Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
Economics References Committee
25/10/2016
2016 Census

McLENNAN, Mr Bill, Private capacity

Committee met at 09:40

CHAIR ( Senator Ketter ): I declare open this public meeting of the Senate Economics References Committee. The committee is hearing evidence on the committee's inquiry into the 2016 census. The Senate referred this inquiry to the committee on 31 August 2016 for report by 24 November 2016. I welcome you all here today.

The committee has received 91 submissions so far, which are available on the committee's website. This is a public hearing and a Hansard transcript of the proceedings is being made. Before the committee starts taking evidence I remind all witnesses that in giving evidence to the committee they are protected by parliamentary privilege. It is unlawful for anyone to threaten or disadvantage witness on account of evidence given to a committee, and such action may be treated by the Senate as a contempt. It is also a contempt to give false or misleading evidence to a committee.

The committee prefers all evidence to be given in public, but under the Senate resolutions witnesses have the right to request to be heard in private session. It is important that witnesses give the committee notice if they intend to give evidence in camera. If a witness objects to answering a question, the witness should state the ground upon which the objection is taken and the committee will determine whether it will insist on an answer, having regard to the ground which is claimed. If the committee determines to insist on an answer, a witness may request that the answer be given in camera. Such a request may, of course, be made at any other time.

On behalf of the committee I would like to thank all those who have made submissions and sent representatives here today for their cooperation in this inquiry.

I welcome Mr Bill McLennan. I invite you to make a brief opening statement, should you wish to do so, and then the committee will ask some questions.

Mr McLennan : I submitted a CV, which I understand has been circulated to all senators. I have had 40 years in statistics and I have worked overseas as well. The first point that I would like to suggest is that the contemporary advise that I put forward from the Parliamentary Counsel at the time of the Census and Statistics Act was being drafted leaves me strongly of the opinion that the Census and Statistics Act does not provide the Statistician with the power to direct people to provide their name in a census. I know there are other views around, and I also realise that the only way to settle this is to have a legal process. I am not proposing that. However, given that the revised act is not the Statistician's—it actually belongs to the Parliamentary Counsel—I would be surprised if his view did not hold in any law case.

As a matter of common sense, there are two separate exercises we are looking at here today. One is the taking of the census and the other is the matching processes afterwards. They are separate issues and should not be considered as one. In terms of taking the census, you do not need to have any compulsory collection of name. They have always been collected in the past on a voluntary basis and with high cooperation from the public. The criterion here to collect compulsorily is really to have a large number of names from which the matching process can be undertaken. It is pretty obvious that that is not covered by the Census and Statistics Act.

Second, I have grave doubts that what the ABS wants to do with matching is actually covered by its current legislation. What they are trying to do is to become the role of data integrator of various government statistical administrative records. I am not even sure that that function is compatible with the statistical function. They conflict slightly. Therefore I am not too sure of the legal situation of the matching that has been done to date.

Further, and more importantly, I do not think the general public has any real idea of what the ABS plans for data matching are, and that the ABS has no idea what the impact of such a change would have on the vital cooperation that the ABS needs from the community if it is to continue to do its job effectively. If people do not want to answer questions, they do not, or they answer incorrectly or do something that is not helpful. You have to have cooperation.

It also means—and I think this is pretty evident from the internet chatter—that the matching processes will almost certainly mean that the secrecy provisions of the Census and Statistics Act will be breached.

Further, many of these submissions that I have here in this folder are from people talking about the need for a social contract between the ABS and the community with respect to the census. How that social contract was perceived to exist was broken by the decision to compulsorily retain names in the 2016 census to undertake data matching without properly engaging the community in the decision.

I ask the committee to recommend that the ABS undertake a comprehensive privacy impact assessment of its data matching plans—I mean a comprehensive independent assessment and not the process that it went through in late 2015—before it proceeds with any data matching, whether using the 2016 census data or data from its many household surveys. I would also recommend that the ABS be required to report all data matching it undertakes in its annual report in some significant detail, so the parliament and the public can review and understand what is going on.

Third, many of the problems that the ABS got into in running the 2016 census—and I could add many of the problems that it has run into in other major statistical series, such as the now obviously broken monthly Labour Force Survey—is because it has walked away from the approach which in my day was called total survey design. Given the recent monthly Labour Force Survey kerfuffle, it is provident actually that I provide a brief description of what this means in the paper I submitted to the committee, entitled 'Is the monthly Labour Force Survey broken?' In brief, the total survey design is aimed at making sure there are no surprises when changes are implemented.

Fourthly and lastly, if the committee is tempted to consider that the census be undertaken less frequently, for example in order to save money, I would like here and now and with what follows to scuttle that idea. I would suggest they look carefully at the advice of the chief law officers to a High Court decision, McKinlay v The Commonwealth, in 1975, which led to the requirement for five-yearly censuses in the current statistics legislation. The High Court decision was that the Constitution required that the population of the various states be ascertained during the life of each ordinary parliament for the purposes of determining the number of members in each state in the House of Representatives.

Of course the requirement to conduct a population count in all states every three years, besides being very costly—say half the cost of running a census—would have almost been impossible, timewise, to implement. There was not really any practical solution to the High Court decision. In a fit of bravery I suggested to the powers that be that perhaps they should tell the High Court that the decision could not be implemented, but I did not get much success with that suggestion. A compromise was reached when the Attorney-General and the Solicitor-General said:

.... it necessarily follows that the states' respective populations be reliably determined. For this some method of counting the population such as a periodical census is essential.

The law officers' opinion led to the provisions in the Census and Statistics Act which required five-yearly censuses and quarterly population estimates. Unlike what some people think, it is not for the purpose of collecting the census: it is for collecting the right numbers to determine the number of seats in the House of Representatives. On the question of accuracy of population estimates in the context of the quinquennial censuses, the law officers said:

The temporal disparity between quinquennial censuses and triennial elections means that the statistical population estimates which are based on the antecedent census tend to become unreliable and thus to afford ground for a court to hold the number of each State's members in the Representatives is not in fact in proportion to its population.

That has never happened. The accuracy of the state population estimates rely in a measure on the accuracy of the census counts. It does make me wonder why it was suggested recently that the frequency of the population censuses might be altered.

CHAIR: I have a couple of questions, then I might hand over to Senator Xenophon. Firstly, have you had an opportunity to look at the ABS submission to this inquiry?

Mr McLennan : It is very long, and I did read it.

CHAIR: The ABS submission states that they have been compulsorily acquiring names since 1910 as part of the census. Are you familiar with that or do you dispute that?

Mr McLennan : I personally think that is wrong. I have checked with a colleague who knows more about censuses than I do, and in our knowledge it has always been collected on a voluntary basis.

CHAIR: They also say that they have sought advice from the Australian Solicitor-General on this issue, and that advice has confirmed that they have been acting within the law in relation to the census.

Mr McLennan : I think they talked to the Australian Government Solicitor, and AGS has an opinion. These days the AGS opinion is just like an opinion from any other lawyer. Back in the days when I was running most of this, we could only go to the Australian Government Solicitor and his opinion we had to follow. That was changed many years ago, when giving that advice effectively became privatised. It is just advice. You pay the government solicitor for advice just like you would pay somebody else for advice. The advices have to be considered on their merits, not because they came from the Australian Government Solicitor.

CHAIR: In relation to your experience as a former senior ABS officer, would you care to reflect on to what extent the operation of the ABS at the present time is affected by funding cuts? Do you see that is impacting on the quality of the work that the ABS is doing?

Mr McLennan : That is a difficult one. I have not been in the hot seat for a number of years. There is no doubt that there have been significant cuts to the ABS. My perspective on it is that it is the job of the managers in the ABS, headed by the Statistician and his senior staff, to work out which projects are the most important, if necessary, discuss with government, et cetera, and make sure that with whatever money they have the most important ones get done and the least important ones drop off; or you may not do particular large collections—you might cut them down in size and save money.

However, if you keep cutting, you have to think that there is going to be some impact somewhere. There is a lot of evidence of this in the press. The lady from The Australian Financial Review made some comments on this matter not so long ago. It is going to have some impact.

From where I sit and looking at it from the side, I often wonder why they are making certain decisions that they do make. I think a lot of that is to do with lack of experience, if you really want to know. They are not lacking numbers—or they are maybe lacking numbers too—but they are lacking numbers of people with experience. There has been a gross loss of experience out of the ABS and, I think, from my reading, it has happened right across the public service, not just the ABS in terms of loss of knowledge.

Senator XENOPHON: I just apologise to the committee: I am not physically there today, because I have a funeral to attend later on but I will be back on this afternoon with the committee. Mr McLennan, thank you very much for your submission and for giving evidence. I have some questions that go to the ABS's power to collect names and addresses. You have argued in your submission that the ABS does not have the legal authority to collect names and addresses because, as I understand it, you do not consider it to be statistical information as such. Is that in essence what you are saying?

Mr McLennan : What it says in the act is that, if you collect the information, you then have to compile statistics and publish them. The bureau is about collecting, compiling and publishing statistics—that is its job in a nutshell and it always has been. The act, as it is currently written, states that that has to happen, so you start collecting names, then say you yourself: how am I going to compile statistics about names? Do we want to know how money Bill McLennans we have got in Australia? Hopefully, no more than one, I can hear people say.

Obviously, under collecting names in the census, we are not doing anything in this last census to collate statistics and therefore we are not publishing them. The act says: the ABS shall collate and the ABS shall publish. So it falls down, so you cannot really say that there are any powers to compel people to do it. That is a layman's answer for you. I just think the argument is so straightforward.

Senator XENOPHON: Can I just go to that issue in terms of the collection of names: you are saying that actually providing your name on the census form is unnecessary for the work of the Australian Statistician.

Mr McLennan : No, the collecting of names on the form has always been done on a voluntary basis and then, as soon as the census is taken, within six months or so—don't hold me to the six months but it happens quickly—that information is destroyed. It does help to look to see if you have got names to work out the relationships within a household to make sure that they have been filled in correctly in terms of editing the returns. That is one simple example.

Senator XENOPHON: But you are saying that the legal basis for requiring names and for them to be kept—in other words, making it compulsory—could be counterproductive to the cooperation that the Australian people give to the census.

Mr McLennan : It is certainly does not help them, does it? I thought long and hard about whether I would put my name on the census form but I did in the end. I was rather exposed on this.

Senator XENOPHON: You are saying that this census will use names and addresses to create a statistical linkage key for a unique identifier. Do you have particular privacy concerns in respect of that?

Mr McLennan : I did not say anything about privacy in my introduction, because there was so much in the reports of people who know more about privacy than I do—they are bound with clients. It certainly does affect the privacy side, no doubt about that. Our experience in my day—and I started working on the censuses in 1960 after the end of my first year at university. I spent four months in Sydney doing nothing else but hand coding all the census returns for that year, including whether it be labour force, whatever. You soon learn that people are very helpful, but they do not like to have the government push over them too much.

Senator XENOPHON: If the provision of the name or something was voluntary, do you think that would be more likely to get a greater cooperation of those Australians who have a wariness about providing their name in the first place?

Mr McLennan : Let me put it this way. My contention—and I answered the question a moment ago—is that for all of the censuses that I know of, in the recent past anyway, names have not been collected compulsorily. They have been collected voluntarily and the censuses have worked well. That is the answer to your question.

Senator XENOPHON: You have talked about trust and goodwill. Do you link the issue of trust and goodwill towards the ABS to the quality of the data that is likely to be collected?

Mr McLennan : Yes, I do. No one likes to be told they have to do something—or maybe it is just me.

Senator XENOPHON: No.

Mr McLennan : Most Australians are not like that. They do not like it if you say, 'If you don't do this we're going to throw you in jail or fine you $180 a day,' or whatever. That gets people's backs up. When you are dealing with the Australian population you have to get them onside. You have to get them to cooperate. Whether they do it in a vindictive manner or not, they are more likely not to give you the information that you are looking for.

I was heavily involved in the Labour Force Surveys. There was one classic case when we were in a pilot test and one interviewer, from Tasmania I might add, by the way, wrote in to us and said, 'I went to all of these houses and they all told me the answer to the question was no, but I can assure you from their actions their real answer was yes.' If you do not handle people in the right way, you do get an impact on the information they give you. That is well known.

Senator XENOPHON: Whilst the intention of the ABS was to have a greater degree of accuracy in the census by requiring the provision of names, do you think that, paradoxically, an unintended consequence was that there was less corporation in respect of the census?

Mr McLennan : There could be. That has to be shown.

Senator XENOPHON: I have one final question, and if there is time later I am happy to come back, but I know my colleagues may have questions. In linking the data between the census, section 8 of the act seems to suggest that the census is a distinct activity taken every five years. Do you think that it actually confers a power to link the data across each census?

Mr McLennan : That is a good question. I am not absolutely sure of that. I think it depends on what you do with it. Linking is a difficult process.

Senator XENOPHON: It is also a significant process, is it not?

Mr McLennan : Sorry?

Senator XENOPHON: It is also significant in that the level of linkage of data, and the extent of it, goes beyond what has been done in any previous census. Is that a fair comment?

Mr McLennan : You are mixing up two concepts. There are two things going on: there is the taking of the census, and that has nothing to do with matching.

Senator XENOPHON: Yes, I know that.

Mr McLennan : After the census is over you have a matching process.

Senator XENOPHON: Yes.

Mr McLennan : In the past they have done some matching on an experimental basis, to work out how it works. That has been well reported in a methodological study that is on the ABS website. I looked for its address this morning to give to the committee, but for some reason or other my old fingers cannot get the internet to work properly today. But, what that report says, as you would expect, is that if you start this data matching and move it over time, as you get further and further away from the source material—just like with longitudinal surveys—the results get worse. The methodological study said that the business of moving censuses forward produces roughly the same accuracy as a longitudinal survey. In my view, the accuracy of longitudinal surveys is notoriously poor.

Senator XENOPHON: Thank you. I do not have any further questions at this stage.

Senator HUME: Mr McLennan, thank you very much for your submission. It is very detailed and well researched. I can see you have written it with an extraordinary amount of enthusiasm: there is an awful lot of inverted commas, bolds and statements like 'elephant in the room', 'underhand' and 'shifty', and all sorts of quite colourful language.

Mr McLennan : That is me, I am afraid.

Senator HUME: I speak colourfully too, so I understand that. I might have missed this—when did you retire from the ABS?

Mr McLennan : I have a note here, somewhere. It was so long ago, I forget. June 2000.

Senator HUME: That was 16 years ago. A couple of things that you have said today are that you have a friend who knows a lot more about the census than you do, there are people who know a lot more about privacy than you do and that the Australian government Solicitor-General's office has said that the ABS can collect name information on a compulsory basis but that you doubt that advice's credibility. What was the motivation for you in putting together such a comprehensive review and presenting it to this committee? What was it that made you do this?

Mr McLennan : The comprehensive review that I have submitted is just a summary of a number of papers that I have been writing and sharing with the Australian Bureau of Statistics. I deliberately did not rewrite anything. I just put the papers together and you can read a trail.

The start of the whole process was when I saw what they were doing with the legislation. I have had a lot to do with the ABS legislation. I am not a lawyer, but I think I know as much about the Census and Statistics Act as most lawyers do. I thought the ABS was making a bad mistake, and I did write to the Statistician saying that. I got a reasonably terse response which rubbed me the wrong way, so I have persisted with looking at it.

The other reason why I am writing these things is because I get rung up by people all the time. I have been rung twice this morning, looking for comments on the Labour Force Survey.

Senator HUME: This is an ongoing crusade for you?

Mr McLennan : No, it is not an ongoing crusade. I am known in the economics side and the statistics side in Australia, that if people want to ask me a question I will talk to them and be prepared to try to explain difficult statistical things in simple terms. So people come to me.

Senator HUME: Thank you. Those were my only questions.

Mr McLennan : I like talking about statistics, by the way—as you have probably gathered!

Senator GRIFF: Mr McLennan, you stated in your document:

…neither the Australian Privacy Foundation (APF), or any other non-government privacy agency, were consulted about the 2016 PIS.

Is this unusual, in your assessment?

Mr McLennan : I thought it was extremely unusual, but on the other side there was no legal obligation on the ABS to consult with people like the Australian Privacy Foundation. It just seemed to me, probably, to be a process mistake on the part of the ABS. When you are going through something like this, trying to make a big change, you want to get people onside and you have to talk to them all. But you cannot talk to them all, so you talk to representative groups. That is what you do.

The Australian Privacy Foundation has a lot of contacts with a lot of people who know a lot about privacy and other things. If you do not consult them you not only get the Privacy Foundation offside you also get all the other people offside. Therefore, I think it was probably a poor decision on the part of the ABS. I do not think it was done with any vindictiveness—not vindictiveness—it was not done for any reason, that they are not going to not talk to Australian privacy people because they are all privacy nuts or whatever. I think it was probably just a mistake.

Senator GRIFF: There was that damning privacy impact report by privacy expert Nigel Waters relating to the changes in the 2006 census. You also mention that privacy concerns were raised in 2011.

Mr McLennan : Yes.

Senator GRIFF: Do you consider that there are any factors differing between the 2006, 2011 and 2016 censuses that would get a different outcome in that analysis?

Mr McLennan : Short answer: no. In one of my papers I actually referred to that matter, though maybe not the one in front of you. I said that, if I were querying the ABS on this, I would say: 'Well, this is what happened in 2006 and this is what happened in 2011. What's changed to make it okay now?' The short answer is: not much.

Senator GRIFF: One thing that has changed, going back to the actual act, is that the regulations changed prior to the current census, where 'name' is now included in the regulations.

Mr McLennan : I think 'name' could have been included in the regulation before. All that says is that you can ask for name. It does not make it compulsory.

CHAIR: Senator Xenophon, do you have any further questions?

Senator XENOPHON: Just a couple. Just going back to the chronology of this—and I ask this of you as a former chief statistician, although I note that you retired 15 or 16 years ago from the position—in terms of keeping in touch with the public, a lot has changed since 2000, but the time line of events suggests that there were problems earlier in the day. There were a number of problems, and I think the problems started with an outage at about 10.08 am on 9 August, but it was not until 11 pm that there was a public message from the ABS—I do not think it was before that—to advise that the form would be out of action for the rest of the night. Do you have any views as to how you think, as a former chief statistician, the bureau should communicate with the public to ensure confidence in the process and in the census itself?

Mr McLennan : I was not involved at all in this exercise, but I have been involved very much in public relations and that sort of exercise for censuses. As you well know, the public relations exercise in respect of this census was, shall we say, not up to par. Anyone who saw the skit that was put on television would, as I did, cringe when they saw it. You really cannot collect a big census properly unless you spend some money—a significant amount of money—on PR to get your message across.

As regards the processing itself when it went belly up, I think ideally the ABS probably should have reacted a bit quicker, but it is the first time collecting the information electronically, subject possibly to people trying to breach your system, was ever undertaken, so maybe the bureau were slow to react—I think that is probably what the words are. It probably could have been caution; they were not fully sure what was going on. I am making all that up, because I was not involved at all and I do not know the people who were involved, so I cannot even comment from that angle.

Senator XENOPHON: Thank you. My final question is a broad one relating to those people that completed their census form accurately but did not provide their name. Do you have a view as to whether that compromises the essential integrity of the information collected, and do you have a view as to whether those people ought to be tracked down and prosecuted and whether that would be a fair thing to do?

Mr McLennan : That is a good question. I will have to fess up that, when I was statistician, we did prosecute a few people.

Senator XENOPHON: But that was for not completing the form at all.

Mr McLennan : Yes, that was for not completing the forms at all. But we only did it in extreme circumstances. I was going to say early on, and I did not get around to it, that, given that I think there is a significant doubt whether it is legal or not—in fact, I am sure is not legal to collect the name compulsorily, and I have gone to print for that many, many times—under those circumstances I think it would be a gross injustice to have people prosecuted and fined. That is a personal opinion.

Senator XENOPHON: If you accurately complete the census but do not put your name—you put everything else in there that is required of you—does that compromise the census?

Mr McLennan : Well, I do not think it does. But you are asking me the question whether they could prosecute, and I think under the law they could.

Senator XENOPHON: Yes, but in terms of the integrity of the census, not providing the name is itself a problem.

Mr McLennan : The integrity would be breached if the people filling out the census form were miffed about having to give their names, so they did not give their names and then they botched a few questions deliberately. That could easily happen and nobody would know, and that would affect quality. But I do not know that has happened. I could not give an estimate as to if it did or did not.

Senator XENOPHON: Yes.

Mr McLennan : Even the post-numeration survey will not pick that up because they only check on coverage—you know, are any houses missed, or are houses counted twice? They do not go in to see, 'well, let us have a look at your questionnaire: what was your answer to the census and did you give us the right answers?' You would need a survey like that, Senator, to really give you the correct answer to your question.

Senator XENOPHON: Thank you, Mr McLennan.

Senator HUME: I have just one more question: surely people intentionally erroneously filling out a census form has nothing to do with whether they had to compulsorily put their names on it or not. In previous years, we have had however many people claiming to be Jedi Knights—I do not think that was done because they had to put their name down compulsorily. It was not because they were 'miffed,' in your words.

Mr McLennan : I read that the other day—that there are more Jedi Knights than Jewish people in Australia, or close to the same number.

Senator HUME: That is exactly right. So do you not think that it is a rather long bow to draw to suggest that because people had to put their names down compulsorily, they potentially would have filled out the forms erroneously intentionally?

Mr McLennan : I do not know that it is a long bow. If I got put out—

Senator HUME: You did not put yourself down as a Jedi Knight though, surely.

Mr McLennan : No; I put myself down as Bill McLennan, hoping they would misspell it or something! But I did not give my date of birth; I gave the age at last birthday because that is nowhere near as intrusive.

Senator HUME: But that has been done in previous years, and it has nothing to do with whether you had to put your name down compulsorily.

Mr McLennan : No. But because it has never been compulsory before, I cannot give you an example in the past where it was compulsory, and what happened—because it has never happened before.

CHAIR: Thank you very much, Mr McLennan. I do not think there are any further questions, so thank you very much for appearing before us today.

Mr McLennan : Thank you.