Industry Committee on July 27th, 2010

[Charlie Angus, NDP, Timmins—James Bay, ON]

Thank you, Mr. Chong.

Looking at the questions then, and the reasons.... We're trying to find out why a political decision was made, and the example my colleague mentioned was Jedi Knights. A week ago, we had Dimitri Soudas, who is not known as a statistician, trashing the work of Statistics Canada and claiming that 21,000 members of the Jedi Knights were in the survey.

Mr. Boyko, when you look at survey data and see a spike like that for the Jedi Knights, is that not common--there are certain areas where people might provide misinformation? Are you still able to use the data? Is Statistics Canada's credibility damaged because 21,000 people said that they were Jedis as opposed to Druids?

[Ernie Boyko]

We're talking about 21,000 out of a population of respondents of 12 million households, or three million households in the case of the long form.

When you see responses like that, you can deal with them analytically and it does not destroy the credibility of Statistics Canada any more than it did the Office for National Statistics in the United Kingdom.

There are people who respond a certain way, and we don't really worry about that. The large bulk of people who responded to the question on religion responded appropriately, so there was not a huge worry.

[Charlie Angus, NDP, Timmins—James Bay, ON]

Mid to late last week, the Conservative talking points changed, and they suddenly invoked Denmark in their reason for scrapping the long-form census. We had asked Mr. Clement who he spoke with. He doesn't seem to have spoken with anyone in Canada, and I don't know if he spoke with anyone in Denmark.

Mr. Drummond, my understanding of Denmark is that the reason they can data-mine is that every single citizen has an identity number. With that identity number is included their credit history, where they are living, what their job status is, and all manner of trackable information, which makes data-mining much easier.

Now, I'm not sure if the Conservatives are going to this number that we all have to sign on to so they can track our credit history and know where we're living at a given time, but they seem to be saying that Denmark is a better solution for us than Canada.

Are we talking apples and apples here, apples and oranges, or are we talking about a scramble to come up with some excuse as to why they made this move?

[Don Drummond]

Well, this is the alternative to a census like the one we have in Canada or in the United States. You use administrative records. Several Scandinavian countries already do this, and this is something that the United Kingdom has indicated it wants to study for 2016.

You are absolutely right, they track it through a variety of administrative sources, and they track absolutely everything. You may not be aware directly, so it may not be as intrusive to your privacy, but somebody is literally following your every single move: if you move, if you change jobs, if your income status changes, if you buy a new car, and so on.

It's more expensive, we do know that, because the countries that implemented it had to increase their overall public expenditures. I think it's a debating point as to whether it's less intrusive. My own view is that their system is actually more intrusive than the type of system we have.

[Charlie Angus, NDP, Timmins—James Bay, ON]

So it would be more intrusive than what we have now.

The other interesting question is at what point do people draw the line on their own personal moral privacy and are willing to be dragged off to jail? We haven't seen anybody being dragged off to jail yet, but apparently there are certain questions that trigger that.

Mr. Anderson said this morning that renovation was one such issue. He seemed to take personal offence to renovation. I find it surprising from a party that stood up in the House and waved their renovation tax credit. I would have thought that they based that on data that they may have gotten from the census, and are very supportive of renovations and having tax credits.

Nonetheless, let's say, for example, that 100 of Mr. Anderson's constituents are deeply offended by anybody asking them if they've ever renovated their house. Is it not possible to have those questions removed? Is there not a system, before the census is brought forward, where there's a review of the workable questions? As we understand it, the minister can even say, “Under my watch there is absolutely no way that you're going to interfere with people's privacy and ask if they've ever renovated their home”.

If we remove the question on renovation, does it change the nature of the census?

[Don Drummond]

You're absolutely right on the process. The questions are always put to cabinet. As Mr. Bernier indicated, the 2006 census would have gone to cabinet in 2005. At the latest, the questions for the 2011 census will go to cabinet in December 2010. They're proposed by Statistics Canada, but they are under the ultimate control of the cabinet. They can absolutely be changed.

I don't think it's fruitful at this time to go through every one of those questions, but I would refer you to the set of principles we have put out as an advisory committee. One of the key ones is that if you can get it on a reliable basis from another form, then don't put it on a long form of the census. You can get a number of these things.

Much of the discussion has been around a binary choice. You can continue it exactly as it was or you can move away from it, and I don't think it needs to be that way. It can be moulded into something that causes less difficulty. Keeping that in mind, as an advisory council we tried to determine how many complaints there had been through the privacy commission, the Prime Minister's Office, the Privy Council Office, and Statistics Canada directly. There were about 100 on the 2006 census out of over 30 million who would have filled out the long form--roughly one-twentieth of who would have filled out the long form.

[Charlie Angus, NDP, Timmins—James Bay, ON]

It would be interesting to see how many complaints Mr. Clement has received on cellphone fees. It would be amazing to see if he acted on those tens of thousands of complaints, as opposed to the 100 he may have received on this.

Since these questions go to cabinet and our colleagues across the road are concerned with questions about renovation and water, should this not have been an internal discussion at their caucus meeting, since Minister Clement signed off on the questions he's now denouncing? You say that cabinet would have had to approve them. Then why are we having to discuss questions that they would have already looked at?

[Don Drummond]

Well, of course we have no idea whether they approved a set of questions, because they proposed a different form. So I have no basis on which to speculate what happened in the cabinet discussion.

On a normal basis, yes, whatever questions were in the 2001 and 2006 censuses would have been approved by cabinet. I have no idea what the process would have been leading to this.

[Charlie Angus, NDP, Timmins—James Bay, ON]

But they could have taken out any questions they found offensive.

[Don Drummond]

Absolutely--and proceeded without those questions, absolutely.

[The Chair, Conservative, Michael Chong]

Thank you, Mr. Angus, and thank you, Mr. Drummond.

Mr. McTeague.

[Dan McTeague, Liberal, Pickering—Scarborough East, ON]

Thank you, Mr. Chong.

I don't know where to begin here.

Mr. Bricker, perhaps you could give us an opinion, from your recent survey, as to how many Canadians would not comply if it were simply a voluntary system. That's a very simple question.

[Darrell Bricker]

It would be 19%.

[Dan McTeague, Liberal, Pickering—Scarborough East, ON]

Okay.

Mr. Drummond, with an inaccuracy level of 19% to 20%, how difficult would it be for the Minister of Finance for Ontario, who said last week on July 20 that in every spending and tax decision, the census information assists them? You sat, sir, in the federal finance ministry, I believe as deputy minister. Without accurate, reliable information, or with information that's skewed 19% to 20%, how difficult will it be for the minister, Jim Flaherty, when he makes his next budget decisions or future budget decisions, should he still be minister, in terms of reliability of information?

Clearly, one in five not reporting is going to create a significant skew. In your professional opinion and with your experience, what does that do for data collected generally? How is it going to impact on, let's say, a decision by our Conservative colleagues to slash social spending if they don't have accurate information on what social spending should be for the makeup of the country?

[Don Drummond]

Again, the fact that one in five might not respond is not terribly relevant. Having something like 25 million Canadians respond to a survey would be overwhelming. Angus Reid would do a survey with 1% and consider that to be a huge survey. The labour force survey is only 40,000 households, and we base a lot on that, so the size is not.... The key question is whether the 19% who didn't respond would be representative of the overall population. So it's the skewness within the response rather than the overall response.

Again, we can only draw on the results of previous surveys, particularly the experiment in the United States in 2003, where they did a pilot project on a voluntary basis and benchmarked it against the actual survey. The people who did not respond were highly unrepresentative of the total population.

But you don't really know, without having done that, exactly where the non-representation would be. We certainly overwhelmingly would speculate that they would be at the two extremes of the income distribution, and they would tend to be in first nations communities and our recent immigration.

If we knew what that bias was, even that would not be a problem, but we have no reason for doing it a priori. If, for example, in 2011 we did one more run of the mandatory survey and we ran a voluntary one on a pilot project at the same time, then we could benchmark them. Then I think we'd probably be okay, because we would know how to adjust for those biases.

[Dan McTeague, Liberal, Pickering—Scarborough East, ON]

This is perhaps a question for you, and maybe Mr. Bricker, but I'll go to you first, Mr. Drummond.

Do you feel that the Government of Canada...? You've heard it from several Conservative members here, suggesting that the census is an intrusion, a violation of the privacy of Canadians. Do you think those kinds of comments will favour or help or encourage people to actually participate in future censuses?

[Don Drummond]

I do worry that the longer this continues, the lower the response rate we'll get. We have been reasonably successful without enforcing fines, and there has never been prison. People feel it's a duty to fill this out, and whether they want to or not they do it and off it goes; they don't deliberate on it a lot.

I do worry that we can create a sense that, “You know what? You don't really need to fill this out. It's not that important.” Again, I don't really care if that brings the overall response rate down, but it may do it in a differentiated fashion, of course--people in different socio-economic conditions.

[Dan McTeague, Liberal, Pickering—Scarborough East, ON]

What does it say when you have one particular position taken by the minister on the long form, to remove the mandatory aspect of it, and not the same with, say, the agricultural form, on the form relating to, say, the short form, if...?

Is there a possibility that this patchwork of obligations not only creates confusion above the government's narrative that the intrusion of privacy is encouraging people not to be there...? What does that do, apart from creating absolute confusion?

[Don Drummond]

I think the perception could be that it would create difficulty in conforming to other voluntary and mandatory....

The other thing that's important to note is that in almost every survey, as was indicated, the reason we can get reliable data from very small samples is because we target the people to survey using the long form. For example, I believe that if we did not have a mandatory long form we would have to increase the size of the labour force survey. Keep in mind, the labour force survey is only the second of two compulsory household surveys; there are no other compulsory household surveys. In some sense, if we get a lesser response on the census, we will have to increase the sample size in the labour force survey, which is just as intrusive and--

[Dan McTeague, Liberal, Pickering—Scarborough East, ON]

My take is that this is going to damage beyond repair the ability for us to collect critical, credible data down the road.

Do you also think that this perhaps is an attempt to try to privatize this process, make a few people happy? A couple of people might be contributors to the Conservative Party. Ultimately there are people who around this table have said that they can do this kind of information.

If it's done from a perspective of the private sector, how credible is it in creating analysis for public policy-makers down the road?

[Don Drummond]

Well, I think it is critical. I won't speculate on the motive to privatize. I don't think you need to speculate on that. I mean, there's a legitimate reason to look at it. It is intrusive. That's why I'm amazed, when people are asked whether the questions that are floating around are intrusive, that some people are saying they're not. I'm kind of surprised by it, because let's face it, the long form takes 20 or 30 minutes. It asks you private questions.

I personally think it's worth it, and the advisory council thinks it is worth it, because there are benefits from it. But when you're asking people for private information that is taking that 20 or 30 minutes of time, is it the absolute best set of questions? Is it the minimal set of questions? Are they the right sort of penalties? I don't think we're quite there. I think some of those things could be changed.

That's why I'm kind of surprised here. We can go, okay, there are some things that are perhaps not as good as they should be in the long form of the census, but why are we abandoning it? Can we not deal with it and improve that? There are legitimate reasons to question it. There is a legitimate debate that needs to happen here about the intrusion of these questions.

[The Chair, Conservative, Michael Chong]

Thank you very much, Mr. McTeague and Mr. Drummond.

Monsieur Bernier.

[Maxime Bernier, Conservative, Beauce, QC]

Thank you, Mr. Chair.

I will ask my question of Mr. Veldhuis from the Fraser Institute.

You wrote recently--I will quote you--that “The long-form questionnaire...is a truly intrusive instrument”. As Mr. Drummond just said, it is an intrusive questionnaire.

Can you explain to me why you think it's so intrusive?

[Niels Veldhuis]

Thank you, Mr. Bernier. I certainly can explain. I do believe that the long-form census is too intrusive.

If you look down the list of questions, they are on everything from a person's sexual orientation, whether they live with a same-sex common-law spouse or an opposite-sex spouse; how much time they spend playing with their kids; how much time they spend with their elders, be they grandparents or parents; and all sorts of questions on their home, such as whether it needs minor repairs or major repairs; their religion; and their ethnicity, whether they're black, white, Arab, or Filipino. I could go on and on.

I, at least, feel that the government really has no business forcing or threatening Canadians to provide this very personal information.

[Maxime Bernier, Conservative, Beauce, QC]

Thank you.

You're also doing a lot of studies at the Fraser Institute. Are you using the data from the long-form questionnaire?