Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
 Download PDFDownload PDF   View Parlview VideoWatch ParlView Video

Previous Fragment    Next Fragment
Economics Legislation Committee
Australian Bureau of Statistics

Australian Bureau of Statistics


CHAIR: Gentlemen, welcome to the Senate Economics Committee. Do you have an opening statement for us, Mr Kalisch?

Mr Kalisch : No, I do not, Madam Chair.

CHAIR: Senator Ketter will kick off the questions.

Senator KETTER: Welcome to estimates, Mr Kalisch. I want to start off with the labour force survey. I want to ask about the May figures. The ABS jobs figures for May were significantly higher than expectations. A number of commentators and market participants questioned the accuracy. Do you consider those figures to be accurate?

Mr Kalisch : The labour force survey provides our best estimate of labour force features and trends at the time. The ABS regularly provides commentary to a number of users to recommend that they use the trend estimates, rather than the seasonally adjusted estimates, on a month-by-month basis. The ABS has been producing trend estimates since 1987, because even in the 1980s it was also apparent that there was some month-to-month volatility in the numbers. That volatility is due to a number of sources, so we do recommend that people use the trend figures. They are the best estimates and really reflect the nature of Australian labour force estimates. They are amongst the world's best in terms of labour force survey estimates. We do stand by those numbers.

Senator KETTER: But do you accept that they are more volatile than what they would normally be, and if so why is that the case?

Mr Kalisch : The volatility comes from a number of different dimensions. One is the sample size. We do survey about 26,000 households, which is about 50,000 people across the population, each month. We do maintain people in the survey for eight consecutive months, so we do rotate out one eighth of the sample each month. We also find that even with going back to the same households there are differences in terms of people responding to the survey from month to month within the household. That does drive a little bit of volatility. But looking at the volatility over time, the work we have done in our Methodology Division, and we reported this about 12 months ago and we have been updating that work as well, has shown that at a national level the volatility in the month-to-month labour force series is probably less now than it was in the 1980s and 1990s, and particularly around the GFC—the GFC was particularly fraught—and potentially around major labour market turning points. So at a national level we are not seeing any greater volatility; in fact, if anything, there is slightly less.

Senator JACINTA COLLINS: A report in The Weekend Australian on the 13th of this month indicated that education minister Simon Birmingham had referred a detailed report prepared by the Victorian Catholics to his department for assessment and asked the Australian Bureau of Statistics to check the claims of systemic bias. Are you familiar with that report?

Mr Kalisch : Yes, we are.

Senator JACINTA COLLINS: Did you receive a copy of that report?

Mr Kalisch : We received a copy of the report, yes.

Senator JACINTA COLLINS: Did you receive any other material in that request?

Mr Kalisch : The request was very much along the lines of I had a phone call from the secretary of the education department—

Senator JACINTA COLLINS: Michele Bruniges?

Mr Kalisch : Michele Bruniges, saying that her minister would like us to comment on a paper—we understand it was written by the Victorian Catholic Education Commission—and we provided some advice back to the department of education around some of the methodological issues that they had raised.

Senator JACINTA COLLINS: So you produced a discussion paper, or a comment paper?

Mr Kalisch : I would not call it anything quite so elaborate. We provided an email response and had spoken to some of the officers in the education department within 24 to 48 hours of that request.

Senator JACINTA COLLINS: What date was that request?

Mr Kalisch : I would have to take that on notice unless anyone has the information here.

Senator JACINTA COLLINS: Perhaps we can continue asking questions and if that information becomes available as we go you could assist us. So within a couple of days of the request you replied. Can you make available the information in that reply?

Mr Kalisch : The reply I suppose noted some issues around the commentary that was provided by the Catholic Education Commission of Victoria. It talked about the potential use of small area data and whether there would be some benefit in using personal level data.

Senator JACINTA COLLINS: Does small area data equate to the small areas that the Gonski report was referring to? I think in those days they were mesh blocks.

Mr Kalisch : It was not necessarily mesh blocks, but how accurately it captures the nature of school populations depends on the nature of the area they are capturing, and potentially there are some differences because people are coming from out of area or there are changes in terms of I suppose the composition of the school population compared to the local area. That advice was provided on 9 May.

Senator JACINTA COLLINS: In general terms, basically, you said that what the Catholic Education Commission of Victoria was describing was accurate.

Mr Kalisch : There were a couple of aspects where we said there were some—

Senator IAN MACDONALD: Was this in the way of advice to government that is being requested? It is very interesting, I might say, but is this a matter that we should be dealing with at estimates?

CHAIR: It is certainly on the cusp as advice to government. Perhaps we could be a little bit more careful with the questions and answers.

Senator JACINTA COLLINS: I have not heard a public interest immunity claim yet.

CHAIR: Let's just tread lightly, but we can continue with the line of questions.

Senator JACINTA COLLINS: I understand that Senator MacDonald is advising the officer to tread lightly, and I am sure the officer will take his advice on board. But equally, it is a balance that he needs to deal with because Mr Kalisch heads an agency that is an independent agency as well, so he has a couple of issues to grapple with there.

Senator GALLAGHER: He is quite experienced at estimates.

Senator IAN MACDONALD: It is a question of whether it is advice to government. As I say, I am finding it very interesting, but I think asking public servants to disclose advice that they have given to the government is, perhaps, something that the minister might consider—

Senator Sinodinos: Let's keep going on the basis that there would be no disclosure of advice to government in the sense of policy advice. We are not talking about policy advice. You are asking about a request for certain information?

Senator JACINTA COLLINS: I am asking for the ABS's views with respect to a paper produced by the Catholic Education Commission of Victoria, and Mr Kalisch was providing the answer.

Senator IAN MACDONALD: Perhaps the way around it is, 'Have you seen the paper,' and, 'Yes,' and, 'What do you think about the paper?' Of course, that is the question of opinion. It might also not be relevant.

CHAIR: I do not think there is a point of order at this point.

Senator IAN MACDONALD: If nobody else is worried about it, then why should I be?

Mr Kalisch : In effect, all I am saying is that there are a few perspectives around some analytical techniques that could or could not be used by the education department. Perhaps to further assuage you on the—

Senator JACINTA COLLINS: You suggested new techniques?

Mr Kalisch : No, these are just some aspects that were raised by the Catholic Education Commission, and we were commenting on the feasibility, or otherwise, of their potential use.

Senator JACINTA COLLINS: And what was your assessment of that feasibility?

Mr Kalisch : We said that there were a couple of things that they could look at if they chose to. We were not going to provide that advice to the education department. Ultimately, it is up to their choice as to what techniques or approaches they use in their analysis. It is their call. They understand the policy. We were just talking to them about what might or might not be possible.

Senator JACINTA COLLINS: Ahead of this request on 9 May, had the ABS had any previous dialogue with the department of education about similar matters since the Gonski review in 2011 recommended that these matters should be comprehensively reviewed?

Mr Kalisch : I am not aware of anything, but I have only been Statistician since the end of 2014.

Senator JACINTA COLLINS: Could you take that on notice?

Mr Kalisch : I can take that on notice.

Senator JACINTA COLLINS: If you had a look at the paper, obviously, you are aware of the general problems that are being raised here.

Mr Kalisch : We are aware of some of the suggestions from the Catholic Education Commission. They probably fall into two categories. There is a set of dimensions that you could, potentially, look at in terms of other analysis or more refined analysis. Ultimately, it is uncertain as to whether that would make any difference to the outcomes, but there is potential for further analysis. There is a second set where there really is no information source that could be used to address the concerns of the Catholic Education Commission. Just to give you one example, they suggested greater use of wealth data, whereas there really is no comprehensive, frequently collected source of data for local areas.

Senator JACINTA COLLINS: In relation to wealth?

Mr Kalisch : In relation to wealth; we have enough difficulty doing wealth estimates on a population or even a state basis, let alone going down to small areas. I think the other dimension was that they suggested more frequent information at a local area, which would, in effect, mean that we would need to run a census every couple of years. There were a couple of things that were way off beam in terms of feasibility.

Senator JACINTA COLLINS: But there were some things that were way on beam?

Mr Kalisch : There were a few things that we suggested to the education department that they could consider.

Senator JACINTA COLLINS: So, more frequently. In this article I was referring to, Senator Birmingham said that 'socioeconomic data had been used in the school funding system since 2001, and had been refined, expanded and broadly accepted as a credible way to measure capacity to contribute'. But you have taken on notice what refinement has indeed occurred in recent times, because you are not aware of it?

Mr Kalisch : I think the aspect or the question you would really need to direct to the education department. They are the ones that are responsible for the analysis. We provide some advice when requested.

Senator JACINTA COLLINS: And the data.

Mr Kalisch : And the data. And they use our data.

Senator JACINTA COLLINS: Yes. Getting to the data issue, I have been attempting to gather some data through the Parliamentary Library, and I have been told by the Parliamentary Library that the census data that would assist my inquiries can only be obtained by purchasing customised data from the ABS itself. We have asked the ABS to appear with the department—on Monday, I think it is. We have an extraordinary time line for this inquiry; God only knows why, because the department themselves say spring is okay for the legislation—but anyway. In that context obviously it is impossible for us to commission the ABS to do work to get the type of data that might help inform the considerations about the new provisions in this bill. So what I am hoping to do today is explore with you what data the ABS might be able to provide to a Senate legislation committee and/or indeed the Senate which would assist our considerations on this issue and how that might fit within your current customised data purchasing arrangements. Am I making sense to you?

Mr Kalisch : Yes. But as of right here and now I am not sure what your data request actually is.

Senator JACINTA COLLINS: Okay. I can go to the detail of that and you may be able to help with some of that. Some of the direct methods that Catholic Education Commission of Victoria was referring to I would like to see in relation to, for example, the income of families with primary-school-age children as opposed to secondary-school-age children. I would also like to see the data around workforce participation of families with primary-school-age children versus secondary-school-age children. So I am assuming the ABS—from the census—has collected data about the age of children. It might have to be households rather than families; I am not sure. But I am sure that there is data in the last census that would assist in understanding whether there are different capacities to contribute issues for families with primary-school-age children as opposed to secondary-school-age children.

Mr Kalisch : We would need to precisely look at these dimensions and see what was possible through a customised request, or alternatively through the TableBuilder products. We have TableBuilder, which does enable people to get more refined data and, I suppose, clearer data.

Senator IAN MACDONALD: At a cost, though?

Mr Kalisch : TableBuilder per se—we can provide access to the Senator's officer. They can essentially use it—

Senator JACINTA COLLINS: I am assuming the Parliamentary Library would be on top of that one.

Mr Kalisch : I think we will just have to see the nature of the request.

Senator JACINTA COLLINS: The question is more—

Mr Kalisch : The Parliamentary Library may well have the ability to use TableBuilder Pro, so they may well have a licence.

Senator JACINTA COLLINS: They have already told me that the data I seek is not available through their routine arrangements and that I would need to purchase customised data from the ABS.

Mr Kalisch : I would prefer us to make that judgement.

Senator JACINTA COLLINS: Okay. That is good. So how could we progress this between now and Monday?

Mr Kalisch : If you could provide me with your particular request then we can make sure on Monday that we can provide you with either an answer as to what is possible—

Senator Sinodinos: You would have to put it in writing.


Senator Sinodinos: Formalise it.


Mr Kalisch : We have taken down a few things that give us a bit of a hint, but if you could—

Senator JACINTA COLLINS: It gives you a head start. I might put it in writing and then try and allow sufficient time to allow whatever dialogue needs to occur between that to refine it any further if that is necessary. Then the purchasing arrangements become an interesting issue. How do you manage a request from a senator compared to a request from, say, a marketing company?

Senator Sinodinos: With more respect, I think.

Mr Kalisch : Yes.

Mr Kalisch : It is from a senator.

Senator JACINTA COLLINS: Yes, but Senator Macdonald raises the correct point. I have the Parliamentary Library saying to me, 'You can only get access to this if you purchase it'. And I am saying: 'That sounds a bit odd. I think I will visit that question at estimates.' So now I am trying to understand what we need to do to overcome that issue, because I would have thought—as an independent public agency—there is a significant public interest that information that arises from the census is available for our consideration of legislation with or without fee arrangements. If there has to be a fee between the ABS and the Department of the Senate at the end of the day, so be it, but how do we overcome that conundrum?

Mr Kalisch : Well, I do not think we have actually got to the first step yet of establishing that there needs to be a fee. So we will look at that aspect. We do have the opportunity to waive fees in exceptional circumstances.

Senator IAN MACDONALD: Perhaps on notice you could tell us just what your rules are around the areas that Senator Collins raises.

Mr Kalisch : Yes. We can certainly provide that to you on notice. The other dimension I would note as well is that some of that income is important for us to continue our business. We do get a certain amount of an appropriation from government that enables us to do a certain amount of work. If there is anything beyond that—and some of these customised requests, if it is in that nature, can be quite detailed. This one—

Senator IAN MACDONALD: Help the ABS out and pay.

Senator JACINTA COLLINS: Do the PBO make requests?

Mr Kalisch : We have an MOU with the PBO.

Senator JACINTA COLLINS: That might be the way around it—if we do the request via the PBO. On this occasion we will not bother for now because it will slow down things anyway.

Mr Kalisch : Yes.

Senator JACINTA COLLINS: But that may be the issue that overcomes the funding arrangements. Moving then to issues around ABS data integrity, can you outline your data privacy policy?

Mr Kalisch : Perhaps I can just talk about the privacy dimensions per se. We do have a number of dimensions to our privacy policy. We have the aspects around the legislative provisions in our two acts—particularly, the Census and Statistics Act, which talks about the secrecy of the data that we keep. We have various provisions in our legislation around the inability and exclusions of the ABS to provide that data to other agencies, including the security agencies or courts or tribunals or parliamentarians. We also have dimensions within our legislation that talk about the requirement that we can place on people to provide data to us—households and businesses. In our legislation we have aspects around producing statistics that do not disclose the identity of individuals and businesses. There are other privacy and security aspects around the way we maintain data, the way we have various controls on officers within the ABS, as to what access they have to data and what access they do not have. We also have some policies, particularly in our data integration area, which keep names and addresses separate from the content.

Senator JACINTA COLLINS: You are currently working through a process of integrating the census data with other datasets. Is that creating issues?

Mr Kalisch : At the ABS we have integrated census data with other datasets for over 10 years. It is not new. This is something that we have been doing as part of the Census Data Enhancement program in the 2006 census and the 2011 census. We are also going to have a Census Data Enhancement program in the 2016 census.

I will just give you a sense of some of those examples where we have used census data with other administrative data. We have used census data essentially to provide us with the ability to produce statistics about the outcomes of migrants. We can link it with the migrant settlement datasets and look up the employment and income outcomes of people who have come into Australia on different visa classes.

We also, in the 2006 and 2011 censuses, linked the census data for Indigenous people with the information that is available on death certificates. That has enabled us to provide more accurate estimates of mortality for Indigenous people and then ultimately more accurate estimates of life expectancy, which is particularly pertinent in terms of the COAG Closing the Gap target.

We have also used census data to look at some of the outcomes for people who have undertaken various mental health programs and initiatives and to look at the outcomes for them in terms of employment and other characteristics. We have also used census data, again, to look at outcomes for people who have undertaken various education programs.

Senator JACINTA COLLINS: These people who have potentially filled out the census have no idea that their data is being matched with some other datasets that might talk about their mental health condition or their employment circumstances?

Mr Kalisch : Senator, if I can take you back to your days in government when you were involved with community services or other programs, one of the big challenges that you would have had at that stage was understanding: what are the outcomes from these programs? Are they actually delivering—

Senator JACINTA COLLINS: No, Mr Kalisch, I understand the policy purpose for these things. What I am trying to understand, though, is how merging de-identified data with identified data is then to be managed from the privacy side. The very first issue, obviously, is that, when I, Joe Blow, member of the public, fill out my census form, I do not necessarily have any conception about what other data that information might be being merged with.

Mr Kalisch : I would be surprised to find too many people who did not understand about the privacy dimensions, particularly in the lead-up to the 2016 census. In the previous censuses the ABS, I understand, was also transparent in terms of producing documents that talked about the Census Data Enhancement program, and the—

Senator JACINTA COLLINS: But does the census documentation itself alert people to this purpose?

Mr Kalisch : Yes.

Senator JACINTA COLLINS: Okay. The enhancement process, you say, has been running for some time now. It is not new.

Mr Kalisch : No.

Senator JACINTA COLLINS: I was wondering what sort of ethics discussion had occurred around this process.

Mr Kalisch : The ABS does not have an ethics committee.

Senator JACINTA COLLINS: Should it?

Mr Kalisch : There are probably competing views on that.

Senator JACINTA COLLINS: What is your view?

Mr Kalisch : My view is that I do not see that there is a compelling reason for us to have an ethics committee at this stage. We do go through privacy impact assessments and various other things related to the everyday linkage program. We also have various approaches in the way that we do data integration that maintain the secrecy of personal information.

Senator JACINTA COLLINS: A privacy impact statement?

Mr Kalisch : An impact assessment, yes.

Senator JACINTA COLLINS: Sorry, assessment. For example, can you guarantee that every ABS employee with access to census data has the appropriate classification? Is there a minimum standard required? How do you guarantee that staff who access census data are using it appropriately?

Mr Kalisch : I will ask Mr Palmer to provide you with a sense of some of the protections we have in place.

Mr Palmer : The ABS ensures that only staff with appropriate security clearances have access to security classified information. We control access to sensitive information on a need-to-know basis, so only the staff who need to access that information have access rights to those datasets. We provide training to all staff so that they know what is expected of them. We have a number of controls in place to detect and prevent fraud. We require staff to undertake training around disclosure of conflicts of interest and to make appropriate declarations, and we investigate those. We also screen prospective and existing employees, contractors, consultants and suppliers who operate on ABS premises. We require all ABS staff and contractors to sign undertakings and indeed declarations that require them to be bound by the secrecy provisions of the Census and Statistics Act—or it could be the ABS Act. I think it is the Census and Statistics Act. So all our staff are very clearly aware of their obligations to maintain the secrecy of information that is collected under the Census and Statistics Act.

I use the word 'secrecy' instead of the word 'privacy' because our legislation has those secrecy provisions, and they require that we do not disclose any information in a manner that is likely to lead to the identification of an individual. ABS staff are very much aware of that. We ensure that those secrecy requirements are met through our disclosure practices, which see us confidentialise data appropriately, and through, as I said, our need-to-know access of systems.

Senator JACINTA COLLINS: Do you monitor and audit?

Mr Palmer : We do. We monitor access to our databases, yes, and we conduct audits.

Senator JACINTA COLLINS: Audits of whether all of those accesses have been appropriate?

Mr Palmer : Yes. Perhaps not all—auditing is usually around a sample—

Senator JACINTA COLLINS: You do a random audit?

Mr Palmer : but certainly we do regular reviews of access to a number of our more sensitive data holdings, which look at who has accessed them and whether that is considered appropriate.

Senator JACINTA COLLINS: Have you found problems?

Mr Palmer : Yes, we have. In 2014 there was a very public problem in the form of an ABS staff member who abused the trust placed in them. Actually, the access that they had was in line with their duties, but they misused that access for fraudulent purposes.

Mr Kalisch : That case—if I can just elaborate—was a case where someone was subsequently jailed for insider trading. They were disclosing final statistics that were market sensitive, so it was not individual information or information about a business, but it was necessarily market-sensitive data.

Senator JACINTA COLLINS: I understand. I understand that there are a range of potential sensitivities around this data, which is why, I suppose, I asked whether there is an ethics discussion process, because there are a range of potential ethical questions, probably far broader than what someone might think was an ethics matter, in fact. As you said, insider trading is a good example of essentially fraud or corruption that you might not necessarily think was an ethics issue in the sense of your data.

Mr Kalisch : The other mechanism I would also draw to your attention is the disclosure review committee that we have, which is an internal committee that looks at the output that we produce even in terms of some of the microdata and tabular releases to make sure that we do not breach the Census and Statistics Act provision about the secrecy of personal and business information. We have a committee that reviews the outputs that we produce as an organisation just to make sure that we are producing those in a manner unlikely to disclose those sorts of personal or business characteristics.

Senator JACINTA COLLINS: In the monitoring that you do, do you also monitor all the data that is sent outside the ABS?

Mr Kalisch : With regard to the microdata use, yes. We look at, I suppose, the context in which that data is being used, so the institutional context in which that data is being used, the professional context, and we also look at the output before that is released.

Senator JACINTA COLLINS: Do you get guarantees about who is accessing?

Mr Kalisch : We have undertakings from institutions and from people—

Senator JACINTA COLLINS: And do you test those undertakings?

Mr Kalisch : We do test them to a point.

Senator JACINTA COLLINS: Do you ask for evidence of who has accessed that data?

Mr Kalisch : We audit it. It is through either a virtual data lab or people coming into the ABS offices, so we can see the electronic trail.

Senator JACINTA COLLINS: Have there been any breaches of those policies reported?

Mr Kalisch : Not that I am aware of while I have been the Statistician. As I say, they are fairly tight.

Senator JACINTA COLLINS: If you did have data breaches or classified data was accessed improperly, how would that be reported?

Mr Kalisch : To most points, we try to stop it before it is actually released, so trying to the point of looking at the output that is proposed. If there are breaches, then there would be consequences. People, particularly if they are, say, in parts of government or in the research community, would lose access.

Senator JACINTA COLLINS: But there is no public reporting of the integrity of your information management system?

Mr Kalisch : If there were a breach of that nature, I would be confident that we would report it.

Senator JACINTA COLLINS: But where would you report it? That is what I am asking.

Mr Kalisch : We would report it on our website, in our annual report, to this committee and to the police.

Senator JACINTA COLLINS: At what magnitude? Sorry, I am trying to—

Mr Kalisch : Really where people have given an undertaking and they have broken that undertaking.

Senator JACINTA COLLINS: So breach of an undertaking would be the sort of thing that you would report. And to your knowledge no such thing has happened to date?

Mr Kalisch : Not that I am aware of.

Senator JACINTA COLLINS: I do not want to take too much more of the committee's time. I just want to go back to this paper that you were asked to respond to. You gave a response which was in relation to what might be feasible to collect to improve the SES measure. Is that the correct context?

Mr Kalisch : The context was that we were providing some advice to the education department about the feasibility or otherwise of the concerns that were raised by the Catholic Education Commission.

Senator JACINTA COLLINS: There were two issues. Earlier we were discussing their suggestions about alternative data that could inform the department of education, but the other issue is the concerns that they have raised about the existing data used to inform the education department's formula. I am interested in your response to the various concerns about how the SES scores are currently calculated. If it helps, I can run through each one of them.

Mr Kalisch : I suppose the aspect on which we were quite clear to the education department was that their use of ABS data was in effect a judgement that they needed to make. We see this in a number of instances.

Senator JACINTA COLLINS: 'This is not our judgement.' That is the answer?

Mr Kalisch : Yes.

Senator JACINTA COLLINS: When they say, for instance, in the first point here, 'The education and occupation dimensions in the SES scores are not relevant to capacity to contribute,' you say, 'Your judgement, not ours'?

Mr Kalisch : Yes.

Senator JACINTA COLLINS: Not, 'We have census data that tells us, "Yes, they're probably suggesting there's a problem here."'? You did not provide that type of assessment?

Mr Kalisch : It is not up to us to, in effect, provide the policy advice to the education department. They need to think about the data that is appropriate and relevant to their policy dimension. That is their call.

Senator JACINTA COLLINS: So it is their call to ask you for the census data that might inform a consideration of that matter? And you do not feel that they asked for that when they asked you to critique this paper?

Mr Kalisch : No, it is their call about what data they look to use and how they look to use that. We can provide some analytical assistance, and we will certainly do that and have provided that on request. But ultimately the way in which ABS data is used by any agency—and I am not singling out the education department, but, if the industry department or the infrastructure department or the regional development department were to use data in a particular way, that is ultimately their call because they are in the best place to understand the nature of the policy that they are looking to implement.

Senator JACINTA COLLINS: I can understand that part of the concern. For instance, the second point here is that the education and occupation dimensions and the SES scores overlap with other factors in the SRS model. I can understand that you might say, 'Well, that could be the case, but you'd have to review whether that is appropriate or not in your own circumstances.' But the first point, that the education-occupation dimensions in the SES scores are not relevant to the capacity to contribute—well, I thought that was something that the ABS could fairly easily have a view on, particularly around the myriad other data that might be more relevant to that factor.

Mr Kalisch : I suppose I would just query that this would really be our role. We would certainly advise the education department of the data sources that they could look at, but ultimately they have to make the call about the data that they use.

Senator JACINTA COLLINS: Okay. I am asking these questions because it was the minister who put you in it—

Mr Kalisch : The minister, through the department, asked us to comment on a paper written by the Catholic Education Commission.


Mr Kalisch : We have also—

Senator JACINTA COLLINS: And the paper raises issues well beyond just what the feasibility is of using direct income data.

Mr Kalisch : I would have to say that we also looked at what our remit is and what our comparative advantage is, and we commented on the aspects that we thought we were well placed to comment on. Not all those aspects—we saw a line between, 'We will provide some analytical advice to the education department,' which we did, 'But we are not going to stray into the policy dimensions.'

Senator JACINTA COLLINS: Okay. The third point here then partly relates to my question to you, which is about there being no consideration of family household wealth in the SES scores. As you said, there are issues about collecting data on wealth. But did your response to the department go to my other question, which is: if not wealth, what other indicators are there available to us that might be relevant to a household's capacity to contribute? It could be income or workforce participation and there may be others that I just have not plucked immediately.

This report refers to wealth; it probably should have referred to a myriad of other factors that are at the ABS's disposal. Did you provide any information to the department about what measures, other than wealth, may be relevant and could be put to their disposal?

Mr Kalisch : Just to give you a couple of examples, we did provide information to them that they could consider looking at equivalised income rather than household income.

Senator JACINTA COLLINS: What is that?

Mr Kalisch : Equivalised income is essentially taking account of the size of the family when you look at household income. To some extent it is almost getting at the participation aspect that you are thinking about—particularly if you have two adults and a number of children. In effect, it is looking at capacity to pay.


Mr Kalisch : That was one suggestion that we did point out to the education department. The other aspects were that we said, 'You can look at income data from the census or, alternatively, you can look at income data from the tax office.' So that is another data source you could think of that might or might not be of preferable to use.

Senator JACINTA COLLINS: Is the income data that comes from tax records matched in any way with families or households?

Mr Kalisch : In what way?

Senator JACINTA COLLINS: As you said, you could look at tax data but for individuals that is not necessarily going to inform family income on its own.

Mr Kalisch : I would have to take that on notice. I am pretty sure that with some of the aspects the tax office would be able to construct some family income dimensions.

Senator JACINTA COLLINS: Okay. This could be a very long conversation when we have it!

Mr Kalisch : I presume you are only meeting for a day on Monday?

Senator JACINTA COLLINS: Yes, but I do not think this bill is going to progress that quickly. So the next one was families and households with nil or negative incomes misclassified. Did you deal with that issue at all?

Mr Kalisch : I think that goes to the aspect of potentially looking at tax data as an alternative to the reported data in the census.


Mr Kalisch : There are some swings and roundabouts. Certainly, if I can give you one example: in the most recent Canadian census that they undertook, in 2016, they did not ask an income question as part of their census collection. Instead they drew that data from the tax records. They had looked at this issue over a number of years and thought that actually the tax records, for all of the vagaries that they can also be subject to, were on balance more reliable than the income information that is reported by people to the ABS.

Senator JACINTA COLLINS: You covered family household size for the SES score in your earlier comments, when we looked at equalised income. The area level data issue misclassifies individuals and families. Do you have a view there?

Mr Kalisch : I suppose we put a view to the Department of Education and Training that it is something that you could consider and could potentially explore. It was not clear to us whether it would make a difference or not.

Senator JACINTA COLLINS: If you used even smaller units?

Mr Kalisch : Yes. Some of these aspects were ones where, if you had the data, you could potentially do more refinement on the analysis, but it also was not clear whether it would actually make a difference without doing an analysis.

Senator JACINTA COLLINS: I ask that because, at the moment, they are using units of 400 households, yes?

Dr Jelfs : SA1s.

Senator JACINTA COLLINS: And that has been the case for yonks, yes?

Mr Kalisch : Yes, SA1s.

Senator JACINTA COLLINS: And that has been the case for many, many years?

Dr Jelfs : It has been the case from about 2001, when the model changed.

Senator JACINTA COLLINS: This is a question for the Department of Education and Training: what refinement could they be referring to? You do not know yourselves?

Mr Kalisch : To some extent—this is probably just conjecture from the Catholic Education Commission of Victoria—whether, if you change that approach, whether it would make a difference.

Senator JACINTA COLLINS: I certainly understand that issue. But the concern is the size of the unit, which is 400. The best example of this is inner-city Melbourne. There is an example in Fairfield, for instance, where you have a local Catholic parish primary school that essentially services the kids from the high-rise flats. All the other kids in Fairfield in middle upper class families go to the public school—I think the syndrome is called 'white flight'—but the poor little parish primary school is being rated at the same capacity to contribute as a high-fee independent school because of the way the statistical units are organised. Are you aware of that problem?

Mr Kalisch : Again, that is a policy choice for the Department of Education and Training as to the data that they do or do not use. It is not a statistical issue. It is a policy and analytical choice by the department as to the data they use.

Senator JACINTA COLLINS: The statistical issue is how you shape your units?

Mr Kalisch : We shape our units for SA1s. That is the geography we use. How people then use the data—whether it is SA1s, SA2s or otherwise—is, in effect, their choice.

Senator JACINTA COLLINS: How do you organise an SA1? If an SA1 has a whopping big housing commission flash in it, is any consideration given to whether the rest of that statistical unit leaves the data out of it reliable or unreliable?

Mr Kalisch : We have an Australian statistical geography standard that we use—ASGC—and it is done on that basis.

Senator JACINTA COLLINS: So there is no flexibility at all to factor—

Mr Kalisch : It is not with regard to socioeconomics. It is really looking at: what are some of the natural boundaries that we use for geography.

Senator JACINTA COLLINS: Would one of those boundaries potentially be halfway through a high-rise flat complex?

Mr Kalisch : I would be very surprised because, certainly with the work that I do with the Australian Electoral Commission, we do look at roads and we look at natural thoroughfares like rivers and bridges.

Senator JACINTA COLLINS: But you do not account for socioeconomic factors in that?

Mr Kalisch : No.

Senator JACINTA COLLINS: But you do count rivers, roads—

Mr Kalisch : We look at natural boundaries, so we do not, for example, run halfway through an apartment building.

Senator JACINTA COLLINS: I understand. All right, I will attempt to put into writing my request—

Mr Kalisch : That would be good.

Senator JACINTA COLLINS: to you. Perhaps directly would be the quickest way. That will guide the questioning that we will have further on Monday.

Mr Kalisch : I will probably suggest, given protocol, that you also send a copy to the minister.

Senator JACINTA COLLINS: Certainly. Thanks very much.

Senator BUSHBY: I have a couple of questions about the census. I understand the preliminary response rate for the 2016 census is more than 96 per cent. How does that compare to previous censuses?

Mr Kalisch : We are currently finalising the census data for release in June. On 27 June, census data will be available across about 110,000 different community profiles in quick stats. There will be further data releases in July. Alongside the census data, we will also release a rebased estimate of our resident population and also look to provide measures of the census undercount or overcount. The preliminary estimate that you talked about was something that we released late last year. It was something that we were doing on a tracking basis. We will have more accurate information that will be released on 27 June.

Senator BUSHBY: Is it likely to vary greatly from the 96 per cent in the preliminary?

Mr Kalisch : I think the reality is that ABS is confident about the 2016 census data. From all the checks that we have been doing, the processing that we have been doing, that the data is of high quality and comparable to previous Australian censuses.

Senator BUSHBY: I come back to my question, though. Assuming it is around 96 per cent—obviously, that is a preliminary figure and not the final one—how does that compare with previous censuses?

Mr Kalisch : That is comparable to past censuses.

Senator BUSHBY: Were census refusals high after the outage?

Mr Kalisch : After the outage, which took place for a couple of days, by the time we got to the Sunday after the census night—so, five days later—we had recouped and were back to the track or the path that we had expected in terms of census data provision. So we had got back to our expectations within five days of census night. In terms of actual refusals, we saw fewer refusals in 2016 than we saw in 2011.

Senator BUSHBY: That is interesting.

Mr Kalisch : So, in terms of giving you some numbers, we had 20 per cent fewer refusals—

Senator BUSHBY: Between 2016—

Mr Kalisch : And 2011. In 2011, there were just over 13,000, and in 2016 we had just over 10½ thousand. From those numbers, we then work with our field staff and other initiatives to work out who have legitimate excuses. There are people that have good reasons not to have responded to the census—because they were not at home or were not in Australia on census night—and we work through other compassionate reasons as well.

Senator BUSHBY: You have already mentioned that you expect the 2016 census to be of high quality. How will quality assurance of the data be assured following the outage? How is census data quality being assessed and checked? Who are the members of the independent panel that are doing that work?

Mr Kalisch : There are probably two aspects I will raise, and then I will ask Mr Duncan Young to talk through the detailed quality aspects that we do ourselves. There are two aspects to the data quality dimension. There is the work that the ABS traditionally does in terms of data quality assessment, and that is a very robust, rigorous and time-consuming process that takes a number of months. We have also, as you have pointed out, introduced an independent panel to provide quality assurance on the census results. They have been getting privileged access to ABS processes and information about how the census data quality checking is going. I can certainly take on notice the names or provide them to you now.

Senator BUSHBY: I think the chair would probably be happier if we took it on notice, so I am happy to do that, just in the interests of brevity, given we are considerably behind time. Mr Young, do you have anything to add?

Mr Kalisch : I will ask Mr Young to answer your question about the ABS data quality checks.

Mr Young : We are very excited about the release of census data—only 27 sleeps away now. In terms of the data quality, the first key point, which is what you raised earlier, Senator, is around ensuring a high response rate from the Australian public, and that is what we saw: a very high level of participation from the Australian public. The next key dimension of ensuring quality was making sure that we collected accurate information from the Australian public, and one of the really pleasing things that we saw in the 2016 census was an 80 per cent increase in the number of people reporting to the census online. So, when people reported to the census online, it prompted people if they left questions blank, it prompted people if the answers to questions were unexpected or invalid, it actually helped people fill out the census more accurately by taking them to the next relevant question for them and tailoring the questions based on their previous answers. So if a first—

Senator BUSHBY: It minimised the risk of human error, in a sense.

Mr Young : Absolutely. If the name of the first person in a household is David, then you ask the second person about their relationship to a first person and you ask about the relationship to David. We have seen these things right across the international community lead to higher quality data.

Our process certainly does not stop as the data comes through the door. Our confidence around the data quality has been built over the number of months since census night and the time we have spent processing and analysing the data. We check all the data for improbable or impossible kinds of data responses—for example, children who are older than their parents or five-year-olds with university degrees. Our data quality checks can pick up on those things and make sure we are not seeing unexpected responses coming through.

We also take the data and analyse it at small-area levels and for small population groups. A lot of the time we can compare the data against other reliable sources. If you want to understand whether or not our count of young children is accurate in the census, we can compare it against birth records over the last few years, or, if we are looking at different migrant groups, we can compare it against migrant records. So we go through a rigorous set of checking processes which help us understand that the census data is not only complete but aligns with expectations and will produce the high-quality dataset we are expecting.

The data quality panel—I can mention their names—is chaired by Professor Sandra Harding, the Vice Chancellor and President of James Cook University. Also on the panel we have Dennis Trewin, former Australian Statistician and former President of the International Statistical Institute; Professor Peter McDonald, Head of Demography at the University of Melbourne; Peter Morrison, a former Assistant Chief Canadian Statistician and a person who has let multiple Canadian censuses; Anton Voss, Deputy Secretary of the Tasmanian Department of Treasury and Finance; and Professor Lisa Jackson Pulver, Pro-Vice-Chancellor Engagement and Aboriginal and Torres Strait Islander Leadership at the University of Western Sydney. So, it is a very esteemed panel. We have been providing full support to this panel since census night. We have provided them with full access to all of the information they have requested, so they have had access to the data at a small-area level, access to all of our quality indicators and access to all of the information about our processes, our methodologies and all of the methods we followed to produce the census dataset. We look forward to seeing their quality report, when that is presented to the Australian Statistician, and we look forward to releasing that alongside the census data on 27 June.

Senator BUSHBY: Was the independent panel put together as a response to the outage or is it something you would normally do post-census?

Mr Kalisch : This is the first time that the ABS has put in place an independent panel. It was done largely as a response to the outage and the recognition that we needed to ensure that there was an independent assurance around the census process and the census data. It is something I would suggest the ABS does regularly with regard to every census from now on. I am also alert that a number of my international colleagues—heads of other statistical agencies—are also looking at the initiative we have taken to put in place this independent panel and are considering that for their own censuses.

Senator BUSHBY: Certainly, I agree with Mr Young. It is a very esteemed—eminent—panel. You put together some people with very deep and strong knowledge of the area, so I am sure they are doing their job well. Final question: how will your experience in managing the 2016 census and all the challenges that came with it inform preparations for the census in 2021?

Mr Kalisch : We are already using the experience that we built up within 2016 for how we are managing the 2021 census. There are a number of key features that came out of our own independent assessment of the process and our own reflections and perspectives on what went well and what did not go well, as well as the MacGibbon review and then the census inquiry that was undertaken by the Senate. To summarise a couple of key features, we have put in place stronger governance arrangements around the census. We have already committed to stay digital first—to pursue the e-form approach that was initiated in 2016—and to look at better partnering and risk management around some of the key census aspects.

We have committed to various aspects around privacy impact assessments and the timing of those features if that is going to be required for the 2021 Census. Perhaps one of the other things that I should mention around governance is that we have sought the services of KPMG to provide some independent assurance, some perspective outside the bureau, as well as a number of other people, including a senior person from the tax office, particularly with some expertise around risk management, to provide us with some further guidance, and some support from Telstra as well. This is something that we are using for our broader transformation program in terms of looking at external expertise actually sitting in on our governance structures and getting access to the processes.

The other aspect that I draw upon is that with the independent panel Mr Young mentioned Peter Morrison, who has joined us from Canada with that exercise. We have also taken the opportunity to draw on Mr Morrison's expertise around Canadian census development and to look at some of the initiatives that they have been progressing in Canada and thinking about their applicability within the Australian context, as well.

Senator BUSHBY: I look forward to watching that over the coming years.

Mr Kalisch : I think that is a reflection that the census does take seven years for each census cycle. There is five years of development of the census. There can be a number of ministers and a number of people even in the role of Australian Statistician through that time.

Senator XENOPHON: I have a short line of questioning and, unusually for me, it is going to be a very friendly line of questioning. The mission statement of the ABS is that 'we assist and encourage informed decision-making, research and discussion within governments and the community, by providing a high quality, objective and responsive national statistical service'. That is pretty axiomatic. I want to go to the issue of labour statistics. I think this goes back to 2006—concepts, sources and methods for labour stats. A matter that I think I have touched on previously is the definitions of employment used in ABS household surveys. The labour force survey gives a number of definitions of who is employed, including all persons 15 years of age and over who during the reference week worked for one hour or more, worked for one hour or more without payment in a family business or farm, or were employees who were not at work and who were at work for several weeks.

I will focus on the one hour a week. I am happy for you to take the details of this question on notice. I understand that this was changed a number of years ago to be in line with international definitions of employment. In terms of providing useful information for policy makers, for public debate, for governments and for parliamentarians, by and large, if you work one hour a week that is not going to cut it in terms of meeting your commitments for your mortgage or a whole range of measures. I think you know where I am going with this. Have you as the Australian Statistician, or has the Bureau of Statistics considered having another pathway of measuring employment that would be seen to be more realistic and to reflect the social impact of people being unemployed in the genuine sense? You measure underemployment, but I query the usefulness of the one hour figure, which I understand is something that we adopted because of an international norm a number of years ago.

Mr Kalisch : I will take that on notice as to when that was introduced. My understanding is that the international definition for employment and consequently unemployment has been pretty much in place for decades.

Senator XENOPHON: It is not very useful, though, is it?

Mr Kalisch : The aspect that you need to think about is that at the ABS we produce a range of statistics. We produce employment, unemployment and labour force participation according to the international definitions, which is the aspect you are referring to. We also publish underemployment aspects. We publish information on people who are working part time and who would prefer to work more hours. We also publish information about people who are working more hours than they actually want to work. So there are some swings and roundabouts, in terms of people's choices.

Senator XENOPHON: Overemployment?

Mr Kalisch : In effect. We also publish information on people who are marginally attached to the labour force. That is people who say that ideally they would want work, but are not undertaking active job search. So there are different categories and these dimensions—

Senator XENOPHON: Are these people that have dropped out of the workforce who have just become disheartened and think that it is not worth it any more?

Mr Kalisch : There are probably a range of motivations.

Senator XENOPHON: Do you measure the motivations?

Mr Kalisch : Yes. We ask for different aspects around that. My sense is that we do cover off things like where there are family or health perspectives. We do report these things. I am aware that there has been quite a bit of commentary, including fairly recently, about the employment statistics, and consequently the unemployment estimate, which did draw out these other aspects around underemployment. There are aspects around marginal attachment that we should refer to. Essentially, what I am saying is that the ABS also produces those statistics. We produce regular statistics on underemployment—

Senator XENOPHON: With what regularity?

Mr Kalisch : In terms of underemployment, I think it is at least quarterly. So it is quite contemporary.

Senator XENOPHON: Given the ABS's mission statement, which I know you take very seriously, to provide useful statistical information for the purpose of good decisions being made, the fact that you are not deemed to be unemployed if you work for only one hour a week or more—do you have any concerns that that might be inherently misleading? If someone does two or three hours of work a week, and they are looking for full-time work, they are effectively pretty much unemployed. You are not going to be able to pay a mortgage or buy a car or pay your rent with that sort of income.

Mr Kalisch : Part of the issue with this goes to the nature of how I understand that Treasury, the Reserve Bank, the employment department and other policy commentators and policy developers would look at the data. They all know about this aspect of the definition.

Senator XENOPHON: But it is a bit misleading out there when you hear an unemployment rate of 7.3 per cent, as it is in my home state, which is much too high—but the true figure is that that includes people who do one to three or four hours a week.

Mr Kalisch : I would not get too caught up with the one hour dimension. Certainly the latest information suggests that from the sample of 50,000 people that we draw on in the labour force survey we have less than 50 people who report that they only work one hour.

Senator XENOPHON: This material is all available, isn't it?

Mr Kalisch : I am happy to disclose this. This is not secret information. It just gives you a sense—

Senator XENOPHON: I do not want to waste the committee's time or incur the wrath of the chair, if the information is already out there. Do the current statistics breakdown who works for one or two hours a week or five hours a week?

Mr Kalisch : No. We would not publish at that level.

Senator XENOPHON: Can I ask you for that?

Mr Kalisch : I am happy to take that on notice as long as we do not breach any confidentiality or other aspects. I presume we will not.

Senator XENOPHON: I am not asking you to breach confidentiality in any way. Presumably there is more than one person in the country who works just one hour a week.

Senator KIM CARR: There has been some other commentary about the reliability of the statistics and around the survey itself. There has been commentary about the adequacy of the survey. Are you familiar with criticisms of the adequacy of the survey?

Mr Kalisch : Your colleague Senator Ketter covered off that aspect earlier.

Senator KIM CARR: What was your response?

Mr Kalisch : The response I gave to Senator Ketter is the same response that I will give to you, Senator Carr. The analysis that we have undertaken on the labour force survey suggests that there is less month-to-month volatility in the labour force survey now than there was in the 1980s and 1990s. The other part of the response was also that the ABS has been publishing trend estimates of the labour force estimates since 1987, really to deal with some of these aspects about the month-to-month volatility in the series. We would recommend that people use that as the broad indicator of the labour force.

Senator KIM CARR: But you are satisfied that the survey is valid. What is the margin of error?

Mr Kalisch : The standard errors are comparable internationally. In fact, in Australia, I would suggest that we have one of the highest-quality labour force estimate surveys in the world.

Senator KIM CARR: So what is that margin of error?

Mr Kalisch : It depends on the nature of the estimate itself. We publish it; it is in our labour force estimates bulletin. It depends on whether it is a national estimate, whether it is a state and territory estimate, whether it is on employment, unemployment rate or labour force participation. If you like, I can take something on notice and provide that to you.

Senator KIM CARR: Thank you.

Senator XENOPHON: Thank you very much for offering that. I just want to confirm that on notice you can give us a breakdown of the people who are underemployed, if you like—those who want to work more, those who are deemed to be employed but who work from one to three hours, three to six hours.

Mr Kalisch : We would have to look at it.

Senator XENOPHON: Can you break it down so it is a bit more specific?

Mr Kalisch : We will try to be as helpful as we can, Senator.

Senator XENOPHON: Yes, but it will not be like three to 30 hours or anything like that.

Mr Kalisch : No, we will try to make it meaningful numbers, but the information on underemployment—

Senator XENOPHON: Good, we are always searching for meaning in this place.

Mr Kalisch : The information on underemployment is a little bit more complex than, say, the information on people who are just employed.

Senator XENOPHON: If you can give greater particularity, I would very much appreciate it, and I am sure the minister is very interested in these figures as well.

Senator Sinodinos: Yes.

CHAIR: Thank you, Mr Kalisch and Mr Palmer, for joining the committee today. The committee is going to take a break. I am very glad you pushed us just over the edge, Senator Xenophon, because now I can say that we will resume at 10 past six. We will go straight to the Anti-Dumping Commission and we will see that one through, and after the Anti-Dumping Commission we will go to the dinner break.

Proceedings suspended from 17:51 to 18:10