Proceedings of the Standing Senate Committee on
Social Affairs, Science and Technology
Issue No. 33 - Evidence - December 6, 2017
OTTAWA, Wednesday, December 6, 2017
The Standing Senate Committee on Social Affairs, Science and Technology met this day at 4:15 p.m. to continue its study on Bill C-36, An Act to amend the Statistics Act.
Senator Art Eggleton (Chair) in the chair.
[Translation]
The Chair: Welcome to the Standing Senate Committee on Social Affairs, Science and Technology.
[English]
My name is Art Eggleton, senator from Toronto and chair of the committee. I would like the remaining members of the committee to introduce themselves, starting on my left.
[Translation]
Senator Petitclerc: Chantal Petitclerc from Quebec.
[English]
Senator Cordy: Jane Cordy, Nova Scotia.
Senator Dean: Tony Dean, Ontario.
Senator Bernard: Wanda Thomas Bernard, Nova Scotia.
Senator Manning: Fabian Manning, Newfoundland and Labrador.
[Translation]
Senator Mégie: Marie-Françoise Mégie from Quebec.
[English]
Senator Raine: Nancy Greene Raine, British Columbia.
Senator Seidman: Judith Seidman from Montreal, Quebec.
Senator Frum: Linda Frum, Ontario.
The Chair: Thank you very much.
Today we are discussing Bill C-36, An Act to amend the Statistics Act. This is our third meeting on the subject. We have two panels today, and this first panel will go one hour, until 5:15.
I’m very pleased to welcome Ian McKinnon, Chair, National Statistics Council, a council we’ve talked about quite a bit in the course of our examination of this bill. Also, we have Mel Cappe, Professor in the School of Public Policy and Governance at the University of Toronto. He is a former Clerk of the Privy Council of Canada and has held numerous other public service positions. Welcome to both of you, gentlemen. Mr. McKinnon, would you please start?
Ian McKinnon, Chair, National Statistics Council: Thank you, Mr. Chairman. I’d like to begin by thanking the committee for their invitation to appear here and offer comments on Bill C-36.
Since its inception 30 years ago, the National Statistics Council has always tried, in its role of advising the Chief Statistician, to promote a national statistical system which produced data and analyses which were as relevant, accurate, timely and accessible to as wide a range of users as possible.
The issue of relevance is particularly important and has been an underlying theme in some of the discussions in the earlier sessions of this committee’s hearings.
Relevance implies understanding the information needs of all Canadians, whether government policy-makers, service deliverers, NGOs, companies large and small, communities of interest, researchers or individual Canadians from all parts of this country.
In considering Bill C-36, the National Statistics Council applauds efforts to increase the independence of Statistics Canada in professional matters and the changes that the bill makes to the position of the Chief Statistician.
Those go to the heart of the objective of having an accurate statistical system. In particular, we support the changes which give the Chief Statistician authority over the statistics program, “. . . based strictly on professional statistical standards . . . .”
While our parliamentary Westminster system gives ultimate authority to Parliament and the cabinet, and the act explicitly provides for the cabinet or minister to issue directives related to the operations or the statistical programs of StatsCan, the act also ensures greater transparency by requiring that the directives be tabled in the House of Commons directly or at the request of the Chief Statistician.
The council also has, over the years, been supportive of an open search process for future Chief Statisticians, a topic which was discussed extensively in your earlier meetings.
The bill’s provisions are also consistent with the council’s public statements during the debates over the long-form census recommending the removal of imprisonment as a punishment for not completing the census.
As my last topic, I want to talk about the new Canadian statistics advisory council and the evolving advisory structure for StatsCan itself.
First, let me thank Ivan Fellegi and Wayne Smith for their kind words in the testimony they gave last week about the work of the National Statistics Council over the years.
While I’m a little saddened by the demise of the NSC, which would occur with the implementation of this bill, we have seen a broadening and deepening of the consultative efforts of StatsCan over the years as well as an increase in the number and the professionalism of its technical advisory committees.
The increased independence given to StatsCan and to the Chief Statistician in Bill C-36 also means that there is more need for a council that has formal responsibilities and that can provide advice on more strategic issues.
I am particularly pleased that the new advisory council will make a public report on the state of the Canadian statistical system. This is an opportunity to raise issues and talk about information and data needs that go beyond just the jurisdiction and operational needs of the federal government and Statistics Canada.
As an example, in this era of big data and the increasing capacity to link administrative records and data files, there is great opportunity but also increased need to assure anonymity and confidentiality of data records.
Let me use, as an example, the opportunities to better understand and design services for marginalized populations. By gathering data that are produced from censuses, surveys and institutional records from educational systems, social services and the health sector, we can build a better understanding of best practices, monitor progress and set in place more effective policies.
These, however, all refer to the national statistical needs and systems rather than solely as advice to StatsCan on its operations.
The new advisory council will have the opportunity to address many broader issues like these. As well, they will be able to report on government-wide issues in terms of the development and analysis of statistical information. Once again, this can be a platform for looking more broadly at needs and issues rather than solely at the operational challenges facing StatsCan.
Now, several members of this committee raised the issue of the representativeness of a committee limited to 10 plus the Chief Statistician. I’ve stressed this is my personal view; we haven’t discussed it at length in the council.
My personal view on this issue is that the new council’s form, particularly its size, should be determined by its role. I very strongly agree that a council whose members lived experience rarely reaches beyond the Ottawa-Montreal-Toronto triangle is likely to miss many important perspectives and much knowledge.
On the other hand, given that this new council can play a strategic role in assessing and advising on the state of the national statistics system, it needs members who are knowledgeable and focused on that primary goal.
In the past, when the Chief Statistician has sought advice from the NSC on the kinds of issues that the new council may face, the NSC has often created working groups of a handful of members with experience beyond research who worked in depth on the topic to provide advice to the Chief Statistician.
This also underscores the importance of StatsCan’s deepening its advisory and consultative mechanisms. It can only make the whole system stronger if StatsCan better understands the underlying needs of all of its data users and the people affected by the information it collects.
In summary, we are very pleased with the moves that increase the professional independence of the organization, and we look forward to more success for the Canadian statistical system that may flow from the changes that are being countenanced. Thank you.
The Chair: Thank you very much, Mr. McKinnon, for your presentation and for your service to the National Statistics Council.
[Translation]
Mel Cappe, Professor in the School of Public Policy and Governance at the University of Toronto, former Clerk of the Privy Council, as an individual: I appeared before the House of Commons committee on April 6, 2017, for the study on Bill C-36. Prior to that, the last time I appeared before a parliamentary committee was on August 27, 2010, on the abolition of the long-form census.
I will continue in English, but I am ready to answer your questions in French or English.
[English]
I want to start off by saying that although I was the Clerk of the Privy Council to Jean Chrétien, the first time I was appointed to the rank of deputy minister was by Brian Mulroney, and I ended my career serving Prime Minister Harper. This is not a partisan approach. Frankly, statistical issues shouldn’t be partisan. Now, I don’t detect any sign of that in this committee, but I have to say that in the house back in 2010, I don’t think that was the case.
I have three points, basically.
The first is that statistics are a public good. That’s a technical term in economics, but it’s used by a wide array of real people, banks, charities and public health authorities. The state can collect and analyze statistics at a lower cost than requiring everyone to collect their own. That means it is genuinely a public good.
The second point is that the Statistics Act should minimize coercion. I want to change that. I know I wrote minimize, but I want to say optimize coercion because, as you’ll see in a moment, we should be minimizing the intrusiveness and maximizing the privacy of data as much as possible.
In 2010, I proposed that you could remove jail terms. You can review the questionnaire and minimize the intrusiveness of the questions, and I’d add that as the National Statistics Council has said, you can increase the penalties for the divulgation of private data.
I think it’s in that sense that I’m looking for optimization and not minimization. Minimize the intrusiveness and minimize coercion on some elements, but should someone divulge confidential information, I think there should be serious fines. I think there is a sense of optimization here.
My final point is that the governance of Statistics Canada can be improved. The higher principle is to ensure the integrity of the statistical agency. The public and the users of statistics need to know that there is no political interference or challenge to the collection or analysis of the data. I think the events over the course of 2010 raised questions about this larger and, I think, more significant issue.
I think the committee should consider the UN Fundamental Principles of Official Statistics. To my surprise, there is such a thing, and I think that makes it clear that there has to be an individual responsible for the methodology in collecting statistics, and that should be the Chief Statistician.
There are several principles of that UN charter that deal with independence, methodological integrity and the role of politics, and it’s worth following them.
I said to the House committee in 2010 that they might consider amending the Statistics Act to make clear that the Chief Statistician, who is a statutory officer named in the Statistics Act and appointed by the Governor-in-Council, has the sole responsibility for methodological and technical issues, and Bill C-36 does that. I think that’s a good thing and an improvement.
However, I also think there’s a legitimate role for politics in statistics. By that I do not mean methodologically; I mean in the definition of what issues we should be addressing. That is inherently a political question. What are the grand questions facing Canada over the horizon, over the next decade and over the next millennium?
StatsCan is a department of government that reports to the minister, and many issues over the choice of the census questions are indeed political. The identification of ethnic origin, for instance, is inherently political; similarly, norms on gender identification are changing and are, in themselves, political. Having that kind of opportunity for political direction is not a bad thing in itself.
But there’s no doubt in my mind that the Chief Statistician should be the only person to comment on methodological questions in government and have the obligation to inform the chair of a parliamentary committee, or someone in public, of his or her views on methodological questions. I would urge the committee not to play partisan games with an important institution of governance. Rather, I think Bill C-36 accomplishes the objectives that previous committees have identified.
What I said in 2010 and what I said earlier in 2017 still applies.
Statistics are indeed a public good. It’s highly appropriate for the state to collect and analyze statistics. I support the objectives of this bill. It confirms the legitimacy of the collection and analysis functions of the agency.
In optimizing coercion and reducing penalties for violation of the act, removing criminal prosecution and jail terms is probably a good thing, but keeping some penalties is still necessary to optimize coercion.
Finally, the changes to the governance structure of the agency strike me as appropriate, creating the statutory advisory council and legislating the independence and responsibilities of the Chief Statistician for methodology.
So this is a good piece of housekeeping that modernizes the Statistics Act, and I’m happy to answer any questions that people have, although I can’t imagine there any questions.
The Chair: I didn’t know optimizing coercion was a good thing. Optimizing compliance, maybe?
Mr. Cappe: It’s a question of how you encourage compliance, and at some point coercion might be required.
The Chair: We’ll leave that up to people’s judgment as we go here.
Thank you very much for that. Your experience, Mr. Cappe, over your years in government is very valuable in this.
As is traditional when we have a bill, the first question will be from the sponsor of the bill, the second from the critic of the bill and then the two deputy chairs.
I’ll start with the sponsor of the bill, and that’s Senator Cordy.
Senator Cordy: So is that one question or two questions?
The Chair: Two questions, but concise, not too long.
Senator Cordy: I can do concise. Thank you.
Thank you very much to both of you for appearing. As the chair said, your experience in government is helpful in this. Mr. McKinnon, thank you very much for your work on the National Statistics Council. The minister himself spoke about the contribution the members of that council have made.
In the media and at the second reading of this bill we heard of a possible amendment that the appointment of the Chief Statistician should require the approval of both houses of Parliament.
Now, on Monday in the Senate, we had a hearing with the candidate for the Commissioner of Official Languages, and on Friday we’re hearing from the candidate for the Commissioner of Lobbying. These commissioners and all officers of Parliament report directly to Parliament, so I understand why they are appearing before both houses.
The Chief Statistician reports directly to the minister, not directly to both houses of Parliament, so should the Chief Statistician need the approval and appear before both houses of Parliament? Would this lead to approval of deputy ministers, of going down the line of people who directly report to the minister? I wonder if I could have your comments on that.
Mr. Cappe: Let me start, but I’ll be interested in Mr. McKinnon’s response as well.
It is clear that Statistics Canada is an arm of the executive. It’s an agency of the executive of government, not the legislature. As Senator Cordy noted, unlike the agents of Parliament, the statistics agency is meant to be in part of government. There’s a danger of politicization that I think would be problematic.
Now, I’ve said before, and as I said in my opening remarks, there’s a role for politics in deciding on questions, in deciding on the issues that should be addressed — not the form of the question, but the issues to be addressed. That is exactly an executive function, and it’s in that context that I think there’s a role for government, not Parliament, in assuring that the Chief Statistician is the right person for the job.
Finally, it really comes down to a question of accountability. The government should be held to account for the appointments it makes. The government should be held to account for the operations and the nominations of individuals that it makes to head those agencies. That’s what takes place in a parliamentary committee, ultimately, and the minister in the house, when the — if there were, God forbid, a bad appointment. There has to be accountability. To put the accountability on a parliamentary committee, I think, removes the government from its responsibility.
Mr. McKinnon: I’ll come to the same conclusion but for a different reason.
One of the things that are fundamental to a well-functioning statistical system in a country is that the public has confidence in its non-partisan nature, its openness and fairness.
Something that I play through in my mind, we had a very high-profile resignation of Munir Sheikh as Chief Statistician, and had his replacement gone before Parliament or parliamentary and Senate committees for approval, I’m fairly confident it would have devolved into an acrimonious political debate. That’s fine, but if it were seen to taint the apparent independence and professionalism of StatsCan, that would have an effect on the statistical systems, people’s willingness to give information and respond to the census honestly and in very high numbers and so on.
I use that case as a worst possible outcome, but I could see a parliamentary approval system making matters significantly worse for the agency. Not that they would pick the wrong person, but it would create the impression in some people’s minds that the statistical agency was not neutral and not professional.
Senator Cordy: That’s very helpful. Thank you both.
Last week we heard from Mr. Fellegi, a former Chief Statistician, and he said that it should be in the legislation that there be a three-person advisory panel tasked with finding and presenting candidates to the minister for Chief Statistician.
So I guess I’m wondering whether you feel that it’s always necessary. Should it be in legislation tying the hands of future ministers that there always be a three-person panel set up? Do you think that the process should be left to the discretion of the minister? If there is succession planning in place, if it’s in legislation, you would still have to set up this three-person panel to look for a Chief Statistician.
Mr. Cappe: Mr. Chair, I would ask whether you think it should be in legislation that we require a panel to review potential nominees to the Senate. I mean, you can do that, and we’ve seen this Prime Minister do that, and we’ve had excellent appointees, but you didn’t have to put it in legislation to do that. I don’t think putting it in legislation is going to be particularly useful, whereas a wise Prime Minister, as he or she is thinking about nominating somebody to be Chief Statistician, will think about the best person for the job for the time and will consult with the Clerk of the Privy Council, presumably, and potentially other people, like the national statistical advisory council.
Mr. McKinnon: To take a slightly different view, first, there was a very successful experience of this when Statistics Canada in the late 1960s and early 1907s was in some disarray. A panel of very senior people searched broadly and found Martin Wilk, who in many ways set the modern Statistics Canada in place and resolved many of the problems. So the potential to do this is at times very, very useful.
Other systems actually have public advertisements and let people apply. Australia and New Zealand, which are fairly similar to our system, work in that fashion. I think it is not a necessity, and you may have an issue of tying your hands when there has been a fairly clear train of succession.
On the other hand, I think it’s the sort of thing which ought to be on the table for any cabinet or Prime Minister and something they think about very seriously.
Senator Cordy: Thank you.
The Chair: Senator Frum, critic of the bill.
Senator Frum: Mr. Cappe, if I can start with you, I think we’re all agreed that the intent and purpose of this bill is to improve and increase the independence of the Chief Statistician. I just want to question if that really is going to be the result of the bill.
One of the questions I have is the change of status for the Chief Statistician. Currently the Chief Statistician is a deputy minister under the Minister of Industry, and under Bill C-36, he or she will become the deputy head of Statistics Canada. Can you explain how that is different?
Mr. Cappe: There are several agencies of government that have deputy heads named to them and several people whom we style as deputy minister who may have different titles. We have the Commissioner of Revenue in Revenue Canada, who is the deputy minister of what was the Department of National Revenue, now the Canadian Revenue Agency. So we have seen the evolution of the title but not the nature of the body or the function. This is, I think, comparable in some respects.
I would argue the key issue here is the term. If I recall, it is an appointment for five years. I think that adds a degree of independence that frankly doesn’t exist as far as I can see. It does with the Governor of the Bank of Canada, which is a seven-year term, and that is not a deputy minister of the executive branch because the bank is totally independent, but I think this would be another governance innovation, if you will.
Senator Frum: It is really a difference without a distinction?
Mr. Cappe: If there was pressure from the government of the day, this gives the Chief Statistician a very high horse to get up on and dictate.
Senator Frum: Because of the title?
Mr. Cappe: No, because he or she is independent.
Senator Frum: Because they serve on good behaviour. That is why. That is the issue. It’s not the title. That is sort of window dressing.
Mr. Cappe: Everyone serves on good behaviour, and all the deputy ministers as well.
Senator Frum: Isn’t that the fundamental change? I am trying to understand, actually, what Bill C-36 does and how it makes any difference whatsoever. I don’t think that is the case now, actually. Now they serve at the pleasure of the minister.
Mr. Cappe: That is true. The pleasure of the Governor-in-Council, that is, the Prime Minister.
Senator Frum: So now they have the security of the serving on good behaviour.
Mr. Cappe: Right. So “good behaviour” is a high standard to remove someone, and it is renewable.
There is an interesting question about renewability. As far as I know, we have never renewed a governor of the Bank of Canada. That has reconfirmed the independence of the governor. Renewability makes the person, arguably, more subject to the government of the day. If that is where you are going, I can understand that.
Senator Frum: Actually, it wasn’t where I was going, but I am glad you mentioned that because that is a good point. Maybe it should not be renewable. Maybe it should be two terms.
If the purpose is greater independence, I am trying to find why we need Bill C-36. We now see that the title change is completely irrelevant because they are a deputy minister either way. That doesn’t change their status. You are saying, if they had a renewable term, that even the good behaviour is undermined — the job security that you are supposed to have with the incentive of a renewal undermines the independence.
Mr. Cappe: Ivan Fellegi, who appeared before this committee, served for almost 20 years as Chief Statistician. It is desirable to have continuity.
Senator Frum: But independence and renewability are in conflict?
Mr. Cappe: I agree with that. That is exactly right. So you are looking for the optimal degree of satisfying two conflicting objectives. I think this is a reasonable balance. I think seven years would be too long, quite frankly, and I think three years would be too short. Five years is Goldilocks.
Senator Frum: But non-renewable would be Goldilocks.
Mr. Cappe: No. I am saying that if the right person were there and they were early in their career and in the middle of a transformation of the organization, you might definitely want to renew them. Continuity is valuable as well. I think there is a choice there. But I think this is a fundamental change. This will be the only deputy minister-ranked person who sits around the deputy minister table that actually has a term as far as I can think.
The Chair: Mr. McKinnon, do you want to weigh in on this?
Senator Frum: May I ask one more question?
The Chair: I think I heard the equivalent of two questions.
Senator Frum: That was really one topic area.
The Chair: A brief one, but I would like to give Mr. McKinnon a chance to respond.
Senator Frum: I will address it to Mr. McKinnon. It is a different question.
I want to get quickly to the pure science of statistics. You are a statistician, I take it, and this idea that the methodology is pure science and the only place where the politics comes in is in the direction. However, in Mr. Cappe’s testimony, he made the point on issues like ethnic origin or sexual orientation that it is the choice of questions that is political. But isn’t it the choice of questions that is the methodology?
Mr. McKinnon: I will give you an ambivalent answer. Here it is: The topics are the policy and small-p political issue. What are the priorities of the government to learn about through a census or another survey? The best way to ask a specific question is something which can be tested and where there isn’t always a perfect physics or mathematical answer, but you can see costs and benefits to different ways of asking a question. The general methodology is we go through that kind of process where you test questions to see how accurate they are or the responses appear to be, how easily understood and so on. That part of the process is rightly a methodological issue.
Senator Frum: Would you call them an objective science, statistics?
Mr. McKinnon: Statistics is. In a very pure sense, this is a pure mathematical end of statistics. The exercise of doing surveys is a scientific-like process of test, evaluate and revise, each time improving. It is not the same as abstract mathematical statistics.
The Chair: All right. We will have to move on.
Senator Petitclerc: Thank you both for your presentations.
Mr. McKinnon, I would like to hear from you a bit more in depth. You talked about the new advisory council and said it will be listed in this bill. How important is it that it be actually in the bill as opposed to what it was before?
You mentioned this a bit, but I also want to hear about the number. Is 10 a good number? Do you feel confident that it will accomplish what it has to accomplish with that number?
Mr. McKinnon: There are two different parts. One is I think it is very important that it be in the bill. The National Statistics Council was constituted to provide advice to the Chief Statistician. This was not in regulation or legislation. That meant that we never — only under extreme circumstances — spoke publicly on issues. We were asked for our response on the cancellation of the long-form census, and we provided what was effectively private advice to the Chief Statistician.
One of the major changes in this bill is requiring an annual report on the state of the statistical system and creating the group in legislation which gives it the full authority and responsibility of tackling that big issue in a recurring fashion. For me, that is a big difference.
The second question you ask is about the number of people to involve. My experience, working both in the private sector and in government and with volunteer groups like the National Statistics Council, is that the appropriate number is a product of the nature of the issue you are facing, number one, and number two, the responsibilities you have. It does have to be more focused than the National Statistics Council, which was in many ways a broad collection of experts from across a very wide array of areas. That is not an appropriate organization to perform the function that Bill C-36 envisages for the new council.
Having said that, I will quote from the two previous Chief Statisticians. There are people on the council today who would be admirable as members of the newly constituted group, but not all because they had different purposes.
Regarding size, you do want a range of areas of expertise, and as the former Chief Statistician said the other day, a range of lived experience, but you want to make sure that it functions tightly and, therefore, is much smaller than the current one.
Senator Petitclerc: You talked about pure statistics versus political input or impact or influence. Since we have started studying this bill — and perhaps Mr. Cappe can add to this — we have been talking a lot about autonomy, transparency and independence. That puzzles me. Do we achieve that to the level we want, namely the independence of the Chief Statistician, when we consider the input on the quality or choices of the questions versus the purity of the science?
If we add to that the element of ministerial discretion that can be used, how important is the purity of the statistics to the Chief Statistician, and are we achieving that with this bill?
Mr. Cappe: I think this goes a long way to achieve those objectives, yes.
You can’t put every contingency in legislation, so there is a vagueness inherent in all pieces of legislation that I think is desirable. This bill is very specific, though, about the kinds of issues of methodological choice that are given to the responsibility of the Chief Statistician. There is a clarity that comes in that I think achieves the objective. As I suggested in my response to Senator Cordy, this gives the Chief Statistician a high horse to get on, should that kind of political pressure be presented.
I think that independence is very clear.
There is a political consequence to the political action, and we shouldn’t forget that. We can’t put all the contingencies in the bill, but if we make it clear that making public the directive of the minister has a political consequence, that is important as a constraint on the discretion of the government of the day.
Senator Seidman: I would like to address my question to you, Mr. McKinnon.
We have heard from witnesses who point to the technological issues with Shared Services Canada as a significant barrier to Statistics Canada’s independence.
In your previous comments before the House of Commons Industry Committee on this piece of legislation, you indicated that you served on a working group where you canvassed the former heads of other statistical agencies in Britain, Australia and New Zealand, who said they felt that a significant element of their independence flowed from their control of their informatics structure.
Could you expand, please, on this process? What specifically has been the experience of international jurisdictions with regard to IT?
Mr. McKinnon: I cannot give you a detailed answer on their specifics. I can, however, say we talked directly to the equivalent of the Chief Statistician in the United Kingdom, Australia and New Zealand. All of them felt it was important. They did not talk in highly technical detail.
I want to make a distinction here. In the testimony you heard earlier, Shared Services talked about the technical expertise. That is part of the security of the data, and it is absolutely important.
The other is a deep cultural distinction. StatsCan has a deep culture of confidentiality. I have no way of assessing the internal culture of Shared Services. I take it they are highly sophisticated technically. You have to meld those two over time. I cannot give you a sophisticated view of the current internal situation on Statistics Canada and its interaction with Shared Services.
Senator Seidman: Are there other areas where Canada’s approach differs from its international counterparts?
Mr. McKinnon: The biggest one is not around security. We have a unified system. Britain and the United States have incredibly diverse and segmented systems where different agencies collect the data. There are bureau of labour stats and agricultural statistics, and so on.
Ours, it turns out, is considerably more efficient and, I would argue, a better and more sophisticated one. That is the biggest difference internationally. Canada is often held out as an organizational model around the world of how best to organize the statistical system.
Senator Seidman: That is good to know.
Senator Bernard: My first question picks up on a question that Senator Petitclerc was asking around the council and the number of people.
I am thinking about our move towards a gender-based analysis to all of our work. If we were to bring such an analysis to the council, how do you think that would impact or influence the things that you have identified in terms of the range of expertise and the lived experience? What difference would bringing that gender-based-plus analysis make, and should that be specifically identified in this?
Mr. McKinnon: I think you will find, through an appointment process, that there are many people who would be useful in service on a council like that envisioned in the bill.
Part of that diversity is not only region but also gender. It is important to look at it; we have in appointments to our own council been careful to cover regional, racial, ethnic and linguistic divisions.
The driving force, though, because of its very specific and very technical mandate, has to be that. You can add various layers to ensure that you are not replicating the same experience and expertise over and over again.
Senator Bernard: This picks up on the comment about optimizing coercion. There are some communities that feel they are heavily under surveillance and therefore may not participate — that is, more vulnerable populations.
Could you elaborate on any ideas you might have for collecting data in non-invasive ways, protecting the rights of those vulnerable populations?
The Chair: Both of you might want to add to this.
Mr. McKinnon: This is something I personally care about a great deal. Some of our indigenous communities have not participated in the census; particularly marginalized people, the homeless, have been difficult to count. I think not accounting for them properly and fully harms them. Statistics Canada already does very significant outreach work. I don’t think the issue is a coercion or any penalties. Frankly, if you cannot find a person on census day who is homeless, sleeping under the bridge in Victoria, you’re not counting them because they refuse and they are afraid of paying a fine or not afraid.
Some of the people for whom it is most important that we count them and that we have information so that policies can better serve them are precisely the kinds of populations that you’re talking about. So I give them the highest importance, although I think the coercive elements and the potential fines have little bearing on our success or failure at getting information from them.
Senator Bernard: What solution would you propose?
Mr. McKinnon: For example, very simple things that have been worked at extensively: working with homeless advocates and with homeless shelters; going, as StatsCan has, to many of our indigenous communities to show them the value to that community of things like the census. To know themselves better gets people involved. Then you start having people who not only understand what’s going on but also realize the importance to the community they either speak for or are in.
Mr. Cappe: I would just add that I agree with everything that Mr. McKinnon said. The “how” is important. For some people who are visually disabled, you might need somebody to do an interview. For other people who are hearing impaired, you might need a different way of gathering the data, but it’s not a question of whether the data should be gathered.
Senator Poirier: I just have one question. When we met the minister a couple of weeks ago or last week, I questioned him on the representation on the committee, going from the number of 40 to 10, whether there would be a regional representation. Actually, he told me that there would not be a regional representation, that out of the ten, four or five could come from Prince Edward Island, or they could all come from one province.
I’d like to hear what your thoughts are on that, if you have any concerns about that, or if it should be spread out to represent Canada. Again, being just 10 instead of not having one in each province, I had concerns with that, specifically with the Maritime provinces or the small provinces. I was wondering if we were even going to get one in the Maritimes. At this point it doesn’t seem to be of importance to them to have regional representation. I just wanted to hear your thoughts on that.
Mr. McKinnon: They are the following: When you have a small council, it’s tough — if you start out saying, “We have to have every region represented,” that may not provide you with the best council.
Similarly, as I mentioned in my opening remarks, having a council constituted entirely of people who have spent their careers in the triangle is not just an issue of representation. It’s a competence and knowledge issue. The lived experience, the local knowledge of social and economic issues will be fundamentally different.
Part of having the best possible representation would be to have diversity in the representation as well as high technical competence. Let me end it there.
Mr. Cappe: I again agree, but I would just add that this won’t be the only consultative mechanism that the agency will use.
Mr. McKinnon: I should add that one thing that has happened at StatsCan over the years is that they have tremendously deepened their ongoing consultation with provincial and territorial organizations. I think one of the next big frontiers is making sure that towns and cities become much more sophisticated users.
We have a real problem now where you have a half dozen very sophisticated regional areas that make really good use of the data and other places where the skills are absent. So there is an education as well as a consultation process that needs to be done.
Senator Griffin: Thank you for being here today, expert panel.
My concern is that in 92 years after the 2016 Census, there’s going to be a lack of data available because of the fact that the people either did or did not tick the box. My concern is that in many cases, I think, the consent or lack thereof was not very informed. The trick is how to fix that.
We were told by someone from the department that people can retroactively opt in to have their data released. In the next census, what wording would you suggest that Statistics Canada should use in an opt-in box to enable Canadians to opt in to releasing their records from the 2006 to 2016 censuses as well as the census that they may be doing now? They may never have noticed the box before. So what wording? How are you going to get their attention?
Mr. McKinnon: That’s actually, in many ways, a methodological question. My answer is you get good advice and test it. I would be the last person to try to answer off top of my head because there is a technical answer to this.
Senator Griffin: Okay. I’m happy with that answer. The other way of getting at this, of course, is at some point in time to retroactively legislate that that data will be available. I don’t think the current government has any appetite for that. That’s why I’m trying to beef up the voluntary aspect of this.
Mr. McKinnon: I would comment on one aspect of that. There seems to be very little evidenced that release 92 years in the future, or the distant future to most people, has a significant effect on their willingness to respond, because this has been going on for several decades.
On the other hand, Statistics Canada, deep in its culture, respects deeply the commitments of confidentiality that it has made to people. Therefore, I know part of one of their most admirable qualities, the respect of the confidentiality of responses, would make it very difficult for them at a cultural level for the institution to say, “Remember 10 years ago how we said you would never have to answer? We’re rethinking that.”
There’s actually a very serious outcome issue. If people are saying, “Well, you promised, and five or 10 years later you’re backing off on your promise. How will we trust you when you swore to us that something would be confidential and that it wouldn’t be sold in the future or made public?” That’s the concern.
I have looked at historic census records. I find them fascinating. Personally, I would prefer that they were all available 92 years out for everybody and forever, but I worry about the risks to the views of the institution and the willingness to respond.
[Translation]
Senator Mégie: Thank you for your presentations.
Mr. Cappe, I think earlier you said that the Chief Statistician has a lot of power. Does he have that power now, or will he have it after Bill C-36 is passed?
Mr. Cappe: I was referring to the power of the Chief Statistician after Bill C-36 is passed. His power has been more or less effective in the past, as Mr. Sheikh demonstrated. With this bill, the power will be much more significant.
Senator Mégie: The reason I asked you this question is precisely because one of the resigning chief statisticians complained that he felt that Shared Services Canada had invaded his independence. You see the situation from the outside. We have met with the main stakeholders.
Mr. Cappe: It happened after I left the government. I do not completely agree with Mr. Smith or why he resigned.
[English]
Senator Raine: Thank you very much. Actually, you’ve already answered my questions.
I have one more but it doesn’t really relate to the bill. Is there a way for citizens to interact with Statistics Canada if they feel they were missed in one of the censuses? I haven’t come across that and I wasn’t sure.
Mr. McKinnon: During the period of the census, Statistics Canada does a lot of publicity which, among other things, does tell people that if we haven’t been in touch, here are our phone numbers and so on. To my knowledge, if someone said, “You missed me 10 years ago and had I filled it out, here’s what I would have said,” I would be surprised if they had capacity. I’ve never heard of a case like that.
Senator Raine: Thank you.
The Chair: This bill is about increasing the independence of the Chief Statistician. The bill also provides for a number of statistical analyses that come under the direction of the minister to be removed. But it leaves in that the Governor-in-Council shall by order prescribe the questions to be asked in the population census and agricultural census. Do you see any inconsistency or danger in the fact that some of this is being changed to ensure the independence but that part is still under the government’s control?
Mr. McKinnon: Which section is that?
The Chair: It deals with the population census and the agricultural census.
Mr. McKinnon: The wording of those questions has always been approved by cabinet. One of the changes would be — and this does enhance the independence of StatsCan — to have any involvement with change in either methodology or questions made public. So the people would know who made the decision, which was at the heart of the reason Munir Sheikh resigned, because he could not protect the professional integrity of the institution.
Mr. Cappe: I would add that there are questions. For instance, which ethnic boxes are available to be ticked off is indeed a decision that should be approved by government at some point.
The Chair: That’s it. We’ve run out of time. I thank both of you for being here.
We are now on to the second panel, which will run to 6:15 p.m.
On this panel in the room is Philip Cross, Senior Fellow, Macdonald Laurier Institute; and Chad Gaffield, Distinguished University Professor, University of Ottawa. By video, welcome to Bill Waiser, Distinguished Professor Emeritus, University of Saskatchewan.
Gentlemen, I’ll start with Mr. Waiser, if I might, coming in by video conference, just to make sure that everything goes smoothly on the airwaves.
Bill Waiser, Distinguished Professor Emeritus, University of Saskatchewan, as an individual: Thank you. That’s very kind. I’m a W, so I usually go last.
There was an extra column in the 1916 Census. For the first time in Canadian census history, enumerators asked about military service. If someone was overseas, then the enumerator wrote “O” in the military service column. But if the person was still in Canada, the enumerator put down “C” and provided the name of the camp in the residence column. This information might seem insignificant — an “O” here and a “C” there — but it helped provide a more complete understanding and appreciation of Canadian participation in the Great War.
The census data can also be compared against the information individuals provided on their attestation papers. Race or nationality, for example, was not recorded at the time of enlistment, and the census data fills this void. This kind of detail at the individual level is a wonderful gift for the descendants of Canadian soldiers and nursing sisters, especially during the centennial of the Great War.
But such information may not be available in the future. In 2006, for the first time in Canadian history, census participants were asked to indicate by checking a box whether their responses could be made available for research after 92 years. No explanation was provided on the form about the significance of census records to future historical and family research. Nor was anything said about the consequences of saying no. Canadians completing the census had never been asked this opt-in question before. Rather, nominal census information had been made publicly available after a minimum 92 year waiting period, without a single word of complaint. But because of the so-called informed consent question, Canadians now had to agree to make their census information accessible to future generations.
Only 55 per cent said yes in 2006, then 66 per cent in 2011, and finally 81 per cent in 2016. The consent rate may be improving, but in 2016, 1 in 5 Canadians still said no, representing about 7 million Canadians. If that had been the case in 1916, then much of the rich data available today would have been forever closed to the public.
The opt-in question, first introduced in the 2006 Census, was part of a legislative compromise. In the early 21st century, a stalemate had been reached between Statistics Canada and Library and Archives Canada over public access to historic name-specific census data. Statistics Canada insisted that a promise had been made by the Wilfrid Laurier government to keep this information forever confidential. Library and Archives Canada countered that the confidentiality provision applied only to census enumerators and that name-specific census data could be released 92 years after the census had been taken. The Department of Justice agreed that there had been no promise of perpetual confidentiality by the Laurier government.
By 2005, the time had come for a comprehensive solution that would end the stalemate. I provide a footnote that provides the history of this issue if any member of the committee would like to explore that in greater detail.
Under the 2005 compromise, Statistics Canada agreed to release all post-1911, name-specific censuses after the 92-year waiting period. The public response was simply overwhelming. When the 1911 name-specific census became available online, the site averaged 17 hits per second, not per minute, the first few months.
The other part of the compromise was an opt-in question. The opt-in question would be tested for the 2006 and 2011 censuses and then reviewed. This review was not optional but mandated in legislation. Section 2.1 of the 2005 act specifically required a review of the informed consent question no later than two years before the taking of the third census of population — in other words, 2016 — by any committee of the Senate, the House of Commons or both houses of Parliament that may be designated or established for that purpose. Section 2.2 asks for a report on the matter.
At the time, historians, archivists and family researchers were deeply worried that the opt-in question would undermine the integrity of the national census as an inclusive source of information on individual Canadians and their lives.
Canadians exist in history and across time because they are in the census. They also pointed out that neither the American nor the U.K. censuses have an opt-in question. And they warned that it’s impossible today to know what might be historically important tomorrow, and that future Canadians could be deprived of access to family information that might not be otherwise available.
Most important, under the opt-in question, those completing the census would be making the decision for all members of the family. If they said no to the opt-in question, it applied to their children’s data as well. Statistics Canada countered these genuine concerns by promising a vigorous media campaign to get Canadians to say yes to the informed-consent question.
So what happened? No detailed explanation is provided on the census form about the significance of census records for future family research or for understanding Canadian society, nor has there been an extensive media campaign. Nothing is said about the consequences of saying no. Instead, that section begins with the rather ominous warning about the confidentiality of the census and then simply asks respondents whether they want to make their census information available in 92 years for important historical and genealogical research. That is not informed consent as required by the act.
It’s not clear from Bill C-36 whether the opt-in question will continue to be part of the short-form census or the National Household Survey. The draft bill also does not address the mandated review of the opt-in question, a deadline Statistics Canada has missed. That legally required review must be undertaken before another census cycle. It should be open, transparent and comprehensive and result in a formal, documented report to Parliament. Until that review is undertaken and the opt-in clause removed, hundreds of thousands of Canadians will be missing from the historical record. Everyone deserves to be remembered and to have a place in the archives of Canada, especially this Canada 150 year. Now that would be a birthday present. Thank you.
The Chair: Thank you very much, professor. You’ve given me a lot better appreciation of what Senator Griffin has been raising here at committee. Thank you very much for your presentation.
Philip Cross, Senior Fellow, Macdonald Laurier Institute, as an individual: Besides my work with the Macdonald Laurier Institute, I suspect the real reason I am here is my 36 years at Statistics Canada. I will try to bring an insider’s perspective to a lot of these issues.
Most of the commentary surrounding this bill is focused on the yin and yang of independence and accountability at Statistics Canada. This trade-off certainly does exist. No organization can be truly independent unless it controls its own budget, but that would negate any accountability to the taxpayer.
However, based on 36 years of working at StatsCan, I would argue there has been excessive emphasis on the virtues of independence and not enough on the importance of accountability.
To start, it is a mistake to frame the question of StatsCan’s independence solely in relation to external factors. Calls for more independence play into the myth that StatsCan is a passive, defenceless victim of forces beyond its control. StatsCan, to a very large extent, controls its own reputation and destiny, something that is not discussed enough internally or in external forums such as this hearing or in the media. Internal processes established StatsCan’s reputation for accuracy, bipartisanship and respecting confidentiality, without which its relevance and independence would be open to question.
There are many existential threats to Statistics Canada that do not involve its independence. For example, I can think of few things that would compromise StatsCan’s reputation more than a proliferation of the type of high-profile errors we have seen recently, such as in the labour force survey in 2014 and the census in 2017. If ever the public or users lose confidence in the accuracy of StatsCan data, it would take years to regain. As the Dutch saying goes — trust arrives on foot but leaves on horseback.
Political neutrality is even more important. If StatsCan was ever found manipulating its data or analysis for the advantage of one political party over another, the hit to its reputation would be fatal. A leak of confidential data such as tax returns would shake to the core the confidence Canadians have in divulging information to StatsCan.
It is not far-fetched for StatsCan to meddle in policy-making. It already did so when it publicly endorsed evidence-based policy-making, a blatant example of pursuing its self-interest. StatsCan has no business saying how policy should be made any more than what policy should be. If the electorate wants policies based on intuition, tradition, religious beliefs or even phases of the moon, that is irrelevant to StatsCan’s basic function of making available the best possible data. Nor does StatsCan face the external world alone. When faced with budgetary or other challenges, StatsCan has demonstrated considerable skill in forging alliances with other federal government departments and external alliances with a wide range of governments, media, academics and other data users. The full range of the latter was on display in the 2010 controversy. The list of organizations supporting a mandatory census numbered in the hundreds.
The last resort in preserving StatsCan’s integrity and independence is going to the media about attempts to interfere with its operations up to and including the resignation of the Chief Statistician. The latter card has not always been played skillfully, but it remains useful when not devalued by overuse.
There’s another risk to overemphasizing StatsCan’s independence. Some people think that reflexively taking positions opposed to the government of the day proves independence — an error made by the first Parliamentary Budget Officer. Automatically adopting a particular point of view irrespective of the facts reveals a slavish devotion to dogma and doctrine — the very opposite of independence. The Bank of Canada is a good example of how an institution can be independent yet work closely with the government of the day.
Ultimately, complete independence is not possible or even desirable. Even StatsCan does not want the absolute independence some people champion since it would mean losing its seat at the table when senior government officials discuss their plans and priorities.
Before being granted more independence, StatsCan has to demonstrate that it can be trusted to put accountability of the public ahead of its self-interest. Its record is mixed, at best. It regularly curries favour with select media outlets at the expense of others, something no public institution should do since everyone pays taxes and all media should be treated equally regardless of political views.
A recent example of StatsCan manipulating public opinion was asking senior managers to write down their IT complaints and then getting a friendly journalist to make an access-to-information request. This gives the illusion of involuntarily giving up information when in fact self-serving information was being deliberately planted.
This put its own interests ahead of accountability to the public. The Chief Statistician’s attempt in 2016 to intimidate the public by threatening to resign if computer support was not repatriated also put StatsCan’s interests above the public’s as embodied by the government. All these are the actions of an organization that needs to be held to account, not rewarded with more independence.
A potential risk in the future is StatsCan ignoring the proliferation of private sector data, such as last month’s introduction of ADP payroll data, that could lead to fewer surveys. Only by holding it accountable to taxpayers can one be sure that StatsCan will not simply continue to do surveys out of self-interest or habit.
There has been much discussion about the process of appointing a Chief Statistician and not enough about his departmental removal, which in practice has been the more common problem. Thinking back to the last eight Chief Statisticians, clearly some filled the position well and some were mediocre. Largely forgotten is that the organization I joined in the late 1970s was in tatters, subject to investigations and overhauls of its management practices to rebuild its reputation. The challenge that has proven difficult, if not impossible, to predict ahead of time is who will make a good Chief Statistician because the circumstances and skills required can change rapidly. Making a brief appearance in front of house and Senate committees is unlikely to produce better results.
Inevitably, a Chief Statistician will be appointed whose skills are not those required during his or her tenure. This is not necessarily of great concern, especially with a term limit of five years. Plenty of organizations in Ottawa have to endure similar or longer tenures of poor leadership; only those with weak internal cultures, which certainly does not describe StatsCan, suffer lasting damage.
In fact, organizations often emerge during periods of stress, which can be positive or negative, stronger and more focused. I can think of four instances where this was true at StatsCan, in my experience: the 1970s turmoil, the budget cuts from cancelling and then reinstating the 1986 Census, the 1990s expansion of provincial economic statistics, and the transition to a new Chief Statistician amid the upheaval surrounding the 2011 Census. All were traumatic to live through; all ultimately strengthened the organization.
The Chair: Thank you. Our next presenter is Chad Gaffield.
[Translation]
Chad Gaffield, Distinguished University Professor, University of Ottawa, as an individual: Good evening. Thank you for inviting me to appear before you and contribute to your deliberations.
[English]
I come here today as a professional historian who is motivated primarily by the importance of seeking to make a better future. Over the years, my career has certainly been enabled by the kinds of evidence that Statistics Canada has collected over the decades — in fact, centuries — but I have also had occasion to try to use that in terms of contributing to public policy debate.
I am not alone, as my colleague Bill Waiser illustrates, as well. There is an increasing number of scholars who have been systematically studying Statistics Canada evidence. That is probably one of the key aspects of the blossoming of a Canadian research culture in the last 50 years.
It is often remarked that the census has characteristically been Canada’s largest peacetime operation. I suppose it is not surprising, therefore, that the results of that operation are now considered a major source for advancing knowledge and understanding of the past. That makes sense. In fact — and I think Bill was alluding to this — it is more likely that evidence about a person in Canada at any time since the mid-19th century would be available in Statistics Canada records than in any other type of historical document. It is quite a democratic kind of evidence base.
It is not 100 per cent, and we all know that. But relatively speaking, it is the most comprehensive.
The reality is so important, because we have been learning increasingly about the need to do justice to the historical role of all those in the past, not just the official and non-official elites. Documents that Statistics Canada creates help us, not only in the past but also in the present.
The other thing — and I was alluding to this — in terms of their focus, documents like the census ask the same questions of everyone. It doesn’t discriminate based on who you are; it asks the same questions. This evidence, like all historical evidence, must be used critically and appropriately, but it is exceedingly important to our knowledge of the past and present, and the horizon of possibilities before us.
My take on this is, as I mentioned, as a historian, but over the years I have had a chance to contribute to important public policy questions. In the early 1990s, I was part of a group called the Data Liberation Army that mounted an initiative called the Data Liberation Initiative that had as its goal to get into our schools and university campuses across the country evidence about Canada and Canadians that had been gathered by Statistics Canada. It transformed many of our courses from being quite colonial in that the data was from elsewhere that was being studied to data about ourselves, about Canada. That goes on today.
Then, in the later 1990s, I was part of the Expert Panel on Access to Historical Census Records that was appointed by Minister Manley to focus on the question of the 92-year rule. Our conclusion out of that expert panel of independent people, which was also our fundamental recommendation, was that census records should be publicly released through the national archives 92 years after the census is taken. We did a thorough examination of the history of that and where it might go.
More recently, I had the occasion to lead a major research initiative involving the construction of the Canadian Century Research Infrastructure to cover the first half of the 20th century and therefore enable research in unprecedented ways on the making of modern Canada. This was done in collaboration with Statistics Canada and universities across the country. All of that has enriched immeasurably our knowledge of ourselves — a true domestic understanding of who we are here in Canada.
I also served between 2006 and 2014 as President and CEO of the Social Sciences and Humanities Research Council of Canada. I had occasion to deal with this dossier, among many others, and also see the value of the kind of evidence that Statistics Canada produces for research across all sorts of fields, and in very innovative and helpful ways.
My main message today is to emphasize that Statistics Canada has historically been and continues to be supported by Canadians as contributing significantly not only to knowing ourselves as Canadians but also to decision making in government, businesses, institutions, including religious institutions, and communities, both official and informal, as well as to international concepts and methods in national statistics agencies.
The historical record shows, I think, that the foundation of these achievements has been an informal covenant between Statistics Canada and the Canadian people. This covenant has been based on trust, authenticity, credibility and legitimacy. The result has been, for example, that Canadians fill out census forms, answer the various questions as best they can, and quite consistently treat the whole process with respect and rigour. In turn, Statistics Canada has both ensured that all interested parties can benefit from the collected enumeration forms while also implementing effective methods for protecting confidentiality and the full integrity of the administration of the census.
This is a good-news story, and it is one of the contributors to the fact that, relatively speaking, Canadian society is one of the most desirable places in the world to live, and it is often considered one of the most successful societies. We have a world of work to do, and all historians will emphasize that — certainly I will; however, relatively speaking, most of us feel we are really lucky to be here.
Happily, the proposed legislative changes strengthen the underpinning of the covenant between Statistics Canada and the Canadian people by formalizing commitment to various good practices that Statistics Canada has developed over the years and that continue to be enhanced in keeping with the changing times.
It is essential that we get real clarity on the 92-year rule, because I don’t think it is helpful to continue the ambiguity and confusion that has been introduced.
If it is approved, I think the proposed legislation, if we can get good clarity there, would be hailed as a major step forward and would be cited internationally, in keeping with the world’s admiration for Canada’s continuing leadership in learning about society in order to help make a better future.
At the same time, I can well understand why various witnesses have suggested ways in which the legislation could be further improved. I appreciate your attention to these suggestions. You may not be surprised to know that I have thoughts about each and every one of them, frankly, and a lot of them based on historical examples. I would be pleased to discuss them with you, ranging from the selection and appointment of the Chief Statistician to the description of qualifications, the use of the UN guidelines, how the actual enumerations are done and things like censuses and so on. We can pursue them if you would like.
I want to close by emphasizing my conviction that nothing should be done at this point to threaten the passing of the proposed legislation — again with that specific clarity about the 92-year rule. As we all know, perfection should never become the enemy of the good.
Thank you again for your good work. I look forward to addressing your comments and questions.
The Chair: Thank you, Mr. Gaffield, and thank you all for your presentations.
Given the size of the list, maybe we can start with the possibility of two questions per person. Given that we have two witnesses here and one on video, I would ask that you indicate whom you want to ask the question of, and if anyone else wants to come in after that first person has answered it, you may do so.
We’ll start with the sponsor of the bill, Senator Cordy.
Senator Cordy: Thank you to each of you. Your presentations were very lively and very helpful.
Mr. Waiser, clause 18.1(1) “Census - disclosure after 92 years” says that when this bill passes there will be no tick box on the census form and that the information will be disclosed after 92 years. I think that is a positive thing.
The concern, of course — and I will leave this line of questioning to Senator Griffin — is what do we do with the information gathered in 2006, 2016 and 2011? Is it gone forever? Senator Griffin asks that at every panel, so I will leave that to her.
Mr. Waiser: Senator Cordy, the opt-in question will no longer appear in any future census, short form or detailed household survey?
Senator Cordy: That is correct. That is clause 18.1(1).
I think you made your point clearly about the difficulty that has caused historians and interested people in gathering that information for those lost years. I am not sure how we can get that information back because as someone said earlier, StatsCan can’t go back on their word and disclose information.
Mr. Waiser: No.
Senator Cordy: Anyone can answer this next question.
I think you made your point very well about the integrity of Statistics Canada and that we can’t do anything to change that. I think Canadians have a trust that statistics coming from Statistics Canada are valid and are not political; that is, the information they are getting is accurate. However, that trust can be removed quickly. I thought your comment that trust arrives on foot but leaves on horseback is very true.
Will this bill strengthen Statistics Canada? Is it a better place for us to be if this bill passes?
Mr. Cross: First, there are two separate issues there. Integrity and bipartisanship are separate things. Accuracy of the numbers and whether the numbers have a political motivation are slightly different things.
It depends on the numbers you look at. Some numbers are extremely accurate from Statistics Canada; other numbers drive analysts crazy. The variability of especially the employment numbers, the trades, exports and imports can be revised all over the place. There can be big misses there. There is a real problem with how the statistical system handles the energy sector.
With other areas like the CPI, there is never a revision. There is a high degree of accuracy. There better be because the whole tax and transfer system is indexed to the CPI.
Another point I would make is that a lot of the discussion focuses on the census. I came from economic statistics. I worried about CPI, unemployment and GDP. This is our bread and butter on the economic side. We didn’t use census at all. I find the whole discussion about census sucks all the oxygen outside of our discussion of Statistics Canada. As I say, there is a whole other side of Statistics Canada — I didn’t work with census ever in my 36 years there, or hardly ever. So we should be aware that we don’t want to mix up the two.
We had a bemused view of this when this controversy was going on in 2010. There was this riot going on in this highly visible, important part of the organization, and it didn’t affect us. We went on doing our days. Something interesting was going on over there, but it didn’t directly affect us. I would just ask people to remember that there is a huge amount of Statistics Canada that is beyond the census, and we shouldn’t lose sight of that.
Does this strengthen the integrity? When it comes to the accuracy of the numbers, I don’t think it really has any impact. The bipartisanship might strengthen it a bit, but StatsCan has a strong internal culture for these things.
Mr. Gaffield: I can weigh in slightly. I think it is unnerving maybe to public discussion when the implication can be given that that covenant that had not been centrally questioned for decades and decades all of a sudden is put up as something to debate.
I would think that this bill is very helpful in terms of the transparency, the reporting, basically committing back to the public that there will be a very overt discussion of what happened and why. I would link that as well to the value of the 92-year rule because basically it says, yes, we will have confidentiality, and so on, but in 92 years the bright lights will be put on everything we did and there will be an accounting at some level. It will be 92 years later, but we won’t be able to fully escape. I think that is important. I think the 92-year rule seems to have worked very well.
My third point is that everything related to this is a human activity. Statistics are human creations. They change. I’m a director on the board of directors of the Canadian Statistical Sciences Institute, and we are well aware that statistics change over time. They are human creations. We are obviously trying to do better, and we are learning more about it, but I do not believe for a minute in the notion of anything being pure and so on. We are human beings. This is our creation. So I don’t think we need to rehearse old myths but rather say that what we are always attempting to do is to do things better and better. I think, senator, this is a step in that direction.
The Chair: Mr. Waiser, do you want to weigh in on this second question?
Mr. Waiser: Just one quick follow-up comment. Dr. Cross talked about Statistics Canada providing the best possible data. The removal of the opt-in question will dramatically help this and help restore part of that tarnished reputation of Statistics Canada over the administration of the census.
Senator Frum: Mr. Cross, I will start with you. The purpose of the bill is to increase independence of the Chief Statistician. You have made it clear that in your diagnosis of what might be where Statistics Canada could improve, that has not been the issue you are concerned about.
Let us say it was. Do you see ways in which Bill C-36 actually does improve the independence in any meaningful way?
Mr. Cross: A lot depends on how it is implemented and interpreted. For example, Professor McKinnon in previous testimony talked about how there will be an annual review of the National Statistics Council regarding how the statistical system was performing. That sounds laudable, but I was sitting there wondering how the National Statistics Council will do this. They strongly risk being captured by Statistics Canada in that. How would they ever be able to evaluate it independently of what Statistics Canada told them? I don’t know how they could. It’s something, obviously, we were very interested in.
It’s a question I would ask. Every year there would be a major revision to the GDP. It seemed that there was once a year opportunity where you could ask the question: How is the statistical system performing? I would ask it every year of the director general and we would have a really intense, interesting — but I don’t know how anybody outside could ever access that information.
So I worry a little bit that, yes, we seem to be providing more oversight to the National Statistics Council and that might strengthen its independence, but I am really worried that the National Statistics Council will just be fed a story by Statistics Canada and it will be captured by it.
When, for example, Statistics Canada’s budget is increasing, everything is great. When Statistics Canada’s budget is being cut, oh my God, the statistical system is under great strain and so on. It would be very interesting to see if the NSC can come up with a narrative independent of what Statistics Canada is telling them.
Senator Frum: You started to touch on a theme from the last panel as well about this idea — I think the mythology around the objectivity of collecting statistical data on people. I think you touched on it a moment ago again, that is, how pure a science this is and the idea that you can divorce methodology from politics.
Mr. Gaffield: First of all, I would say there is no pure science. Chemistry is not a pure science. I think we have all understood that what we do is a human activity.
That being said, there is a world of difference between a rigorous, systematically peer-reviewed human activity and something that is a belief, a perspective, I think, rather than an evidence-based interpretation. I am devoted to the importance of rigorous, scholarly and scientific methods, but I would hope we all recognize that the way in which we do that now is not the way we did it in 1800 or 1900 and so on.
I think all of us believe we are striving toward an increasingly sophisticated and appropriate methodology in all our work.
Senator Seidman: Thank you both very much for your presentations. If I might address my question to you, Mr. Cross.
You said in your presentation to us that there are many existential threats to StatsCan. One you spoke of, if users lose confidence in the accuracy of StatsCan data. But you went on to say that even more important is political neutrality, and if StatsCan was ever found manipulating its data or analysis to the advantage of one political party over another, the hit to its reputation would be fatal. I agree with you, of course.
So my question to you is, do you think Bill C-36 could do more to protect StatsCan from political interference?
Mr. Cross: No. I don’t really worry too much about the protection from political interference because that is largely a matter of internal culture. The internal culture of Statistics Canada on this is extremely powerful. I just can’t imagine it, even without the changes in this bill.
When I arrived at Statistics Canada, I would ask my bosses, “What would you do if the government tried to impose a result, saying, ‘We want this result for the CPI or that result for the unemployment rate’?” They would just say, “We would call a press conference and that would be the end of the government.”
That’s the other thing. The organization is aware. It has tools that it can use. The idea that it needs an external big brother to come along and protect it from the bullies — no, it knows how to pump iron pretty well itself.
Senator Seidman: If I might ask you about the Chief Statistician and criteria around the selection of the Chief Statistician. I know you are not so interested in that particular aspect, but the OECD, when they testified before the House of Commons Industry Committee, said that the selection process for the Chief Statistician should be described clearly by law to specify that the appointment should be based on professional competence only.
Do you think that that specificity would be a good thing? Would it have any impact on external threats to StatsCan’s political neutrality?
Mr. Cross: No, I don’t. I don’t think it has any big difference. For example, I think back to Statistics Canada in the 1970s. It was a mess. A disaster. Ivan Fellegi, in his testimony to the house, talked about that. Anybody who lived through that was aware. Statistics Canada has a very good reputation today; it did not have a good reputation in the 1970s. It was one of the worst-performing departments in the government.
That was not because of the selection process of the Chief Statistician. One of the Chief Statisticians in the 1970s was Sylvia Ostry. She was a well-known, well-connected academic. Everybody would have looked at her and said, “This should be a great Chief Statistician.” She was just wrong for the times.
There were all kinds of problems for the agency at that time. One was this very high rate of inflation. Could any Chief Statistician have performed better given that the organization was confronting a challenge it hadn’t faced? We had low inflation in the 1950s and 1960s, early 1970s, and suddenly everything is up 10 per cent, 12 per cent, way beyond anything you had ever published before and you go, “Whoa, that seems to be a lot,” and it turns out it should have been 14 or 15 per cent.
It was just this different world that the statistical system had a lot of trouble adjusting to.
We shouldn’t expect Chief Statisticians to have a huge impact. An individual only controls so much. A Chief Statistician — I saw this with Munir when he arrived. He came in and didn’t try to change the place left and right. He said, “I’m going to focus on one or two things.” One was reducing the error rate. He thought 10 per cent of everything we were putting out in the daily was wrong. That is insane. So he started every meeting with what are we doing about quality control to reduce that error rate? He focused on that, and within six months he had the error rate below 1 per cent.
But that’s it. He focused. You can only change one or two things; if you come in and try to change everything, you’re not going to get anything done.
Senator Petitclerc: Thank you very much for your presentations. I want to address my question to Mr. Gaffield. All the witnesses have made it very clear how important and relevant what we get — and I think in the end, the end priority is what we get as a society, as a country, and historians and professors are in a good place to tell us the value of all that data we attempt collect.
We understood that the Chief Statistician mainly is about the how, how we get that data in terms of methodology, and the what mainly goes to the minister.
So I am interested in, from your perspective and what you are trying to get as data, how do you feel this bill answers that balance on the what and the how when it comes to independence and the quality of data that we need to collect as a country?
Mr. Gaffield: Thank you very much for that question. I think it is really important. I am an historian who studies change looking from the bottom up, as we say, and the top down. Very complex. But I do have a deep appreciation of leadership, and I think it does make a difference who the Chief Statistician is. We have had examples where Chief Statisticians have made a huge difference.
I think in this bill, what I like relates to my own experience. When I was appointed as President of the Social Sciences and Humanities Research Council of Canada, it was a GIC appointment. It was very similar in the sense that my mission was to have total independent responsibility for which research grant applications got funded, the operations of all of that. But my mandate in terms of why I was funding research and what I was trying to do and so on, that was on the side of the government; they held that responsibility. So I think this move actually does strengthen that for Statistics Canada in really important ways.
The other thing Mel Cappe alluded to is a GIC appointment makes it clear that this positioin is not happening in a particular government department but rather as a whole-of-government. That’s the link back to the covenant with Canadian society, Canadian people. It’s a common good. That reinforces that in really important ways.
Certainly when I was appointed at SSHRC, I don’t think I was appointed only because I was seen as a great scholar. Somebody had to think that I had that foundation. But it’s a big, complex organization. For example, cultivating this culture that StatsCan has, that’s a real leadership challenge. There are all sorts of leadership-management skills that go along with that, and getting the right person to do that takes a lot of judgment.
The good news is Canadian people are going to be the judges of that. That was alluded to a bit earlier. I think the pressure to get the right person who has those combinations and so on is really key. I was pleased with how that was presented in this document.
The Chair: Professor Waiser, did you want to come in on this?
Mr. Waiser: I would just like to echo what Dr. Gaffield said about the importance of trust or covenant. For example, doing the census should not be something that’s onerous; it should be something that’s positive, for example, be a part of history. We need to cultivate that relationship and strengthen it between Statistics Canada and the general Canadian population.
The Chair: Senator Griffin, the 92-year question is coming up.
Senator Griffin: My question is in relation to the data gap, of course, and the 92 years when it wasn’t chosen by a number of people to have their individual data released. What impact is this going to have on Aboriginal communities? In particular, I’m thinking of somebody who might want to some day be trying to regain or gain their status under the Indian Act.
Mr. Waiser: That’s a very good question, and I don’t know. When you give your census information, you are promised that it would be kept confidential. It’s not destroyed, and that’s important. The data that they submitted is not destroyed, but they have been given that promise of confidentiality by saying, “No, you cannot release that information.” So I don’t know what the answer is.
Senator Griffin: Does anyone else?
Mr. Waiser: But the good news is that the data is there.
The Chair: Do either of you want to comment on that?
Mr. Gaffield: I think it’s really unfortunate in a variety of ways. The little good news about it all is that it’s going to prove to be a cautionary tale. Historians, urban planners, church leaders, on and on will always be able to be reminded not to take this for granted. In fact, if we do want to work together to build a better future, we have to be attentive to the kind of implications that decisions like that might have.
Senator Raine: Thank you very much, all of you. My question is for Mr. Cross.
I was a little bit concerned when you mentioned a recent example of StatsCan manipulating public opinion with regard to their IT complaints. It reminded me of the testimony that we had from the folks from Shared Services Canada. Of course, we all know what’s happening with Phoenix.
I guess really the end question would be how should the quality of information technology services offered to StatsCan by Shared Services Canada be assessed?
Mr. Cross: I’m not really comfortable answering this. All this with Shared Services happened after I left, so I’ve heard some stories from StatsCan, but it’s not something I have any first-hand knowledge of.
How can it be assessed? My biggest concern would be the confidentiality. That was mentioned by the previous presenters. I would not worry about the confidentiality. Statistics Canada has access to lots of tax records and confidential data. I would not worry about the confidentiality. Again, that goes back to the internal culture being extremely powerful in that area.
Whether Shared Services would share that to the same degree as StatsCan, I’m not sure. But I would trust that Statistics Canada would be involved in putting in checks that would preserve it. As I say, Statistics Canada knows that irrespective of who makes the mistake, if a mistake is made, Statistics Canada is going to wear it in their relations with the public, so they’ll be strongly motivated to make sure confidentiality is preserved. That’s something I’d keep an eye on.
Senator Raine: We’re going through a transition obviously where StatsCan had control of all of the data technically and with the hardware and software. Now they’re depending on an outside agency. The legislation doesn’t impact that at all, but that would be something that the Chief Statistician would really have to be aware of.
Mr. Cross: We talked about Chief Statisticians, and it depends on the circumstances. Some people arrived with a lot of credentials and they were just the wrong person at the wrong time.
The previous Chief Statistician did a very good job. This was an organization in turmoil when I left. He did a very good job of settling everybody down, saying, “Focus on your job, get back to your desk, let’s do our job.” That’s what the organization needed at that time.
What it needs now is somebody who could work with another department. This Chief Statistician originally came from Statistics Canada. He went to work in other departments. He now comes back, so he probably has more skills in working with other departments than the previous Chief Statistician. So hopefully that will work out well. But that is clearly his number one challenge.
The Chair: Professor Waiser, do you want to weigh in on this?
Mr. Waiser: No, I don’t.
Mr. Gaffield: Senator Raine, how to take all the good news out of digital technologies and avoid the bad news is a problem we’re living in our homes and everywhere. It’s an amazing challenge that we haven’t mastered, and I don’t think we’re going to master it any time soon. It is really a journey. We are learning more and more and getting better and better. My colleague was alluding to this. We have to be really attentive to this, keep really focused on it and not underestimate the damage that can result in the same way in terms of our own issues and our own lives.
The Chair: Moving into round two.
Senator Frum: Mr. Gaffield, I just want to clarify the selection process. You’re saying you do support some kind of process. Forgive me. Could you elaborate more on how you would see that working?
Mr. Gaffield: My sense is that the bill appropriately doesn’t go too far. What are the very specific criteria for the job?
I think we’ve all said that it’s a very complex constellation, and we have to set the conditions for success in terms of the selection committee and so on. At the end of the day, I think it will be judged by the Canadian people as part of this covenant. That’s a real guarantee at some level that this will be treated with all the care and done in keeping with the good aspects of an institution that we’re really trying to continue at the highest level.
Senator Frum: But given that whoever the successful candidate is serves on good behaviour and not at pleasure, I suppose that’s where I see that maybe there’s a need for some balance, because once in place, that person is there for five years unless they do something spectacularly wrong. Is that right?
Mr. Gaffield: I agree. At the Social Sciences and Humanities Research Council, I felt that sometimes we were able to fund things that maybe were not seen by all quarters to be the best thing. I used to use the example that through the 1990s, the research council funded a lot of work on 14th century Islamic culture and so on. Some people said, “Why is Canadian society doing that?” Well, pretty soon, after 9/11, it became essential in terms of helping us understand what we were reading, hearing and trying to come to grips with.
I think it is very important that someone gets selected, and in five years if it doesn’t work out, well, they can be thanked. I totally get that. But I think that on the way, on the fly, their competency should really be the test.
The Chair: Do any of the other panellists have any further comment on that question? Seeing none, I will draw the meeting to a close. Thank you to the three of you. You’ve helped, with your testimony, to inform us well as we get prepared to deal tomorrow with Bill C-36 at 10:30 a.m.
We’re going to have the Chief Statistician of the United Kingdom on video conference tomorrow, so we’ll get an international perspective. That will be in the first half of tomorrow’s meeting. In the second half of tomorrow’s meeting we will deal with Bill C-36 clause by clause and then report it to the Senate.
(The committee adjourned.)