Skip to main content Start of content

INDU Committee Meeting

Notices of Meeting include information about the subject matter to be examined by the committee and date, time and place of the meeting, as well as a list of any witnesses scheduled to appear. The Evidence is the edited and revised transcript of what is said before a committee. The Minutes of Proceedings are the official record of the business conducted by the committee at a sitting.

For an advanced search, use Publication Search tool.

If you have any questions or comments regarding the accessibility of this publication, please contact us at accessible@parl.gc.ca.

Previous day publication Next day publication
Skip to Document Navigation Skip to Document Content






Standing Committee on Industry, Science and Technology


NUMBER 137 
l
1st SESSION 
l
42nd PARLIAMENT 

EVIDENCE

Wednesday, November 7, 2018

[Recorded by Electronic Apparatus]

  (1620)  

[English]

    Before we get to our guests for the day, I want to do a couple of quick housekeeping matters.
    Because of the fall economic update, there will be no meeting on the 21st. We had actually scheduled either the 19th or the 21st for supplementaries, so that gives us no choice but to do the supplementaries on the 19th. We've extended an invitation to the minister.
    We have our constituency week next week, so on the 19th we'll have the minister. We're still trying to finalize on whether it will be the first hour or second hour; I'm not sure. There will be no meeting on the 21st, and then the 26th will be fun and exciting; we will have Google, YouTube, Facebook and Spotify.
    That will be fun.
    That will be an exciting day.
    That's the 26th?
    Yes, the 26th.
     Are there any questions on that? No? Good.
    Let's focus here. Today we have a briefing with the chief statistician of Canada, Anil Arora; André Loranger, assistant chief statistician, economic statistics; and Linda Howatson-Leo, director, office of privacy management and information coordination.
    Normally, I'm pretty easy with the time, but I just want to make sure that you are paying attention, because I will follow the time sharply. I don't want to have to cut people off, but I want to make sure that everybody gets the chance to ask whatever questions they would like to ask.
    Having said that, Mr. Arora, you have seven minutes to enthrall us.

[Translation]

    Good morning, Mr. Chair, Mr. Deputy Chair and committee members.
    Thank you for this opportunity to share with you information on Statistics Canada's pilot project on using financial transaction records to provide timely and quality data on our economy and society.

[English]

     Before I begin my remarks, I would like to immediately dispel three major inaccuracies of the pilot project to enhance the statistical system using payments data. First, no data has been collected by Statistics Canada as it pertains to this pilot project. I repeat, nothing has been collected. Second, trust is the foundation of how Statistics Canada operates, and we will continue to earn the trust of Canadians. Third, I can assure you that this project will not proceed until we have addressed the privacy concerns of Canadians and the Privacy Commissioner has done his work.
    As you may be aware, Statistics Canada, like many national statistical organizations around the world, is undertaking a comprehensive modernization effort. This modernization effort will redefine how we gather and deliver data by using leading-edge methods, by leveraging existing administrative sources, by excelling at our core competencies of data integration, e-collection and big data processing. Continuous evolution and innovation has made the agency a world leader in the field of statistics.
    In all of this, of course, I want to underline that Statistics Canada respects the rightful privacy of Canadians and has always devoted itself to doing so. We understand and respect the concerns being expressed by Canadians about accessing their personal information.
    The modernization of Statistics Canada began in earnest in the summer of 2017, when the vision for modernizing the organization was publicly announced. This was followed by a budget 2018 announcement of funding in the amount of $51.3 million to support Statistics Canada's modernization.
    I want to stress that the issue in front of us today is not simply an academic one. Statistics have far-reaching implications for all Canadians. Diminished quality will have direct impacts on Canadians.
    For example, estimates of household spending are used in part to drive the consumer price index. The CPI is in turn used to index pensions and old age security, directly impacting the income of seniors. It is also used to help establish wage rates and labour contracts, employment insurance and policies designed to address things like poverty. Provinces and territories also depend on quality data to apportion HST revenue that funds necessary public sector services such as health care and infrastructure.
    The allocation of these funds is in large part determined by the level of household spending on taxable goods in each province and territory. The Bank of Canada uses our statistics to set interest rates and monitor inflation.
    The methods by which Statistics Canada traditionally collects data from Canadians are falling short of what Canadians demand of us today. Home phones have been replaced by smart phones, taxis compete with ride-sharing apps and withdrawal and deposit slips that used to be filled out at banks have been replaced with online financial transactions.
    The pace at which Canadians are adopting digital services has accelerated rapidly. Today we email money, and we get our food delivered through the use of an app. Eighty per cent of all financial transactions are done electronically, some 21 billion of them in 2016 alone.
    As a matter of fact, following the global financial crisis in 2008-09, there has been an increased demand for more timely and detailed data to better understand how income, wealth and consumption are distributed in Canada, which segments are more vulnerable, and how resilient these groups are to changing economic conditions.
    The Governor of the Bank of Canada, Stephen Poloz, said recently:
We know that cross-border supply chains have complicated the task of gathering accurate data on trade. Digital technologies are making it even easier to fragment production globally. And digital ordering, payments and service delivery are making it easier for transactions to occur that fall below customs reporting thresholds or are missed altogether.
    In today's digital world, Canadians can order goods at any time from anywhere, so a simple question such as, “What did you spend on clothing?” becomes very complex to answer.
    As household purchases become increasingly complex and the volume of transactions multiplies, the burden on citizens to explain, track and report these activities and transactions via surveys is becoming unsustainable.
    Over the last year, we've been actively engaging Canadians about the type of data they need from Statistics Canada. We've consulted Canadians at 176 different sessions over the past year, and they told us there's an overwhelming need and demand for the work Statistics Canada undertakes. In fact, we recently completed a week-long national consultation in every corner of this country, with a full range of stakeholders.

  (1625)  

    Users of Statistics Canada were very clear. Canadians want more data from us, not less. They're seeking data at the city and neighbourhood levels. They expect Statistics Canada to be able to tell them whether single parents, senior citizens and low-income households in their cities have the necessary resources to meet their basic needs for shelter at a time of increasing interest rates. Businesses need better data on consumer spending patterns to grow and serve their customers.
    While our traditional methods are becoming more challenging, fortunately the information we require to develop precise income and spending measures exists in administrative records.
    In that context, I'd like to move on to the second part of my remarks, to briefly describe the pilot project and outline our discussions with the Canadian Bankers Association and financial institutions over the course of this last year. Let me be clear. This is a pilot. It was still under discussion. It has not been implemented yet and no data has been received by Statistics Canada.
    Early in 2018, Statistics Canada initiated a pilot project to determine if the financial information held by financial institutions could be used to help address data quality concerns and data gaps. An important part of this work was to determine if the digital information captured by the payment system had statistical value, and if it could be used to address these emerging data gaps while protecting privacy and confidentiality.

[Translation]

    Statistics Canada met and corresponded with the Canadian Bankers Association and financial institutions on multiple occasions—12 times since April 2018—to outline a high-level pilot project and determine the conditions under which the required data could be obtained.

[English]

    Statistics Canada, the CBA and the financial institutions have been committed to a process that would protect the privacy of Canadians right from the onset.
    The project is designed to follow rigorous methodological principles and privacy by design elements. Selected households would be assigned an anonymized statistical number developed by Statistics Canada. The bank would then be asked to go through its payment information and extract records from a statistical sample of selected dwellings or addresses.
    The current design proposes that the institution create two files. One file contains the anonymized statistical number and the personal information, and a second file would contain the anonymized statistical number and the financial information without the personal information. The current design proposes that these two separate files would be transferred to Statistics Canada.
    Statistics Canada would process the two files separately. Once the demographic information, such as type of household or age of the head of the household, is added, the personal information received from the banks would be deleted. We would take the second file containing the financial transaction data and code them to expenditure categories. We would then join the household demographic information with the coded financial transaction data using the anonymized statistical number I talked about.
    Let me be clear. It would be impossible to associate the financial records and transactions with a given individual or household from this joined file. We have been clear on the need to be fully transparent with Canadians that information was to be provided to Statistics Canada, and we've explained the reason for doing so. We asked that the banks inform their customers that Statistics Canada was requesting this information in August of this year.
    While the notion of data for 500,000 addresses may seem large, there are over 14 million households in Canada. The chance that a given address is selected as part of our sample is one in 28. The chance that the dwelling is used in the actual sample is one in 40. The long-form census by comparison has a one in four sample. We will rotate this sample from year to year so that a history of information for any one household is not possible.
    The sophisticated design has been guided by the very helpful input of the Office of the Privacy Commissioner.
    To put this in perspective, in 2016, as I said, the Canadian payment system processed over 21 billion transactions. Our pilot project sample proposes to access less than 2% of these transactions, each of which has been anonymized and stripped of any personal identifiers.
    My third and final point is around the current status of the pilot, and next steps.
    Mr. Chair, as I said at the outset, Statistics Canada takes the privacy of Canadians very seriously. We have a strong record and reputation on privacy and we understand the concerns of Canadians. Statistics Canada has worked hard to build the trust of Canadians for 100 years, as they provide us with some of their most personal information.

  (1630)  

[Translation]

    We have heard the concerns of the Canadian Bankers Association and the banks and the Office of the Privacy Commissioner as well as parliamentarians, Canadians and Quebeckers.

[English]

     The result of the investigation of the complaints received by the Office of the Privacy Commissioner will further guide the design of this project. We continue to engage with the CBA and financial institutions, and their advice will further help strengthen the privacy protections of this project.
    I can assure you that we will not proceed with this project until we have addressed the privacy concerns expressed by Canadians.
    Thank you, Mr. Chair.
    Thank you very much.
    We're going right into questions with Mrs. Caesar-Chavannes.
    You have seven minutes, please.
    Thank you very much.
    Canadians, many in my riding of Whitby, are concerned about their personal data and want to ensure that it's secure and protected. Canadians have also trusted Statistics Canada for years to use their data to improve lives. I thank you for your service.
    I have many questions. They are not meant to be offensive or to offend. Many come from my constituents. I hope you can keep your answers as brief as possible.
    Can you tell me if I got my coffee this morning from Tim Hortons or Starbucks?
    No, I can't.
    Do you want to monitor Canadians' every move?
    No.
    Is this data project a surveillance scheme in disguise?
    No.
    Are you a spy?
    No.
    Could the government spy on Canadians using Statistics Canada data?
    No.
    Could the Liberal government spy on Canadians using Statistics Canada data?
    No.
    Can Statistics Canada be compelled to share identifiable personal information with the government?
    No.
    Can Statistics Canada be compelled to share identifiable personal information with the Liberal government?
    No.
    Can Statistics Canada be compelled to share identifiable personal information with a member of the opposition?
    No.
    Can Statistics Canada be compelled to share identifiable personal information with any politician?
    No.
    Can Statistics Canada be compelled to share identifiable personal information with the courts?
    No.
    Can Statistics Canada be compelled to share identifiable personal information with the RCMP?
    No.
    Can Statistics Canada be compelled to share identifiable personal information with the Canada Revenue Agency?

  (1635)  

    No.
    Can Statistics Canada be compelled to share identifiable personal information with CSIS?
    No.
    How many data files have been hacked from Statistics Canada databases?
    None.
    When has Statistics Canada lost data in transfer?
    Never.
    Why is Statistics Canada requesting household spending and income data from our financial institutions?
    I can identify four reasons.
    One is the declining response rates from our major source of survey, the Canadian household spending survey. It's sitting at about 40% today. It just isn't giving us the timely and detailed data that we need. We are seeing gaps because of the consumption of digital services in Canada. That is a real gap for us. Our modernization agenda is all about experimentation and piloting, of which this is one example, where we're looking for new ways to fill those gaps. We've demonstrated that when we do use administrative data, as we've done with our housing gap that we've had, we can actually provide good-quality and timely data to Canadians.
    How long have you been using administrative data?
    The use of administrative data in Statistics Canada goes back to 1921, when we first started using vital statistics. Then it was supplemented in 1938 with trade data and so on.
    You've been using surveys for quite some time. Why are you switching over to this type of data?
    As I said, with our traditional surveys we're seeing declining response rates. Canadians are busy. Canadians in many cases are not even reachable. They don't have land lines. There are various reasons why we're seeing the kinds of declining response rates. We have to go to administrative sources. Again, using administrative sources is not anything new to Statistics Canada.
    How will Canadians benefit from this initiative?
    Canadians will have better quality data. They'll have it for the areas that they're interested in. Canadians don't want to just know what's going on at the national level or the provincial level or even at the municipal level; they want to know what's going on at the neighbourhood level. Obviously, this is all with the protection of privacy and confidentiality that they deserve as well.
    Will the data collected be given to the government in any form?
    It will only be in aggregate form, where no individual transaction or data can ever be identified.
    Will the data collected be given to the opposition in any form?
    It will only be in the aggregate form, where no one's information can be identified.
     Has Statistics Canada been asked to spy on Canadians?
    No.
    Who asked Statistics Canada to do this project?
    Well, as I said, using innovative ways, new ways, to get at data is nothing new to Statistics Canada. We're a world leader when it comes to this. When we hear from Canadians that our data are falling short of their needs, we look for new ways to get at that information—again, without ever compromising the privacy and confidentiality of Canadians.
    Is it correct to say that this was internally driven from Statistics Canada?
    That is correct.
    Did the Government of Canada ask Statistics Canada to start this project?
    No, not any specific project, per se; I think the Government of Canada said that they have a desire to increase the quality and the timeliness of data, and I think there's been support, as I just mentioned, for the modernization of Statistics Canada.
    Can this data that has been collected be used by the government in any way to spy on Canadians?
    No.
    Can the Liberal government use this data in any way to spy on Canadians?
    No.
    Can the opposition members use this data in any way to spy on Canadians?
    No.
    Can the government use the data collected in any way to see what specifically Canadians are doing with their money?
    No.
    Can the Liberal government use the data collected in any way to see what Canadians are doing with their money?
    No.
    Can the opposition use the data collected in any way to see what Canadians are doing with their money?
    No.
    Can any member of the government access Statistics Canada data that has been collected?
    They can't in individual form, only in aggregate form, which has been completely scrubbed.
    What do you mean when you say “aggregate” form?
    It's rolled up into statistics, i.e., a trend in a particular neighbourhood or for a particular segment of the population. It's never about an individual, either a firm or an individual Canadian.
    Can any politician access Statistics Canada data?

  (1640)  

    Only in aggregate form, never as an individual record.
    Can any member of the opposition access Statistics Canada data?
    Only in aggregate form, never as an individual record.
    Is there any pressure from any politician to access the collected data from Statistics Canada?
    No.
    Is there—
    I will have to cut you off there.
    Mr. Albas.
    Thank you, Chief Statistician, for coming in today and for the work you and your organization do for Canadians.
    I'd like to start with clarifying a few things.
    First of all, many people were under the assumption that it was 500,000 individuals. In your remarks today, you said “households”. Are those 500,000 households or individuals?
    It's 500,000 dwellings. Our unit of measure there is a particular dwelling. From there we create census families, economic families and other prototypes.
    What is the average per household?
    The average household size in Canada is just under three, I think.
    Are we looking at approximately 1.5 million people who will be picked up by this sample?
    Well, it's a little bit more complicated than that. What we're trying to do is to collect information at a neighbourhood level. We know that in some neighbourhoods you'll have smaller household sizes. There could be an apartment building, for example, or a student residence or whatever it is. Others would be more single-family dwellings.
    Obviously, Canadians' health records as well as their financial records are quite important to them. Suddenly having a change in scheme has raised, I think, a lot of concern.
    When you are collecting this information, will it include children?
    Our intent is not to include children.
    How do you safeguard that if someone has a bank account and lives in that household—
    That's why we do need some level of personal information, so that we can detect what are outliers and what are in scope.
    What if someone has a joint bank account for a relative—let's say an elderly mother or father who's on their own—that their name is on also and they live separate in a different household? Will that then be caught up by your survey of that particular household?
    If we're talking about a multi-generational occupied dwelling, for example, then that would form—
    I'm not speaking about that, sir. I'm speaking about specifically other people. Some people will have joint accounts because they have control over those accounts for an elderly parent or grandparent. Will that elderly parent or grandparent be caught up because their social insurance number is tied to that bank account?
     This is still in a design phase. It's a pilot project.
    Would you say yes or no?
    Once the data comes to us, those are the kinds of things that we will look at.
    There are definitions about the conditions under which a census family is formed, an economic family is formed. Those are the kinds of things that we will determine.
    We're not interested in an individual's transaction. We're interested in the household consumption pattern.
    Let's go back to the pilot project.
    Originally, the letter you wrote to the Canadian Bankers Association did not mention anything about a pilot project.
    Also, I'd like you to confirm, sir, that TransUnion, a credit bureau, has given Statistics Canada personal information on 27 million or so Canadians, when it comes to their credit file, dating back 15 years. Is that the case? Does the pilot project you're discussing today include that?
    You've said that this pilot project has not gone forward. I'd like you to explain the reason.
    First, we've been very clear both in writing and verbally with the CBA that this is a pilot project. It remains a pilot project in its design phase.
    In terms of your question about our work with TransUnion, we worked with them over the course of a year. We worked in collaboration with them to explain exactly why we needed that data and how we needed that data. Something they have been very clear about with their clients as well—
    No, sir, I don't think that's the case.
    —is that this is data. You can look at an individual—
    I'd like just to move on here.
    The problem I have here is that the pilot project you're talking about with the banks is not possible without having the credit bureau information. The two are intrinsically linked.
    Let's say you pick a particular neighbourhood randomized in Canada. You pick a household and you then use the credit bureau data to go to all of the banks, and I imagine this would also be for institutions such as credit card companies. Then you could say, “We've identified this person's social insurance number. We know they have an account with you. We would like you to supply us with all of that information.”
    We're not talking about just bank accounts, sir. We're talking about credit card-specific information. Is that correct?

  (1645)  

    Let me provide some clarification.
    First of all, this pilot project with the financial institutions is about expenditure data, individual financial transactions that we're getting, as I said, from a one in 40 sample.
    The transactions we get from TransUnion, for example, are looking at the credit. We're using that data to look at housing statistics and the degree to which people are over-leveraged.
    That is a complete—
    How are you going to be able to get credit card information then?
    For example, some people will have a chequing account at a bank but then they'll use a credit card for much of their activity. You need to have access to both sets of information to do exactly what you're talking about.
    How do you do that then, if you're not using the credit bureau information? The bank that you ask that information for will not know about that credit card.
    We're not interested in matching those two, first of all.
    This is about household expenditure patterns. This project is about looking at what a typical household spends, and what they spend that money on.
    The reason that's important, sir, is that we use that expenditure pattern, by a demographic profile, and their pattern then fits as the weight for our CPI project. It tells you that, in this particular situation, a typical household spends so much on these types of services.
    Every month we go out and get prices to feed into our consumer price index and then we benchmark and weight up that one commodity against how much a typical household spends on clothing, food, mortgage or digital services. That's what this project is about.
    Again, if you were only able to—
    Thank you.
    The time is up. You'll have time. We'll come back to you.

[Translation]

    Mr. Boulerice, you have the floor.
    Thank you very much, Mr. Chair.
    I'd like to thank the witnesses for being here today.
    To start, I want to tell you that the work you do is extremely important for the development of effective public policies that reflect the Canadian and Quebec reality. We strongly believe that accurate data allows us to make the best possible decisions. Otherwise, we're sort of in the dark, and it's extremely fuzzy. That's why, at the time, we so strongly defended the long-form census, a tool that we feel is absolutely essential.
    Having said that, you have sparked a whole shock wave. Many people are concerned that a public institution may have direct access to their personal information through a bank or credit card company. It makes people very nervous.
    I first want to talk about citizen consultation and transparency. In your presentation, you said that you held information sessions and spoke to people. I don't know who you spoke to, but I don't think it was with a lot of people, because the media, journalists and people in our ridings are worried. They learned about this pilot project in the newspapers and are a little shocked.
    What consultation process did you follow to inform people of the upcoming launch of this pilot project?
    First, I understand the concerns of Canadians about this pilot project. As I said earlier, this project hasn't yet begun; it is still at the design stage. We are still working with stakeholders to determine how best to communicate with their clients. Again, the intention is to be transparent. We want to tell Canadians that this is what it takes to have high quality data. We have done this several times before in other cases, and that is exactly what we'll do for this pilot project.

  (1650)  

    The Privacy Commissioner has already mentioned that he will launch an investigation into your pilot project.
    If the commissioner ever concludes that this doesn't meet the standards for protecting the privacy of citizens, what will you do?
    We'll see. Let's let Mr. Therrien do his job first.
    Are you willing to commit to following his recommendations?
    Of course. We'll take his recommendations into account.
    I also asked Mr. Therrien to give us other recommendations, if he has any. We worked with him during the design period of this pilot project and took his recommendations into account. He has received complaints, and he will do his job. If he still has recommendations to make, we will, of course, take them into account in our pilot project.
    You said that, since the communication methods aren't the same anymore, there are no longer many fixed telephones in the home and that people communicate more on social media. So they are more difficult to reach.

[English]

You used the expression that traditional surveys are unsustainable.

[Translation]

    You aren't the only organization facing this challenge, but it seems that the solution you've found is quite intrusive as to people's lives and their personal and banking information. No polling firm would dare do anything like that. Personally, I'm not convinced you've found “the” solution.
    In recent years, has Statistics Canada been giving mandates to subcontractors or third parties to collect information, on which Statistics Canada relies to feed its database and establish its statistics? Was the pilot project designed in-house or did you outsource some activities?
    We designed it ourselves, and the data remains in our buildings and systems. Of course, everyone buys software, and we do too, but we never pass on the data we collect from citizens to anyone.
    Do you use the services of a third party to collect data on your behalf?
    No.
    Do you use Shared Services Canada?
    Shared Services Canada provides services to everyone. We have given them the mandate to maintain our systems, but we control our data and how employees access it.
    Do you think the level of security and privacy protection at Shared Services Canada equal that of Statistics Canada?
    As I said, we control our servers and our data. They don't have access to our data.
    Okay, thank you.
    Mr. Boulerice, you have 40 seconds left.
    I'll give them to you.
    Great.

[English]

     We're going to move to Mr. Sheehan.
    You have exactly seven minutes.
    Thank you very much, Mr. Chair.
     I want to thank the chief statistician and staff for appearing before this committee.
    We had unanimous consent to ask you to come here to answer some questions. We appreciate your appearing on short notice.
    My question is around your engagement to date with the financial institutions. Could you please delve into that a bit more for us?
    We started working with the Canadian Bankers Association early this year, around January or February. As I said in my opening remarks, we've had about a dozen interactions with them, either in person or on the phone. We've shared with them why we need it. We've also shared with them the design of the project as we've had input from the Office of the Privacy Commissioner. They've asked about the authorities. They've asked about the privacy concerns and so on. We've certainly tried to answer all their questions, and we continue, as I said, to work with them on the design of this project.
    The design I just laid out, the separating of the files and bringing them in, etc., is something that's very clearly shared with them. As I said, we met with them and their members in person in June and August and laid out where we are. Based on the last meeting we had with them, their advice was that this is now the time to move from understanding the purpose and the design to now going deeper with each of the institutions. It's now time to get at what the unique system needs and what the unique file formats are, and so on. They provided us with the contacts of the relevant people in the individual institutions, and they asked us to spell out very clearly the authorities under which we would operate.
    Like any other provider of information, we work with them. We make sure we understand their concerns and their clients' concerns, and we make sure these are addressed so there isn't a negative impact on their clients. That's just the way we operate, and it's no different from how we operate with all our providers.
    As those financial institutions work with us to tell us about their specific needs and the concerns of their clients, maybe there will be some design elements that we will have to include that are unique to their systems, but we will never, never, never compromise the privacy or the confidentiality of the individual transactions we'll get access to.

  (1655)  

    In your engagement with these financial institutions was there any discussion of informing their clients about your pilot project?
    Yes, absolutely. In August, we made it very clear to them. We asked them to ensure that as we roll out this project, informing their clients will be an important element. We asked them to let their clients know that Statistics Canada may—because we're talking about a sample of one in 40 clients—look at their data. We let them know that yes, this is a legitimate purpose, and we told them what it will be used for and how it's to be communicated to their clients.
    They have their own systems. They have their own consent, etc., that they do. That's the kind of work we still have to do, and we were fully intending on elaborating a public communications and a rollout plan. That's where we are today.
    Thank you very much.
    I'm going to share some time with Majid as well.
    Thank you, Mr. Chair.
    Thank you, sir, for coming here and sharing your insight into the pilot project. Let me ask a couple of questions.
    Can you tell me the benefit that you are trying to get from this pilot project? I'd like you to expand on that one and really get into what the scope of the pilot project is. I understand the stakeholders and I understand some of the demographics, but what were you trying to achieve by initiating this pilot project? What is the benefit you are going to give us?
     There are numerous benefits.
    First of all, as I said, the quality of the data that's coming from our traditional methods is not meeting the needs of—
    Can you give us an example?
    Let me talk about seniors, for example. Let me, then, talk about a business person who wants to expand their business or even just somebody who's between two jobs—
    Seniors' pension would be an excellent example.
    The amount that a person receives as part of their old age benefit, their pension, etc., and whether it's enough, in a sense, to keep them sustained is dependent upon the rate of inflation. It's indexed in a sense. It's then benchmarked to how much prices are either going up or not in certain categories.
    The amount that individual will receive on an ongoing basis is determined by the work that we do. The better we are at benchmarking the types of services and the consumption Canadians have, the better the data are for that to be benchmarked.
    It's similar, as I said, for a business. A business person who wants to expand their business wants to know the ability of that in that particular neighbourhood and what the consumption pattern is. What are they consuming? They can then decide whether it makes sense for them to expand or not or maybe go to another area to expand.
    You have people who are holding multiple jobs these days. How are they procuring services? They could be participating in the gig economy. They may be an Uber driver and may be renting out their place under Airbnb, and to be able to understand what a social safety net is and what the vulnerabilities are.... These are the kinds of questions policy-makers ask of us.

  (1700)  

    Talking about vulnerabilities, I'm definitely impacted as a member of Parliament and as an individual by the increase in the interest rate.
    How will your analysis be able to help me, as a consumer, be able to deal with a situation that may arise as a result of the interest rate?
    You have about 15 seconds to answer that question.
    You'll have better data at a local level to understand the implications of, let's say, rising interest rates.
    Thank you very much.
    We're going to move to Mr. Lloyd.
    You have five minutes.
    I appreciate the testimony today.
    As we in this room all know, this issue's been quite a political firestorm over the past few weeks, and I just wanted to know, did you or your office coordinate in advance of this meeting today with any members of Parliament, ministers or ministerial staffers regarding your appearance at committee today?
    We, of course, have kept our minister's office in the loop on the fact that we're coming to this committee, but these remarks are my own.
    I appreciate that.
    Next, some of my constituents have raised this issue: Does Statistics Canada provide aggregate data to private for-profit companies?
    First of all, most of everything that we produce in terms of aggregate statistics is done so that people can use good data rather than the alternative, and we put it on our website.
    Are there instances where data has been sold to companies?
    We do, from time to time, get a request from a company, for example, that says, “Your standard table puts it this way, by this way, by this way. Really what we want to do is to have a different rendition, or we want it for a different geography, because my geography is different from your geography.”
    But do these companies pay Statistics Canada for that data and you configure it in the way that they ask you to configure it?
     Our starting point is that we recover the cost of being able to customize that request to their needs using the aggregate data that we have. We feel the taxpayers should not be subsidizing the cost of a unique request or an individual's request, so it's a cost recovery.
    Thank you. I understand.
    I am part of the millennial generation, and when I use Facebook or Google or any of these platforms, there are always these long terms of agreement. There are various instances where they ask me for my consent when they want to know where my location is. I have the option to consent to that or not.
    Why is Statistics Canada in this pilot project not giving Canadians the same option to consent?
    We know that people who consent look different from those who don't consent. When you get local areas for which you need good-quality data, or if you want to now know about a target population—how seniors are impacted or how single parents are impacted by certain things—the cross product, if you like, of that low level of geography and a particular targeted subpopulation requires a higher level of quality of data.
     Participation you're saying.
    It's no different from the census long-form debate this country had. As I said, we know the profile of those who consent is different from the ones who don't consent. Our ability to essentially account for those who don't consent doesn't allow itself to be substituted or weighted up, if you like, to those who do, because those who consent don't look like the ones who don't.
    We know the youth don't participate. We know in many cases single parents are too busy and they can't participate. So we miss data in many cases for the very audiences that policy-makers are trying to get a better handle on.
    With that, though, you believe Canadians should not be allowed to refuse consent for Statistics Canada information.
    Look, as I said before, we understand the concerns of Canadians. We get it. We deal with sensitive data all the time.
    But if a Canadian writes to you and says they explicitly do not want Statistics Canada to have this information, would you respect their request?
    We have, give or take, 400 programs in Statistics Canada. Many of them are mandatory and many of them are voluntary. The difference simply is, as I explained earlier, what the level of precision is, what the level of accuracy is, what the level of geography is to which this information is needed.

  (1705)  

    In some cases it's good enough to have it at a Canada level or a provincial level. In some cases, for information that really matters, for issues that are important to Canadians, we have to have high-quality data.
    Some jurisdictions in the U.K., though, have introduced nudge models where you are automatically considered to have consented unless you explicitly say that you do not consent.
    Why has that not been considered as a model for Statistics Canada?
    First of all, what Statistics Canada is doing is consistent with what most modern statistical agencies in the world do. They have carve-outs from privacy acts, PIPEDA, etc., for statistical purposes, and it's only for statistical purposes that we can in fact go and get at representative data, even in some cases without consent.
    Thank you very much.
    We're going to move to Mr. Longfield.
    You have five minutes.
    Thanks, Mr. Chair.
    Thank you, Mr. Arora. This is a great conversation.
    I used to use the CANSIM reports when I was in business. The Canadian paper industry was going through transition. I reported to a board of directors in Europe that asked where the emerging markets were that would substitute for the business we were doing in the paper industry. We looked at wind turbines and steel mills. We got some of those customized reports, and paid what it cost for those. In terms of our business it was very important to have good data.
    Europe seemed to always be one step ahead of us on data, and I had to report to Europe. As a Canadian business director, it made it difficult for me.
    Now I look at the labour force surveys. I wait for the Friday that they come out because in my riding we have 3.6% unemployment. There's a lot of stress on business to find workers. We need data for the business community in my riding.
    The not-for-profits that are working on poverty elimination are looking for data. They were more concerned than I was when the long-form census was cancelled.
    When we went from the long-form census being mandatory to voluntary, what was the impact on data? The people in the not-for-profit sector said, “Now we're drifting. We don't know where we're at.”
    Could you let us know what happened there?
    Sure. I've done a few censuses in my life.
    When we went back and looked at the results in 2006, 2016, and the 2011 national household survey results, we can see as I was saying earlier, the impact of people who respond and who don't. We know for approximately 1,126 communities out of 5,000 communities, the data for that level of detail was just not good enough for the kinds of needs that Canadians have.
    Right. Communities, because of their size, became invisible, and some of our most vulnerable communities didn't get serviced. Is that what you were saying?
    In many cases we had to take it up to the next level of geography because we weren't confident with the level of information at that smaller level. As I said, we do know that the prevailing traits, if you like, of the population start to get amplified when you don't have enough detail.
     Right. I'm thinking that in terms of expenditures, when you're in northern or remote communities, you're going to have a different mix of expenditures that would cause us to look at different types of social programs in order to support them. Really, that's what we're trying to get to.
    Clearly somebody living in Whitehorse or Resolute Bay has a very different cost and profile for their clothing, food and shelter than somebody living in downtown Toronto or Vancouver.
    I was working on another project with our downtown. Guelph is on the list of designated places to grow. We're looking at getting 8,000 people into the downtown. We're looking at creating 4,000 jobs in the downtown.
    The downtown businesses in Guelph are saying they wish they knew the neighbourhoods around the downtown. Are they professional people? Are they people who could shop in the downtown? What kind of shops would we want to attract into the downtown?
    Are those standard reports, or is that where you're trying to head with this pilot project, to be able to give, let's say, business improvement areas information on the neighbourhoods around them so that we can have the right types of retail opportunities?
    I think you're pointing out exactly the kinds of questions that are being asked of us by Canadians, associations, businesses, social support groups and so on. They're saying, “Just giving me the data for Vancouver isn't good enough. What do I do with that? I know that Surrey looks very different from Richmond and downtown Vancouver. Those data don't help me expand my markets or even understand whether people will actually be able to consume the product if I do put something in place."
    You see the changes in demographics, immigration and so on. Knowing the tastes and habits of people in that area, how they are changing over time, and how different subgroups in that area consume products and services differently than others is what this project is designed to do—

  (1710)  

    Again, there is a wall. We can't know what address has what.... When I was in business, I would have loved to know which paper mill was no longer going to be buying my product.
    Not from Statistics Canada you won't.
     I couldn't get to the paper mill. I couldn't get to the level of detail. I could only get aggregated detail. In terms of policy, that's what we're looking for.
    Yes. Again, when I say we take the privacy of Canadians very seriously, it's not just a “trust me”. There are actually processes and procedures in place.
    Thank you.
    Thank you very much.
    We're going to move back to Mr. Albas, for five minutes.
    Thank you again, Mr. Chair.
    I want to get back to your opening statement. You said the chance a given address is selected as part of our sample is one in 28. The chance that the dwelling is used in the actual sample is 1 in 40.
    To me that seems to say that you're actually oversampling. Is that the case? Do you really need 500,000 households?
    One, we don't want the banks to even know which dwellings are going to be in the sample.
    By what ratio are you going to be oversampling? How many Canadians do you actually need for this project to go forward, and why are you using 500,000?
    What we need in order to make sure that we have enough information for our census areas, essentially at the neighbourhood level, is about 350,000 dwellings. However, in the design of the project, even the providers of that data don't know which dwellings are going to be used in the sample.
    You're going to be asking for confidential information that people back home are upset about, sir. You only need 350,000 households, and you're going to be sampling 500,000. I appreciate your answering the question, but I'm disappointed to hear that.
    When it comes to the actual retention of your data, you talked about having the two different areas, and how they're separate. Will Statistics Canada maintain a master key that can reidentify the information?
    We will keep, as I said, the two files, as you explained. The identifiable information with the StatsCan number comes in. The actual financial transactions come in. Once we have associated a transaction, the expenditures that my family and I have—
    Will you be able to re-engineer it using a master key to reunite those files if you so choose, yes or no?
    If there is a policy need for us to be able to do that, we have a very controlled process in place. It is only—
    It sounds like the answer is yes.
    Will Statistics Canada maintain a technical capability to access the data with personal identifiers after anonymizing the data?
    Sorry, could you repeat that?
    Will you maintain the technical capability to access data with personal identifiers after anonymizing the data?
    If you have a master key, will you have the ability to have those files back together at some point in the future?
     One, there are no fishing expeditions here. When we have a specific policy need, a case has to be made to be able to link that to another source. That case has to be looked at and agreed to that whatever other source is with that key is joined together and only the anonymized microdata is given to the area that—
    I understand the separation, but why would you not remove the records that have personal identifiers after anonymizing the data?
    Well—
    Why not delete it?
    Essentially, there are retention periods for files such as this. This is still a pilot project. We are trying to assess exactly what the short-term and long-term needs are for these data.
    Again, to your point earlier about 500,000 dwellings, as I said, that is out of a universe of 14 million households that we have in Canada—
    Okay.
    —and a fresh sample is selected every year so that we cannot create a file that keeps that transaction for even the selected households.
    To do this game, you're going to have to require the use of the banks. Obviously, the banks don't have the infrastructure to make that information available.
    You've also talked about real time and frequency, higher frequency use of the information. Are you going to be seeking to port the information through an API directly to Statistics Canada from the banks themselves? Will this be required of other institutions? Will it also be put upon credit unions, ATB in Alberta, trust companies that do deposit-taking activities?

  (1715)  

    At the moment, the project is restricted to nine institutions. We have secure transfer protocols that meet all the Government of Canada standards for secure transfer, and that is how we're bringing the data into Statistics Canada.
    As you can imagine, we have millions of transactions with Canadians where they're providing us with their really confidential and sensitive data. We use the same protocols that we get to bring that data into Statistics Canada, and we have processes within Statistics Canada that we've built over 100 years with input from the Office of the Privacy Commissioner, as well.
    You didn't build the infrastructure for this real time. Again, when someone fills out their census, sir, they know exactly what they're doing. In this case, you're not even advising them that this information may be taken. What if someone moves from one area to another—
    I hate to—
    —and then ends up being sampled a second time?
    Mr. Albas.
    This is an intrusion in someone's—
    Mr. Albas, I hate to tell you, but you're out of time.
    —personal...and I believe government should not have that power.
    We're going to move to Mr. David Graham.
    You have five minutes
    To Mr. Albas' question, would households in any way be informed that their data could be collected?
    Yes, that's what we expect the institutions to tell their clients. We made that very clear to them in August. We asked them to say it was part of their process. We want to operate in a transparent way.
    So it's not done in some magical background way. I appreciate that.
    Would it be accurate to say that the 500,000 households collected for 350,000 samples used is to ensure that the data is properly blinded?
    Sorry, I didn't hear the first part of your question. Maybe the microphone is just a little farther away.
    Or it's just me. That also happens.
    Would it be accurate to say that the 500,000 households collected for only 350,000 used is to ensure that the data is properly blinded?
    That is simply to make sure that even the institutions that are providing us with that information don't know which records were actually used. It's noise, if you like, to further protect the privacy and confidentiality of—
    Is that a standard methodology?
    Many times we use that methodology to mask, if you like, from the person providing that information, not knowing and being able to replicate that kind of analysis.
    I understand.
    To Mr. Lloyd's question, what's the legal basis of your ability to compel data without consent?
    Section 13 of the Statistics Act is the actual provision. That makes it very clear that, for statistical purposes only, we have the right to go to seek administrative—
    How long has that been on the books?
    That's been there since the inception of the Statistics Act—at least since 1971, if not even earlier than that.
    Do statistical agencies in European countries have the power to compel data, in spite of the GDPR?
    Yes they do. In fact, the GDPR has a carve-out in various sections, for statistical purposes.
    Mr. Lloyd asked this earlier. If someone sent you a letter saying they didn't want their data being used, would you even be able to identify the data as theirs to take it out?
    It would take a lot of gymnastics to find out whether their individual record or the household would be.... As I said, once the linkages are made, those linkages are put away in a vault, and we're dealing with the individual record in an anonymized fashion after that point.
    Do the banks already have the information that you're requesting from them?
    Yes, they're the source of this information.
     Can the banks use that information in any way and for any purpose other than to simply process the transactions?
    I couldn't speak to what the banks can or can't do. I don't know what their policies, procedures and processes are.
    That's fair.
     You know that the businesses need the improved data that come out of this research. Could the banks themselves benefit from this aggregated data?
    Sure. The banks use this to expand their client base as well, and they obviously are impacted by inflation, CPI, interest rates and consumption patterns of households.
     Even the formula on which a mortgage gets approved or denied is based on household leveraging and income and so on. Yes, the banks very much use this data. I think they will also benefit from the strengthened lower level of geography and more timely data.
    Are some banks—
    Again, at the aggregate level—they would never be able to get at who said what.
    In your opinion, are some banks concerned that other banks might be able to benefit from their data being in the pool?

  (1720)  

    I don't see how. I'm sorry. Do you mean—
    If a big bank sends in data, a small bank or credit union then might say, “Wow, look at this great data that we got from Statistics Canada from banks.” Is that providing a competitive advantage to smaller banks and is that why the bigger banks are worried?
    The data we get from them would never be shared other than at the aggregate level.
    No, but what about the results of the data^
    The results would be available for everybody, all Canadians. That's why we collect it. We don't collect it for us. We collect it to share it with all Canadians.
    You mentioned—
    Again, in aggregate form....
    Right.
    You talked about the declining response rates of the old system. Can you explain how the old system actually worked?
    Sure.
    This is one example. As I said, there are a number of changes that are happening in today's society that require us to move in this direction. Previously, one of the surveys was the survey of household spending. We used to go out to 20,000 households. We used to actually give them a diary to keep track, and we still have.... We've suspended that project right now. Individuals had to actually take their salary and log every single expenditure, with the receipts, and be able to take their salary and calculate it to about a 2% to 3% variance. In many cases, we found that they didn't know, or they didn't have the receipts, or they weren't completing it.
    It's a growing problem that we've had. Now, 60% of those we go to either refuse or don't give us access to them.
    I have five seconds left. Do I have time for a very quick question?
    You have no seconds left. Thank you very much.
     We're going to move to the final two minutes.

[Translation]

    Mr. Boulerice, you have the last two minutes.
    Thank you.
    Mr. Arora, you said earlier that you intend to communicate with Canadians about this pilot project, ask them what they think about it and inform them. This seems to be wishful thinking. I didn't hear much detail on how you were going to proceed. Can you tell us more about that?
    As I said earlier, we are still developing the pilot project. We want to work with the institutions and the commissioner to develop the communications plan. Of course, the institutions themselves will communicate with their clients, and we will inform Canadians. We have published several documents on our website. We will also hold several information sessions during which we will answer Canadians' questions, explain why this process is necessary and how we will protect their privacy and confidentiality.
    How many information sessions will there be: three, 30, 50? Can you tell us?
    Given the concerns expressed by Canadians, we will, of course, increase the number of information sessions and the amount of information to be provided to them based on their needs.
    Okay.
    In addition, we live in a digital world that offers many opportunities for malicious people—I'm talking about hackers—to break into systems, appropriate information and use it for various purposes. Just recently, we have seen this with Facebook. This company is still able to hire staff who should normally be able to protect people's pages, their personal or confidential information, and so on.
    I'm sorry, but your time is up.
    Already?
    Yes.
    That's too bad, especially since it was a good question.

[English]

     Mr. Albas.
     I do recognize that we've had a lot of discussion here, and the chief statistician should be commended for coming. That being said, even he in his testimony today has pointed out that there are various aspects of this so-called pilot project.
    Is this a point of order that you're making?
    No, it's just that since we're on the business of discussing this particular topic, I'd like to move the following motion:
That further to the appearance of the Chief Statistician, that the Committee on Industry, Science and Technology invite the Privacy Commissioner, the Minister of Innovation, Science and Economic Development, TransUnion, the Canadian Banker’s Association, Ann Cavoukian and the Canadian Civil Liberties Association.
    This is so we can have further discussions on this important topic, sir.
    Actually, that was a motion we dealt with in committee on Monday, so it's not something we can do.

  (1725)  

    No, Mr. Chair, I would simply point out that, first of all, it's a motion relating to this, and it's an invitation, not a study. Also, we have TransUnion included in this, and that was not in the original motion you're referring to.
    If you won't find it in order I'd like to know, because then I would like to make it as a formal notice of motion.
    Again, the substantive motion was dealt with in committee, so I will rule against that.
    Okay, I'd like to put a notice of motion, Mr. Chair, “That, further to the appearance of the chief statistician, the committee on industry, science and technology invite the Privacy Commissioner, the Minister of Innovation, Science and Economic Development, TransUnion, the Canadian Bankers Association, Ann Cavoukian and the Canadian Civil Liberties Association.”
     I would hope that would be translated and that we would have an opportunity to discuss the motion.
    Again, I'm not asking for a report. I'm asking for an invitation. I would hope that members would say that more discussion is necessary. Mr. Longfield said the other day specifically that we can't discuss this further until we've heard from the chief statistician.
    You've submitted your notice of motion. The clerk will have that translated.
    Unfortunately, today we really don't have time to do anything because there is a second committee that's taking over here. That's what I needed to confirm. It's HUMA. It's all about accessibility and there's a certain set-up that's required here, and they need half an hour to do that. We actually just verified with the clerk that they need half an hour.
    Mr. Chong.
    Mr. Chair, I just want to support my colleague. Obviously he can't re-present the exact same motion, but if the motion is different, I hope you will rule it in order. I think it's really important that we do hear from additional witnesses on this matter.
    When we first established the census in Canada hundreds of years ago, it was entirely for the use of government itself, to raise armies or to collect taxes. With the development of the modern economy the data that StatsCan is collecting, the census data, is used by private sector corporations. Canadians are increasingly concerned about the development of large data being used by companies like Facebook, Google and the banks.
    The issue, I think, in front of us as a committee is whether or not it's appropriate for the federal government to use the coercive power of the state to force, compel banks to hand over very personal and granular financial information—
    You're actually getting into debate. We're not going to get into a debate right now.
    Okay. I hope you rule it in order.
    We've accepted the notice of motion that you've put forward to the clerk. It will be translated, and we'll go from there.
    Thank you.
    Again, Mr. Chair, could I just have clarity on this as to exactly why you ruled the previous one out of order? It was based on the conversation at the business of the committee here today. It was substantially different, in that we were not talking about a study but about invitations, just the same as the Liberal members did last time.
     I can't see why you would rule this out of order. I'd like an explanation.
    I haven't ruled your notice of motion out of order.
    No, but the previous motion.
    I ruled the first one out of order because we actually dealt with the substantive motion in committee business. I can't really have that conversation out of committee business because it's in committee business.
    You've put forward a new notice of motion. We'll deal with that as it comes.
    As it stands right now, I just want to point out that no motions were passed to get the statistician here. It was an ask for unanimous consent to bring the statistician in and have a conversation with him. That was the ask. It wasn't a motion regarding a study, regarding anything. It was very specific: Do we have unanimous consent to have the chief statistician come in? It was unanimous. Everybody agreed with it, and that's why the gentleman is here today. That's the business we are doing today.
     Yes. Again, that's why I suggested we do something similar. We were just doing the business here, Mr. Chair. It is clearly tied to the business of the committee today, so I can't see why it wouldn't be in order.
    That being said, we will prepare ourselves for future meetings where we will be taking a much different approach until we have this issue looked at, and I do hope that colleagues consider that.
    Thank you.
    No problem.
    On that note, I'd like to thank our guests for coming in today. It was very enlightening, and we look forward to the next time we meet.
    The meeting is adjourned.
Publication Explorer
Publication Explorer
ParlVU