 this morning, so it could not have been in your hands any earlier.

The APF provided a submission to the committee in relation to the inquiry back on 3 February but it was, of necessity, very brief. It was prepared by several other members of the board but submitted over my signature as chair. I had not previously been directly involved in this particular matter, but I spent my time yesterday preparing for this appearance. We thank you for your repeated invitation to provide evidence and we note the importance of our evidence to you, as indicated by the media release put out by the committee on 13 February.

The paper has several sections to it. What we are firing off is the statement that was made in our first submission, which is that the changes need to be assessed in the wider context, and it is therefore important that we provide a bit of background on that context. We can then more briefly make the points we need to make about this bill. I will commence with five separate segments of that context. I will then move on to the specifics of the bill itself and then I will draw attention to some procedural aspects of the matter as well.

The first point relates to mandatory voting. I think this can be brief, particularly in view of the discussion that the committee had in the previous hearing. We submit that the right to vote is tarnished by also being an obligation to vote, and more than being just an obligation to vote it is an obligation to enrol, to sustain enrolment accurately, to expose personal data, to accept that that data will be expropriated and to be subject to enforcement actions.

It is one thing in historical times, when representative government was the only option that was possible, particularly in a large land like this, but the world has changed. As I mentioned at the outset, I am an electronic business consultant. We have gone through the e-commerce phase, we have gone through the e-government phase and we are moving into the e-democracy phase. There are many, many embellishments that are already occurring to representative democracy and it is going to change its form. From the citizens' perspective that means that there are many, many ways to be politically active, and going to the ballot box is only one of them. Many people wish to exercise that as a right, but of course what this does in the Australian system, as it stands since 1925, is to impose it as an obligation. That creates an environment that is quite different from the exercise of a right.

The second aspect—and we get to bigger and longer segments as we move on—is address. Section 99 of the act stipulates that the enrolment of electors is based on 'the address at which the person is living'. Jumping ahead in the submission, I note that the AEC's definition in its FAQ is of a 'real place of living', which is a rather more complex concept, and I am not entirely clear; I have not found the location where there is a reconciliation between the concept in the act and the concept as implemented.

The idea of address worked fine when suffrage was based on ownership of real estate, because real estate does not move; it stays within the electorate that it is in at any given time and therefore it is quite easy to pin down the relationship between the existence of a right to vote and where that right to vote can be exercised. As soon as you move to where a person lives, you move into a wide range of difficulties. What we have done here is identify a range of those things. I would like to make it quite clear that all of these things are well known to this committee, but we do need to rehearse them.

The way that we put it is that the concept such as address at which a person is living embodies a presumption that a person only lives in one place all the time or is reliably and legally associated with one place at the time relevant to the determination of their voting rights in respect of any particular election. It embodies that presumption, and for many people it is difficult, ambiguous or even unresolvable to come up with an answer to that question. That has been tackled by the AEC over many years, since soon after 1901, but it actually becomes more difficult as time moves on.

I have in here a figure which is a figure from past research, of 15 or 20 years ago, of 20 per cent of people in any given year being at a different place at the end of the year from where they were at the beginning of the year. I see in the evidence from last week that the figure of 16 per cent was used by the AEC, from more recent Bureau of Statistics figures. It surprises me that it has gone down rather than up over a 20-year period. That could bear some checking, but the fact remains that there are many people who move and there are many people who move multiple times within a period of time.

For the purposes of the electoral roll, annual is not sufficient; it is the cumulative over a period of time between elections that is relevant. The obligation to maintain accurate addresses when there is no actual need for the change of address to be notified because there is no election impending is another one of those impositions that you really ask yourself about. Since it is only an offence nominally and it is not enforced, as we have seen from the evidence, we do not spend our time haggling over the detail, but it is an anomaly.

I would quickly draw attention to one point on page 2, which is that, among the different categories of frequent movers, multiple addresses and no tenable addresses—of which there are many, many categories; politicians of course are very well aware of this, being one of the many special cases—casualisation of labour, which is well and truly in train in Australia, will result in significantly more mobility and significantly larger numbers of people who spend four or five nights a week living a long way away from home. Four nights versus three nights actually adds up to a majority. You start asking yourself what 'real place of living' means in such circumstances.

The AEC obviously has means of addressing these sorts of issues, and the notion they use and the notion that the parliament has given them to work with is 'itinerant electors', under section 96. This does not in any way relieve the individual of any of the responsibilities and arguably adds to them. It forces the individual to remain linked to the electorate containing their last address. It is only in extremis that the AEC will permit them to be linked with the electorate with which the person has the closest association. So once again we have interference between what the parliament has set and what the AEC administers and what individuals would arguably like to do.

With address, one of the biggest problems that has got to be confronted here—and directly relevant to this bill—is that the relevant address under electoral law is subject to particular definitions and interpretations, as discussed, which are quite different from those that apply to addresses used by other government agencies. I need to stress this time and time again in all of the submissions that we make to untold numbers of parliamentary committees. The way in which systems are designed, the way in which data definitions are made of data items like address, like spouse, like child, in each government agency for each program reflect the needs of that agency and that program, and they are different in every circumstance. We will obviously lead up to this and use this point when we talk about data matching shortly.

The third of the five aspects of context that we need to talk about is the untrustworthiness of the electoral system. We have to be as blunt as that because that is the nature of the beast and the problem with the lack of trust that exists in the community. I have been in the IT industry for 45 years and it is quite clear that some personal data needs to be handled by the Electoral Commission in order for it to perform its functions. Everybody is in agreement with that. However, it is a fundamental tenet of information privacy that data collected for one purpose should not be used for other purposes nor disclosed to other parties. Electoral data, on the other hand, is authorised by the parliament to be used and to be disclosed to large numbers of organisations and for purposes far beyond the administration of elections. We identify a number of these in the submission and provide pointers to the sources, because clearly this document will be used for additional purposes above and beyond our submission to you. Most members of the public are astonished when they discover the manner in which the parliament has, in some cases intentionally and in others while sleep walking, gifted the data that citizens supplied for the purposes of electoral administration to other organisations entirely. Something that I do not deal with as well as I should in here but which I have picked up from the Hansard of your previous hearing is that Ms Bishop noted the extensibility that has been delegated to the commissioner. And the commissioner clarified that, while he has at this stage only designated a small number of organisations, yes, he has the power. Parliament has gifted him the power to extend that—powers that should never be in the hands of employed officials and that should remain in the hands of the parliament.

I have not at this stage been able to identify the full list of data items that are released and under what circumstances. I have found derivative documents but I have not actually found the authorities. This is indicative of the problems that a member of the public has when they try to understand just what the extent of the breach is that the parliament has achieved with electors' data. There are protections but those protections that exist are incapable of shutting the door after the horse has bolted. They are weak, they are weakly administered and they do not recover the trust that has been lost.

People's cooperation with government agencies and specifically with the AEC depends on trust. A small proportion of people have got good reason to be concerned about this. People who obstruct or avoid mostly do so quietly and hence are fairly difficult to detect, and that was at the heart of quite a few of the discussions that you had with the AEC the other day. The point that we make here is that visible conscientious objectors are always just the tip of the iceberg of discontent. What we would argue is that it is surprising that the participation rate is as high as 92 per cent—or whatever the number is at the moment—and we would predict that, with the distrust that exists in the community, the participation rate will drop further.

Something that we draw strongly to attention is that the AEC has failed to draw to this committee's attention and has failed to investigate positive ways in which higher participation levels could be encouraged. There are two particular points that we draw to attention, one here and one a little later in the submission. The one here is proper controls over use and disclosure—that is to say, removal of large numbers of the authorised disclosures and authorised uses that the parliament has provided. That would recover some of the public trust. The other point will have to do with consent based cross-notification instead of imposed cross-notification, as is proposed in this bill.

The fourth aspect of the five contextual matters—this is at the top of page 4—is silent electors or, perhaps more broadly, the needs of many people to suppress data. I think it is clear from the discussions previously in the last hearing that the committee is well aware that lots of people in society have something to hide. In general, everybody has at least something to hide—some people more than others. For some people it is for nefarious reasons; for many people it is for reasons not of their own fault and reasons which the public respects. I provide a list here of the standard examples that we use when we argue these positions.

Mrs BRONWYN BISHOP: Do you want to suggest [inaudible]

Dr Clarke : Yes, certainly. Are there any particular ones that draw-

Mrs BRONWYN BISHOP: Just whatever, just so we have got it in the Hansard.

Dr Clarke : One that frequently arises is the second of the bullet points at the top: people concealing themselves from previous partners or friends. Stalking and, in recent times, cyberstalking, is quite common. Stalking is not only of celebrities. Victims of domestic violence are of course the extreme end of that problem. We are not suggesting that these are things that the AEC is unaware of and does not deal with, but there are a great many of these circumstances and very few of them are directly supported by government. Protected witnesses and undercover operatives, which is a subset of the very last category, are the only forms which are directly supported by governments in Australia. The rest of those people have to fend as they can, and in this case the point that we make in this section is that, where people suffer from these difficulties, they have to fight with the electoral commissioner, they have to apply, they have to disclose a considerable amount of distressing information—which is a further source or vulnerability for them—and they then have to fit into the very narrow constraint of 'because it places the personal safety of the elector or members of their immediate family at risk'. That is the only head that they are allowed to argue from. Then they depend on the grant of the discretion by the Electoral Commissioner. This is not something that represents care being taken by the federal parliament or by the Electoral Commissioner of the many people in Australia who are at risk.

Mr GRIFFIN: This is not really part of the bill, though. Are you suggesting it should be?

Dr Clarke : I will suggest at the very end that this is one of the changes that should be made in order to achieve greater public trust, and in order to improve the participation rates on the roll, which are alternatives to this punitive approach that is being adopted in this bill. But yes, you are right; I am going outside the bill—

Mr GRIFFIN: I do not have a problem with that, but I wanted to make sure I had that right.

Dr Clarke : We are still in the contextual phase at this point. This bill adds to a whole pile of problems, it builds on problems and exacerbates existing problems.

Mr GRIFFIN: My point though is that some of these problems are actually in existence in themselves, and the bill in itself—well, I think we can argue the toss about how much it might exacerbate.

Dr Clarke : Certainly I will submit that to you shortly. I believe I have made that point. I am now on to the fifth of the five contextual aspects on page 5. I have used the AEC's term 'continuous roll update', or CRU, for convenience, because I have relied upon their data because it is extremely difficult to perform the research especially in the space of one person-day in order to identify all of the organisations that are involved in this. The AEC is the source of data for vast numbers of other government agencies. We have established that. But the AEC is also authorised to expropriate data from other agencies and to re-purpose it. The first thing that they re-purpose it for is electoral administration. That is one line of argument that represents one balance between the interests of privacy and the many other interests in society, and all privacy protection is about balancing interests. But then it is further re-purposed because, once that data is in the electoral roll, it is then re-purposed and re-disclosed once again by the Electoral Office. What I have not been able to do is to nail all of the authorities that have been granted, but we have established that the parliament has gifted the Electoral Commissioner full power to decide whatever he likes in the way of data acquisition.

There are declarations by the AEC about three key things. We draw to attention that those three key organisations that are used as sources. The first is a Commonwealth agency, Centrelink. Centrelink is merely a funnel for the 100 welfare programs that are run in Australia, which are formally administered by in the order of twenty different agencies. So, when we say that there is one Commonwealth agency involved, there are twenty, and there are 100 programs that are being sucked in through those Centrelink accesses. The second set is state and territory government agencies. The mechanics are that the data is acquired from NEVDIS but that data is sourced and is acquired in the first instance from citizens by motor registries. The third is a completely different category again, which is a government business enterprise. We have crossed out to the grey zone of government in the form of Australia Post.

Contrary to the sentence I have said here, 'It is unclear whether this list is complete': I am sorry, the evidence—which I read subsequently to finalising this submission, at 6:30—last week said that that list is currently complete but totally extensible. That means that at this stage private sector sources are not used but there is absolutely nothing stopping the Electoral Commissioner from deciding that he will become a subscriber to Veda Advantage, the credit bureau, and that he will become a subscriber to ACXIOM, the consumer profile aggregator in Australia and elsewhere, and absorb that data into the electoral roll as well.

This data is matched. I could have written, and have written in the past, lengthy treatises on data matching—how it works, what its deficiencies are and what controls ought to be imposed upon it. I have kept it mercifully brief here, I think you will agree; it is only three short paragraphs. The point about data matching is that it is extraordinarily error-prone. It is based on, firstly, name; secondly, usually, elements of address and, thirdly, date of birth. Date of birth is commonly unreliable. People fib about their ages. Many people are not very pleased about having to disclose their ages, and that includes males as well as females. Address in this context cannot be used because the whole purpose of the study is to come up with different addresses and therefore you cannot match on it. So you have you got to reduce quality of data matching in this data-matching program compared to all the other data-matching programs that go on in government.

Name is enormously variable in its recording and is routinely 'scrubbed'—that is the term used—in order to try to muck around with the data, modify the data, in order to make it seem right. It is differently scrubbed by every different agency, so we have differential collection for different purposes in different ways with different data-quality measures with different data-scrubbing measures, and then we bundle all this together and match it. The false positives that arise from this are enormous, as indeed are the false negatives, because there are enormous numbers of occasions where matches could in principle be discovered which in fact are not discovered by the algorithms that are used. It is extraordinarily error prone. In circumstances like these you would think enormous care would be taken, enormous justification would have to be provided, proportionality would be taken account of and it would only be done when there are very serious benefits to be gained. Unfortunately that is far from the case. The fundamental reference in the literature is one that I wrote 18 years ago, but I have provided you the reference to that.

All of this is not the subject of consent, it is authorised by law. That is the framework that we are working with before we even move to this bill. What does the present bill do? Despite the attempts by the electoral commissioner to represent this as a minor adjustment and a very minor extension to what the AEC is already doing, it is a massive change of direction. We have submitted that, if this were to go through the parliament, this would represent a very large step into territory where the Australian government has never previously ventured, because the results of the data matching are to be used not merely as a basis for contacting electors with a view to achieving updates to the roll but to make changes unilaterally by fiat. This inverts the process from its present consent based approach, which Americans call 'opt in', to an imposition, or as the Americans call it 'opt out'.

There are two particular perspectives that can be taken to this. One is that it is authoritarian, that it is an exertion of power by the state over the populace. Alternatively it can be seen as being paternalistic, because it transfers power over people's lives to an organisation that is asserted to be more capable of looking after them than they are themselves. The primary justification advanced for this extraordinary change in the balance of power between the individual and the state is that:

Providing the AEC with authority to directly update an elector's address will mirror other roll maintenance processes where the AEC can directly act in such a manner.

A couple of examples are given by the AEC. The examples are not in the least bit commensurate with the new proposal. The objection process that is one example that is provided is quite limited and what is more, as we point out later, is based on positive evidence, unlike this approach which will be based on negative evidence. This was raised in the discussion the other day, so I will not flog that particular point. In the second case, it is the deaths register. In the case of death, the elector is no longer able to perform the action themselves, is no longer able to submit a complaint of breach of privacy. The concern about paternalism is far less in the case of dead people. In other words, the word 'mirror' in that justification is completely inappropriate.

The AEC presented two further grounds. One, in my words not theirs, is that it is just a further application of the existing objection process. That submission by the AEC is completely spurious. The receipt of a return-to-sender or left-address message is positive evidence that the old address is no longer relevant. An inference, on the other hand, that is drawn from an error prone matching process that is then coupled with the non-receipt of a response to a letter represents negative evidence. It is far from sufficient to be treated as confirmation of the inference that was already a risky inference from the data matching. The two mechanisms are simply not comparable. Further, the suggestion was made by the AEC that:

A proportion of these individuals were effectively disenfranchised in the old arrangements by prescriptive legislative requirements that they did not clearly understand.

That applies to the new provisions, just as it does to the existing ones. It does not represent any form of justification for the proposal. The other ground that was put forward was—again in my words—the community expects to interact like this. This is spurious, but I have to say it is for a different reason: looked at blandly, the community does indeed expect that government agencies will move with the times and take advantage of and make available improved electronic mechanisms. That is not in dispute.

Many people would welcome a consent based cross-notification arrangement. Those kinds of arrangements have been proposed or half-heartedly implemented in many contexts. I have mentioned the public kiosk era. I did work in my electronic business consultancy prior to the advent of the internet—we are talking here about 1988, 1991—for the state of Victoria right through the e-commerce phase and e-government phases. There have been multiple attempts to do this. I and colleagues—not the APF, I am sorry—in a consulting capacity have consulted various government agencies at federal level proposing, and in some cases helping through to implementation, ways of achieving that. What the AEC is proposing is not consent based, it is imposed.

There are also attempts to claim that governments are much more trusted than businesses or charities, because only 12 to 16 per cent of people said that they avoid government agencies that cross-reference or share information. The AEC even repeated that in verbal evidence the other day. The claim is completely undermined by the fact that the government agencies are legislative monopolies. They are unavoidable. People have to take advantage of them—or in some cases not take advantage of but be taken advantage of—by interacting with government agencies. It is the law. For that reason the questions that were asked by the Privacy Commissioner when the Privacy Commissioner did surveys are, from a research perspective, nonsense. They will always produce nonsense answers, because you will have different responses and different interpretations by the respondents. Some respondents will say: 'I don't have any choice. What you mean? It's a nonsense question. I'm forced to accept it.' You will get confounding of the variables and the measures are meaningless. In such cases, the change from 16 to 12 per cent between two different surveys would be a meaningless statistic which the AEC depends upon.

Each of the grounds that the AEC has presented is valueless. These are not the only things in the AEC's submission that contain far less quality logic than should be in documents presented to the parliament. They impliedly claim that a 0.1 per cent objection rate to the New South Wales Electoral Commission supports the proposition that the public do not care. Few people ever object for many different reasons, and apathy is a large part of it. Many people lack the skills. Many people come to the AFP and simply say, 'We'd love your help to prepare a complaint.' But we cannot support them. We do not have the resources to do that; we are a voluntary organisation. Many people have the skills but lack the confidence to do it, and do not like putting their heads over the parapet.

You also have to take account of the fact that there is inadequacy in government complaint mechanisms. They all look as if they are designed—possibly they are just poorly designed—in such a manner that it is extremely difficult to break through and get responses. I am currently raising with the Australian Bureau of Statistics the complete failure of one of their complaints email addresses to actually function. It has not worked in the 22 months it has existed. It is typical of the problems here. As I have explained, the AEC's website is a nightmare to try to navigate to find information. Again, people do not complain for many different reasons.

A second aspect was that Canada was argued as being a positive example that supports the AEC's case. I have not had time to research the Canadian system, but in the AEC's own evidence they admit that the system there is consent based, not imposed. The Canadian approach is completely different from what the AEC are proposing in this bill. Again, it is a piece of illogic in their submission. The third point that we noted as a weakness in their argument was that they refer positively to use of data for a range of purposes in countries that have a civil or national register. Australia has rejected civil or national registers successively since 1987 every time they have been brought forward. In what way does that support the argument?

What we would have expected to see, and what we believe you should be requiring of agencies that come before you with proposals like this, is an analysis of the sources. We have not seen provision of information about the experiences to date with matching. Where is the data? Where are the appendices that show the extent of false positives that have been detected, the number of occasions on which they have drawn an inference, gone out and checked it and found it to be wrong?

There has been no evidence presented to you and no evidence presented to the public. Every match with every source of data has its own vagaries because the purposes vary so much, and because of that there should be multiple, separate appendices—one for each of the different matches—because they will provide quite different experiences: different kinds of inaccuracies and different results. Not a skerrick of that evidence, if it exists, has been presented.

I now turn to page 7, and I will now try to pick the key points out of this page. All of them are important points, but I should not dwell too long in evidence. I have mentioned the data-scrubbing issue; I must now mention the problem with the lack of controls. The AEC tried to claim that it is all very well because it is subject to controls. But it is not subject to controls. There is only one set of controls that exists over data matching, and that is for the parallel data matching program which was operated originally by DSS—and subsequently by Centrelink—under a very specific act, back around 1990 or 1992.

What exist apart from that are mere guidelines, which were established in the early 1990s. They have been revised minimally, and they have never been revisited. Since the early 1990s a huge amount has been learnt about how controls should be exercised over data-matching programs. Privacy impact assessments of various kinds and developments in techniques—even the creation of guidelines on PIAs—has taught us a great deal about how these controls should be done. That has never been reflected even in the guidelines, and those guidelines are completely unenforceable. The AEC can ignore them in just the same way that the AEC could have been told to do things by the Privacy Commissioner—though apparently it was not—and the AEC could ignore them because the Privacy Commissioner has no power. So the suggestion that there are controls is completely incorrect.

I also explain here, in the middle of the page, why the second letter improves the quality—or recovers some of the lack of quality—almost not at all because the same sorts of problems that beset the first letter will beset the second letter. The second letter confirming that we, the AEC, have gone ahead and done it does not really help matters at all.

The summary is that a great many errors are going to arise in this to the serious detriment of the quality of the rolls and, indeed, to the consternation of some electors.

ACTING CHAIR: We are going to have to move to questions fairly quickly.

Dr Clarke : I will now move just quickly through the rest of that. Page 7 is conclusive in relation of the bill. I do want to draw to your attention to the procedural matters on page 8. Firstly, it would be an expectation in circumstances like these that a risk assessment would have been performed by the agency. Risk is firstly looked at from the viewpoint of the Commonwealth, the electoral roll and the quality of the AEC itself and the quality of the rolls. Of course, there should also be a risk analysis performed from the perspective of the people whose data is being played with. There is no evidence of such risk assessment being performed before you or before the public. The department of finance and the ANAO expect those sorts of things and so should the parliament.

Secondly, it appears that no privacy impact assessment has been performed. That is normal business practice—it is an expectation of the Privacy Commissioner, and it is an expectation of the government. But all the AEC said in evidence was, 'We have been in consultation with the Privacy Commissioner over this particular bill.' We would like to suggest to the committee that you ask the Privacy Commissioner why he did not raise with the AEC the issues identified in this submission and why he failed to prevail upon the AEC to perform a PIA on a proposal, particularly when it clearly raises significant privacy issues.

The third procedural point is that no consultations appear to have been undertaken with NGOs at all. The APF would obviously suggest that it is an appropriate body, but, if the APF is disliked for some reason, there are other organisations which represent the public interest. It appears that the AEC, rather than having informed discussions in advance with appropriate organisations, would prefer the consultation to be by duel before successive hearings of this committee.

Fourthly, it would have been very easy for the AEC to consolidate the legislation. It is very likely that the AEC has already consolidated the amendments into the legislation for its own purposes in order to check that it is asking for the right things. Why has that not been provided to the committee? Why has it not been provided to the public to make it easier for us to analyse the effects?

The conclusions paragraph draws the obvious conclusions that we recommend should be implemented in order to overcome the trust impediments that exist and in order to send the AEC back to the drawing boards in relation to the proposal in this bill.

Senator RHIANNON: Thank you for the information you have put before us. It is very interesting. Would you explain the status of the second submission. Is it from you or from the foundation?

Dr Clarke : No, it was reviewed by the foundation overnight—

Senator RHIANNON: So, in essence, your first preference is to withdraw the bill and the second, if it is not withdrawn, is to make those changes?

Dr Clarke : Correct.

Mrs BRONWYN BISHOP: Listening to your very thorough evidence this morning highlighted just how bad was the research in the submissions we have received in this attempt to ramrod this through. It just backs up my belief that this is being done for political purposes. If this bill were to be amended and incorporate many of the things that you have suggested, is it possible to have such a bill and an activity like this without destroying the integrity of the roll?

Dr Clarke : No, it is not possible. Even if there were rollback and even if a few of the existing authorisations for disclosure were rolled back, that would not be enough. Public trust is undermined and it is going to be further undermined by this. So, unless there is a wholesale change in attitude and change in authorisations, public trust has been lost and will remain lost.

Mrs BRONWYN BISHOP: You talked about something I explored at the last hearing. You point out that many people have multiple addresses. They have an address for one purpose and a different address for a different purpose. For the commissioner to be making an arbitrary decision about which is the valid one for the purpose of the act and then writing to the new address, not the old address—not even sending to both addresses but just to the new address—is, to me, creating a built-in failure. Would you comment on that.

Dr Clarke : It is a putative new address; it has not been established as being a new address. Even if the match is correct, it may not be a) the address that the elector wants such letters to be sent to nor b) the appropriate address.

Mrs BRONWYN BISHOP: Supposing we had someone who wanted to vote fraudulently or a large number of people who wanted to vote collectively fraudulently, if they became aware of the fact—and there are ways they could become aware—that an address had been changed and therefore somebody on the roll had been removed from an address, they could actually register that other property as their address and vote, showing that as their address, could they not?

Dr Clarke : Under this arrangement, the ex-flatmates have the letter and the Electoral Commissioner does it for them, so it is not exactly difficult for that kind of fraud to occur.

Mrs BRONWYN BISHOP: So you could literally stack people into that address and they could all have a valid vote.

Dr Clarke : It would take some hard work, but the fact is that, yes, as you describe, the commissioner has set up that opportunity.

ACTING CHAIR: The senator has one more question before she goes.

Senator RHIANNON: Thank you, Mrs Bishop. Dr Clarke, I was interested in the context of your opinion on what has happened in New South Wales and Victoria. I may be wrong, but I understood that what is proposed federally was fairly similar to what has changed with how the rolls are managed in New South Wales and Victoria. What is your attitude to those changes that have occurred in the states?

Mrs BRONWYN BISHOP: Before you answer, can I just add something. On the Notice Paper is a new bill from the minister introducing automatic enrolment. I described the bill we are examining now as a precursor to introducing automatic enrolment. That was denied, but the bill is on the Notice Paper this morning. So the New South Wales and Victorian system is proposed to be rolled out here very quickly. I think we might look at that bill as well.

Dr Clarke : Yes, I suspect that might be the case. I have to be careful in going beyond the knowledge that I have in the room at the moment because there are others on the board who know significantly more about the New South Wales and Victorian positions than I do. But my understanding is that, yes, New South Wales and Victoria alone have gone ahead with a similar system to what is being proposed in this bill. I would have to check, but I do not have a recollection of us having the opportunity before the New South Wales parliament and the Victorian parliament to undertake analysis and to voice a submission on behalf of the public on those. Clearly, to the extent that it corresponds with this bill, we would oppose it for exactly the same reasons. Our comments on the justification would have to reflect whatever justification was put forward by the New South Wales Electoral Commission. Our opposition is not purely to the AEC doing this; our position would be that any government agency that sets out to impose upon rather using consent based change of address is moving beyond what we have expected of government administration into a paternalistic and/or authoritarian stance.

Mr GRIFFIN: Other than the philosophical objection, you are not aware of any problems that been brought to your attention with respect to the operation of those two systems, which do constitute well over half of Australia's electorate system?

Dr Clarke : No, we are not a complaints-handling body.

Mr GRIFFIN: Of course, but a number of the issues you have raised in your submission relate to concerns you have about the likely operation of the system at a federal level and a broad concern about the nature of these sorts of changes. Given that there is a situation where our two largest state jurisdictions have gone down this track in broad terms, I am wondering whether you have received any information with respect to the operation of it. As you are considering the policy questions, I would have thought in that process you would be aware of any concerns that may have been raised.

Mrs BRONWYN BISHOP: With respect, we may give you the opportunity to do just that. I am aware of only one report on what has happened in New South Wales and that was written by the ABC's commentator in an article that he wrote, which said, if my memory serves me correctly, that of all the people who were compulsorily enrolled only 12 per cent elected to be entered onto the federal roll.

Senator RHIANNON: The legislation was certainly thoroughly debated in the New South Wales parliament. I happened to be there at the time. It was signed off by all the parties. Dr Clark, just so I do understand, would it be correct to summarise what you said as, if the New South Wales and Victorian legislation is largely similar to what is proposed before us, the Privacy Foundation would be opposed to it?

Dr Clarke : Yes, that is correct.

Mrs BRONWYN BISHOP: Has anyone got that Antony Green article? I had it in my other papers. I do not have it here.

Dr Clarke : Sorry, I am not aware of the information that the members of the committee have drawn to my attention.

Mrs BRONWYN BISHOP: I think the important point to make is that, just as there has been no preparation by the federal Electoral Commission to justify it, as you very adequately pointed out this morning. I do not think there was any done in New South Wales either. There has been no really thorough investigation of its impact other than this single article, which is the only one that I have been able to find to date. I find that of concern in itself. From the transcript you would see that I find the idea of corrupting the integrity of the roll in this way absolutely abhorrent. Obviously I am in agreement with your summation that we are moving to an authoritarian or paternalistic situation. I thoroughly believe the individual has an obligation personally to put details into the roll to maintain its integrity. I am flagging this. There is no reference to this committee presently, but should there be a reference to this committee for the next bill I would be very grateful if you and the foundation—and you have indicated that other people on the board have additional information—might have a bit more time and would be able to give us a further submission, which I think is highly valuable to all of us. I think the material you have raised today, in terms of the depth of the research, has been totally missing from any of the debate to date, so I am very grateful for your presentation and would like to flag the possibility that there may be another opportunity.

Dr Clarke : Thank you, Mrs Bishop. We would endeavour to do so.

Mr GRIFFIN: Dr Clarke, I might just try to make a couple of points or ask a couple of questions going through your submission, so I apologise if it is a bit rambling. There is a general point in relation to the broader issues you raise—which go, if you like, beyond the legislation or relate to the legislation but also go beyond it. I understand about mandatory voting opposition. On point (2), the issue of addresses, I just want to make sure I am clear about that. The point I seem to be getting from that is that you are saying there that, because of the nature of living arrangements in modern society, address in itself is actually problematical with respect to being used as a basis for enrolment. Could I get you to follow that through with me a little bit. The argument around integrity which Mrs Bishop has referred to around the question of the potential for corruption of the roll is an argument based on people utilising the capacity to become enrolled at another location as a way to influence electoral results. That is a much wider question. It is something that we disagree on in terms of the likelihood of it happening and the practicality of it. But the point you seem to be making here is that address in itself probably is not really a way to go with enrolment. Could you comment on that. Have I got it wrong? That is not out of the question.

Dr Clarke : No, I have been insufficiently clear in section (2). The first point that is made in here is that address is difficult, in the case of human beings, as a basis for an electoral system. Let us consider an alternative approach that could have been adopted by the parliament for the purposes of assigning individuals their one vote. Suppose the parliament said that an elector may submit a request to have their vote in a particular electorate but must provide evidence that satisfies the commissioner, and that the default is the place where they most commonly live. If that kind of approach were taken, it would then roll up a whole pile of the difficulties together and solve a lot of them. It would not solve all of them, but it would enable the commissioner to say, for instance, that the grey nomads have said, 'I spent most of my life not in the last place that I was living but, in fact in this particular area. I'm going to be in Northern Australia almost all the time, but that's where I really have my affinity.' The electoral commissioner can see that there is sufficient evidence to justify you being allocated there. It is one person, one vote in an appropriate place. It is an alternative approach to the way the Electoral Act works at the moment.

Mr GRIFFIN: You understand that one of the problems you have—and, I argue, one of the problems this bill seeks to address to a degree—is that, because of that address entitlement and the fact that people do move around quite a lot, the question of actually being able to exercise that one vote is to a degree jeopardised by the nature of the operation of the system as it currently stands, combined with the individual circumstances. As you know—I can in effect point to it later on—one of the arguments around the question around other data sources is that you can have some inconsistencies. So in broad philosophical terms there is the issue of where you put the balance point around the question of ensuring that people have the opportunity to cast a vote versus the question of the nature of what criteria determine that right.

Dr Clarke : I was just suggesting that there would be a far smaller percentage of people who would be denied their vote, because they would remain stable in their preferred electorate for the purposes of voting despite the fact that they were highly unstable in their physical location.

Mr GRIFFIN: Yes, and I guess the question there would be: what would you require as the evidence to prove that linkage? For example, would it be sufficient if someone had been enrolled at a location and therefore that was seen as their showing that that was where they saw themselves as being linked, and then they subsequently move around? What level of activity would be required from the elector to ensure they maintain the franchise?

Dr Clarke : Reasonable levels that provide some support for the proposition. I agree that previous enrolment would be a very easy one because nothing additional would need to be provided, but where a person had, for example, said, 'I am really sorry that I left my home town; it was a country town, and that's really where my heart is, but I haven't lived there since I was 17 and I can't show an entry on the electoral rolls because I didn't have one,' there might be some need for some kind of additional evidence to be attached to the application form. But I believe that for many people to whom this matters, where they say, 'I actively want to take advantage of this,' they would not have a problem with coming up with some form of evidence that showed it.

Mr GRIFFIN: Sure. On the third point, the untrustworthiness of the electoral system, there are a number of points that are made in there. There is a key point I want to focus on which again, while not part of the legislation, is, I think, a factor that is worth considering. The argument there seems to be that there are currently too many uses for electoral data as it currently stands with respect to organisations being able to access them and that that in itself has an impact in relation to an individual's privacy. So in that situation you would be arguing that there ought to be a review of which organisations actually can access the data and in what circumstances and for what purpose.

Dr Clarke : Correct.

Mrs BRONWYN BISHOP: Do we have a list of who is entitled to access it somewhere?

Dr Clarke : The best one I have found is an appendix to the annual report of the AEC, but the one I came across was from 2008-09.

Mrs BRONWYN BISHOP: Do you think we could get a list of to whom the AEC makes its information available?

Dr Clarke : There is one further source. In the first submission we provided the URL for a publication of the Victorian Electoral Commissioner which lined up the Victorian situation versus the Commonwealth. It is from 2006, but I think we can say with confidence that everything that is on the list is still true, but there are doubtless a few more which you could probably find fairly quickly.

Mr GRIFFIN: With your point on silent electors, effectively what I am getting from that is that you believe that that is an underutilised option for people and that you are concerned about how it is possibly being interpreted with respect to what people need to satisfy in order to be able to be a silent elector.

Dr Clarke : Correct. Section 104 is too constraining, and the AEC are not in a position to apply it in ways that would be appropriate and use it as a trust-building mechanism. They are forced by the parliament to have a distrustful arrangement.

Mr GRIFFIN: I think that is an interesting point in terms of allowing people to avail themselves of that form of privacy protection. With the CRU—basically the data matching question—on the present bill I have a couple of questions. On page 6, at the top of the page, you make the point about the basis of data from births, deaths and marriages registries, with advising death as an example, and there is the point about the elector not being able to appeal on the basis that they have passed away et cetera. Surely provision of data in those circumstances would be something which would address concerns about fraud to a degree, because it should ensure that people are removed from the roll who, in fact, are no longer able to cast a vote. Would you have any problems with that?

Dr Clarke : I would have to check our previous submissions, because I think that our position was that we were actively supportive of the use of deaths registries, in a feed-forward method, passing information to the electoral office. That one is a good, positive balancing of interests. The dead do have privacy, and in particular the people associated with the dead have privacy. Privacy does not die with the individual, but that is a good balance socially for society. So I think we have actively supported it in the past.

Mr GRIFFIN: I may have misunderstood that.

Dr Clarke : I was not intending to oppose that; I was showing that that does not support the AEC's current argument.

Mr GRIFFIN: I guess the overall point of what I would argue is intended here is to utilise available data sources to ensure that people are enfranchised rather than disenfranchised and, in the process, provide opportunities to ensure that the roll is more accurate as a general sort of premise, whereas your conclusion would be that in fact the result would be the reverse or that, for a level of accuracy, you would get a greater degree of problems.

Dr Clarke : We start with the fundamental tenet that that is repurposing—an expropriation of data—and that there has to be strong justification for authorisation to be given for that to be done. The justification of the authorisation, we would suggest here, is insufficient. There are considerable inaccuracy problems that result in large numbers of false positives and false negatives and therefore the data quality issues and the quality of the inferences that are drawn is sufficiently low. We think that is another strong argument against doing it.

Mr GRIFFIN: I would argue in those circumstances—let us say that an elector is in a situation where, via a data matching problem, there is an error made and their enrolment is changed—then the question really is that if they then come forth and seek to exercise a vote, what actually occurs around the question of an objection process or a consideration process regarding where they are eligible to cast a vote and the question of what opportunity there is to correct the record at the time?

Dr Clarke : Many electors would find it galling that the AEC had taken it upon itself to make a change against their knowledge and when they go to vote they find they are disenfranchised in the electorate where they expect to vote. Now they are being told that they do have the franchise, but it is somewhere else—

Mr GRIFFIN: That may be the case. I guess that in the same way as there is not a great degree of evidence of how this may work in a positive sense there is not a great degree of evidence of how this may work in a negative sense. Your opinion is that many electors would find that is the situation. I guess we have to look at it and say whether we should find out whether that is the case.

Dr Clarke : The AEC had the opportunity to provide you with evidence and they did not do it, because they have done the data matching for some years now.

Mr GRIFFIN: I do not think that they found evidence of that.

Mrs BRONWYN BISHOP: They did not even look for evidence. I find it very galling when the AEC come along and try to justify the basis for doing this as because the ATO has used various forms to do something or other to make sure people are paying tax it legitimises their argument for automatic enrolment relying on data sources that are there for another source. There is evidence that there are many more people with Medicare cards than there are Australians who are entitled to have them, and it is equally so with tax file numbers and so on.

With regard to the AEC's sharing of information, do we know at all whether it sells any of the information? Do we know what the basis is? Does it have the power to sell information to other agencies? I was quite taken with part of your first submission, where it says that the foundation believes:

… that the pressure for increased accuracy comes more from the secondary users than from the requirements of electoral administration.

If the AEC is selling information or supplying it for whatever reason that I am not privy to, there can be quid pro quos and the secondary users can say, 'We want it to be more accurate. What can you do to make it more accurate for us?' I find that the accountability of the AEC, despite the existence of this committee, really has a lot to be desired.

Dr Clarke : I have the expectation that there is no sale. I am not aware of it being authorised and I am not aware of it being done. But, certainly, the quid pro quo argument is a reasonable suspicion because the AEC is unfettered in so many of its discretions. The parliament has gifted it with the power to make an awful lot of decisions that should properly still be in the hands of the parliament. So that kind of horse trading is conceivable.

Mr GRIFFIN: I am not sure that answer is representative in the context of all this stuff, Dr Clarke.

Dr Clarke : It does at least slow things down a bit!

Mrs BRONWYN BISHOP: If I can go to another point, we have given the AEC a lot of powers, which they never use. There has not been one prosecution for failure to change your address. There has not been one prosecution for multiple voting. There has not been any use of their powers to investigate electoral returns, and that was most obvious with regard to the Craig Thomson affair and the HSU where they made no use of the powers that the parliament had given them to call the people to account, to ask questions. It all seems like it is all too much work.

Dr Clarke : I think there is a certain amount of split personality here, in that there is still the remnant perception that it is a right to vote rather than an obligation to vote. There is still the notion that this is something positive that we should encourage people to do. Clearly we would be very supportive of that. We think that is the appropriate perspective, so we do not go rushing to suggest that the way to improve quality is to put more impositions of offences and to start prosecuting. What we would suggest is a more positive approach altogether. We have not discussed the consent based approach here. We are directly supportive of consent based approaches.

Mrs BRONWYN BISHOP: Would you like to expand on that.

Dr Clarke : When a person changes address, if they think their way through the number of organisations they need to advise it is quite scary. In the private sector there is little we can do about that—we need to keep track of the ongoing relationships we have, with a telephone provider and so forth. But when it comes to government, quite a lot of people do have the reasonable expectation: 'In this day and age you'd think they could do this for me, wouldn't you—that I could either go to a central point or, alternatively, go to one agency and have them pass on to other agencies.' Doing that on a consent basis, with clear conditions as to how this is performed, the APF directly supports. And personally speaking, as an electronic business consultant, I not only support it, I have endeavoured to implement it. That, to me, is part of the positive agenda that goes with the right to vote perspective, with encouraging, advertising, and not prosecuting. That is the much more positive approach that we would strongly support.

People in Australia divide into many different groups. There are many people who are very positive about being on the electoral roll, being up to date and voting. There are people who a bit apathetic about it but do it because they know they should. Then there are people who just find it a pain and do not want to be there. And then, of course, there is small element who want to sit completely outside the system. Each of those groups needs to be, we would believe, encouraged rather than attacked, as is being done here.

Mrs BRONWYN BISHOP: So you would say that if someone was, for instance, changing their address for the purposes of their licence there could be a clause that said, 'I consent to this information being shared with the AEC'?

Dr Clarke : I would go even further. I think the word 'request' comes up really well in those situations: 'Would you please do me this favour.'

Mrs BRONWYN BISHOP: So 'I would request'.

Dr Clarke : Yes. There have been designs that have actually put those boxes in a few organisations' forms and online web forms, but the systems still have not made much progress.

Mrs BRONWYN BISHOP: Have you any examples of that that you could let us have?

Dr Clarke : I am not sure whether Australia.gov has implemented that. We wrote a specification for that I believe 10 years ago. I cannot point to one right now, no; I would have to go searching online to find one. But they are entirely implementable. The question is what the take-up rates would be. And, remember, this is a cost to any given agency to build an extra interface out to six other agencies or to one central point, which then has to be built and maintained by somebody like DOFA, some central agency which will then operate as a hub and pass that out and also maintain the software. There is a cost involved by somebody, and if there is no advantage to the organisation it does not particularly want to do it.

ACTING CHAIR: There is a cost in automatic enrolment too.

Dr Clarke : Indeed.

Mrs BRONWYN BISHOP: Absolutely. I cannot find that article in the papers; I will have to go back to the office to find it. But it did show, I believe, that the actual number of people who were compulsorily enrolled had a dramatically less turn up and vote rate than did the rest of the people. We have a 92 per cent turn-up. There is so much room for so much more research to be done to see just where it impacts and how it impacts before you just get this bland stuff: 'This is a good idea. We support it.'

Dr Clarke : Yes.

Mrs BRONWYN BISHOP: I guess what really does offend me the most is that this is a political agenda which is backed up holus-bolus by the AEC without having done any work to legitimise it as a thoroughly researched proposal.

Dr Clarke : Something from a research perspective, putting that hat on for a moment, I would certainly like to see more of is population segmentation. The reasonable expectation is that the AEC would look hard at the different categories of people and the different motivators and pressure points and trigger points and ways in which those different segments can be encouraged. That kind of analysis I would have hoped would have carried through into submissions like this.

Mr GRIFFIN: I have to say I do think there has been quite a bit of work done with respect to that, particularly with issues with younger voters, first-time voters—how to encourage enrolment when they are 17 or 18. The issue as people get older becomes more difficult again. But quite a bit of work has also been done in looking at the question of older voters. It is not all necessary released but I think it informs elements of the conclusions that are reached about how to actually operate the system.

ACTING CHAIR: I am going to have to wind up the hearing. Does anyone want to ask a final question?

Mrs BRONWYN BISHOP: Can I refer to Antony Green's article on the ABC website, 'NSW Automatic Enrolment and its Challenge for the Commonwealth'. It is only nine months since the election—it is written in July.

Mr GRIFFIN: Can I suggest we get that circulated?

Mrs BRONWYN BISHOP: Yes, I think that would be excellent. But I was right:

So far only 12% of SmartRolled voters have taken steps to appear correctly on the commonwealth roll, and the rate of correction has been dropping since the NSW election.

This is by Antony Green.

Mr GRIFFIN: I would like to see the entire article and the context of all that.

Dr Clarke : Could I ask a copy that when it is distributed.

Mrs BRONWYN BISHOP: Yes.

Dr Clarke : Thank you very much.

ACTING CHAIR: Dr Clarke, do you wish to make any concluding remarks?

Dr Clarke : No, thank you, Chair, I have had every opportunity.

Mrs BRONWYN BISHOP: Should we find we have a reference of a second bill, can I place you on notice that we might seek a submission from you.

Dr Clarke : Yes, thank you, Mrs Bishop.

Mrs BRONWYN BISHOP: Thank you very, very much.

ACTING CHAIR: Thank you for your attendance here today. It has been a fascinating hearing. If you have been asked to provide additional material would you please forward it promptly to the secretariat. You will be sent a copy of the transcript of your evidence to which you can make corrections of fact. Thank you for appearing before us.

Resolved (on motion by Mrs Bishop):

That this committee authorises publication of the transcript of the evidence given before it at public hearing this day.

Committee adjourned at 10:47