« We're frozen at #8 on the Palgrave | Main | Saturday morning comics »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Time to start a letter-writing campaign. I'll try to enroll as many of the mathematics friends as possible.

This is just monumentally boneheaded. I assume this is predicated on the census stupidity we are now seeing in the US.

Heck, sign up your non-math friends. This isn't that hard a point to grasp.

This is what you get when you elect people who think government is the problem.

It's going to be a tough slog to fix this one, it's one of the relatively few things that Harper can point to as a win for his base.

Upon reflection, I would up the ante on my last post. Forget the letter writing campaign. Pass the hat for a lawyer. The government is required to conduct a census. See if you can make a legal argument that this form of census is invalid.

This is a most worthy issue for the 8th ranked economics and finance blog to take on!

Given the centralization of authority in Ottawa these days, this would not likely be a decision made by Statistics Canada/Industry Canada. It would probably have been made at the Cabinet level, pursuant to a Memorandum to Cabinet signed my the Minister accompanied by a Discussion Paper prepared by Statistics Canada.

As part of your campaign, you could try to get your hands on the Discussion Paper through an Access to Information request. It should set out the options considered and their pros and cons. You can then judge the quality of the analysis presented to Cabinet and the extent to which the decision is supported by that analysis.

This way you may be able to find out whether this was a route supported by the pros at Stats Canada or a result of political dogmatism, and then be able to target your campaign accordingly.

Good luck!

This was dogmatism. A number of people had complained that the questions were invasive and violated their privacy. There is still a woman litigating in Saskatchewan over this.

@Richard R.: I'm pretty sure that there is no discussion paper anywhere that justifies this move. This is a fundamental mistake, and the people at StatsCan aren't idiots.

@ Stephen Gordon

Do you mind if we cut and paste from this post for a skeleton letter to the editor?

Do you want to complain about this or do you want to change it?

Letters will do the first, it will take something else to do the second.

Cricket bat?

Can't we do both - complain and try and change it? Surely (almost) all social scientists in the country, and beyond, would be willing to sign a petition to try and change this - or at least point out how stupid a decision it was/is. Who knows how to create an online petition? Is the CEA doing anything? Can we link to the Canadian sociologists, historians, geographers....through the CEA? Why aren't provincial governments complaining about this?


Good on all of you. We should oblige good citizens to do everything that seems potentially useful to economists. I'm almost an economist. Could one of you come cut my grass?

I was not asserting that you should not complain. I was just pointing out that it will not produce change. For that, you have to go either to Parliament or the courts.

I thought this was the sort of stuff you learned in a high school statistics class... Unfortunately, the Conservatives haven't, and the consequences will be much more severe than flunking the next unit test.

And of course people never lied on or refused to properly complete the incredibly intrusive long-form census, because it was "mandatory".

A return to current procedure is fundamentally flawed and will produce worse-than-useless results. StatsCan should make lie detector tests mandatory as part of a return to a long-form census, to ensure the accuracy and completeness of all responses. After all, we need the best possible data to develop policy, and without lie detectors the long-form census would provide us with data that are at best badly flawed, and are likely to be completely useless.

no richard's right, there has to be a cabinet level document, telling stats-can to do this.

These might be deeper political waters than it would appear at first glance.

There is a raging conspiracy theory in the US over the census - largely driven by nuts like Michele Bachmann. Google "census conspiracy" - it'll shock you how crazy some people are.

Clearly, nobody at StatsCan would agree to this, so it's a political decision imposed on them. Who among the conservative party faithful would want this? Well, the same nuts that believe the conspiracy theories. So it looks to me like the PMO is pandering to the nuts. They are required to do the census, but they aren't required to do a good job.

If it is a political move by the PMO, then if any attempt to overturn it gained traction,it would be be met with aggressive spin - stuff about liberal ivory tower professors who are paid way too much do no useful work, blah blah.

In any case, I don't think there's much hope in creating public support for mandatory revelation of 'private' data to the government. Certainly not after all the fuss that's made about private data.

Sadly, I think this is a done deal and a lost cause.

BTW, this is the sort of crazy you'd be up against:


Hal Jam, I need you to illuminate the incentive that would induce the typical Canadian (not just the Western survivalist) to take the time to complete the census untruthfully. I also need you to explain how "less-than-perfectly accurate" amounts to "worse-than-useless." I'm not following.

I've been hanging around here for two years. It's interesting to note that suddenly comments start appearing from people whom I never seen. Now, maybe I have a faulty memory and I just don't remember previous comments. Or maybe people really are motivated to comment for the first time on a post about the census. Or maybe the spin machine is already at work.

@Hal Jam

It is the case that requiring people to fill out any sort of census requires demonstrating that the census provides useful information for forming public policy. While I'm not one, I would suggest that making that argument is unlikely to tax any social scientist unduly. Then the only question becomes, how do you best balance the obtaining of useful information against the burdening of the citizenry with compulsory form filling out.

Clearly this decision reflects a desire to impose a burden of next to nothing, and a willingness to accept the consequence of having no useful information from the process.

A lie detector test (which I recognize you are suggesting ironically, but I'll respond to non-ironically) would impose an enormously heavy burden (as well as being non-efficacious, since they don't work very well) and only marginally improve the data (as@Winston suggests).

Both of those shifts in the balance between burden on the citizens and usefulness of the data strike me as wrong-headed, and hard to justify.

I have no doubt that the long form census could be improved and made less onerous. As Frances's post on taxes noted, effective form writing is not a strength of this (or any) government. But to make the data statistically useless imposes a real cost on the ability to formulate effective public policy. As noted, I think that places the balance at the wrong point.

I would note, as a hard core geek, I rather enjoy filling in the long form census, and am a bit disappointed if I only get the short form. I want to be counted, dammit. How's that for an example of the problems of self-selection bias?

If not an discussion paper, a memoranda to cabinet should be kicking around. Section 2 of the pco's guide to making acts covers making policy.

nevermind, I see you weren't disagreeing that the memorandum to cabinet existing, but the discussion paper. But there will be reports generated to give alternate policy options.

Now Clement's office is saying that they'll consider drawing up legislation to release the replacement survey results in 92 years to silence the geneologists... wonder how that will affect response rates (for this and other Statscan surveys).

Clement's office also stated that they are "just beginning" to plan the 2011 Census -- for such a complex operation, is that not leaving it rather late?

What ever happened to the independence of Canada's stat office, something that the UK and other countries are (were?) striving to emulate?

Would Census data not be used to calculate equalization payments, and to derive samples for household surveys that take place after the census is tabulated? Is this the beginning of something even bigger?

I would say that given the HST and other incentives to the underground economy, that now is the time for a Census of Businesses, along with the Census of Population. With the increasing reliance on tax data for business surveys, I fear for the quality of GDP estimates -- and just look at what the Conference Board's error on the China LEI did to the markets this week.

It disappoints me that almost no one (least of all from Statscan) seems to be studying issues like income inequality and business concentration in this country -- Veall's top 1% finding might well be explained by increased concentration in big businesses, along with relatively new compensation methods such as stock options.

At the same time, my public service friends tell me that Statscan is so bureaucratic, complex and expensive that they are completely unable to respond to data needs of key clients (including Clement's shop at Industry). It's quite sad to see professional-looking on-line surveys being done by high schoolers, while Statscan takes over a year and $300,000 to come up with something that does not work near as well. They used to have a mantra of "data stewardship", which now seems to be changing to "data mashup" (perhaps it has a generational explanation).


@ Winston

When I have filled it out, I've filled it out with ridiculous information. When someone knocks on your door in a free country and demands you fill out a form with private information, this is incentive enough for some people to lie on the form. The fact that noone in this discussion even seems cognisant of the fact that people don't like being ordered around seems just bizarre to me. You're social scientists, right?

I understand that economists find personal and private data very useful - I use it myself, and find Canada's privacy laws inconvenient when doing research. In fact, when companies ask for my info, I sometimes provide it to them. But they ask. They don't demand.

I can certainly attest to the fact that people don't even like filling out the short form let alone the long form. If my experiences with the 2006 census are indicative of anything, people of lower income brackets (just identified by neighbourhood, so nothing scientifically rigorous) won't fill it out and many don't trust you once they see that government badge.

Considering that I was specifically ordered not to mention that there were laws making it compulsory, I think they were pretty aware of how delicate it was to demand personal information from people at home. This only seems like the next step in that regard.

It's too bad that people do not seem to appreciate that there are benefits to providing the government with information.


I was wondering what kind of wingnut theory might have inspired this. This is a good one, but I was wondering if maybe it had something to do with census results being used as an input in a "socialist calculation": http://en.wikipedia.org/wiki/Economic_calculation_problem

Notwithstanding the clear statistical rationale for making completion of the long-form census mandatory, I'm not entirely unsympathetic to those who say "I shouldn't have to give the government detailed information about my life".

After all, one of our fundamental rights is the right not to talk to the government (protected, in various ways, by sections 2(b), 8, 11(c) and 13 of the Charter of Rights). While many violations of those rights can be justified under section 1 of the Charter and are therefore permissible (the government, after all, can compel you to testify in court or complete your tax returns or show your driver's license when you're driving - though not in other circumstances - on pain of imprisonment), that doesn't take away from the fact that those are recognized as violations of our fundamental rights (they're just "permitted" violations).

In that light, the argument that "there's a benefit to providing the government with this information " while it is true, entirely misses the point. It's analogous to the arguments we heard last week with respect to the Ontario regulation which purported to require people to identify themselves to police and to show ID (but which, in fact, apparently didn't despite police and government (mis)representations to the contrary), that it was essential in order to provide security and that if you had nothing to hide, you had no reason not to show ID (i.e., there's a public benefit to doing so). Even if true, that doesn't take away from the fact that the regulation (as described by the powers that be) was offensive precisely because it purported to compel Canadians to provide information to the government.


Just listing a bunch of stuff from the charter may give the appearance of adding weight to the argument, but on closer inspection:

2 (b) is : freedom of thought, belief, opinion and expression, including freedom of the press and other media of communication

8 is: Everyone has the right to be secure against unreasonable search or seizure.

11 (c) is: Any person charged with an offence has the right (c) not to be compelled to be a witness in proceedings against that person in respect of the offence;

13 is: A witness who testifies in any proceedings has the right not to have any incriminating evidence so given used to incriminate that witness in any other proceedings, except in a prosecution for perjury or for the giving of contradictory evidence.

We can ignore 11 (c) and 13 since we aren't talking about people being charged with crimes. And the census isn't infringing on anyones exercise of 2. So that leaves 8. The census is required by law, and as a practical matter some sort of census is obvious required for representative democracy to function. Though I don't buy it, I'll concede that there may be an argument that the compulsory census long form violates 8.

If that's the case - and if there is a serious question on this point I'd like to see it sorted out in the courts - then let's save the money and not bother with the obviously stupid optional survey. Just do the basic count required for seat apportionment and call it a day.

"Long gun registry / long form census. What's the diff in Red Deer or Fort McMurray? More oil sands labourers, dammit. We don't care of their background!"

Bob, I'd imagine that those violations are permitted because they're associated in some way with using a government sevice. For instance, the police can compel me to show a drivers licence while driving on public roads or land, can they compel me to show a licence if I'm driving on private property? (courts are associated the with the law enforcement service that we all benefit from, taxes pay for all beneficial services and could be associated with our use of government tender as a means of exchange.)

Whether or not you can be compelled to give the information should be decided on whether you are a beneficiary of some associated service. Presumably the applications of census data to correct problems in society could certainly be argued to benefit everyone.

@Guillaume: Maybe it's my own tin foil hat malfunctioning, but I strongly suspect it's a politically low risk way of throwing a bone to the crazies in the party just like the maternal health/abortion thing.

Why oh why can't Canada have a better opposition?

Patrick, that is an interesting point. Does a census questionaire constitute a search or a seizure?

"With the best Constituency Information Management System (CIMS) of all federal parties, do you think we're going to make similar information available to the opposition? What, do you think we're stupid or something? (ok, don't bother answering that)"

A census questionnaire might constitute a search and seizure, but if it did I think it would easily be justified under s. 1. And at that point Bob, as you know, the question of the benefit obtained from the census is very much relevant. Indeed, if one is putting it in a Charter context I think it is specious to suggest that pointing to the benefit obtained from the census data "misses the point". There are very few "rights" in Canada that are not violable in some circumstances. The only question is when. And when is determined in a significant basis on a cost/benefit analysis under the Oakes test (is the violation minimal, is it justified etc.). I might not like that, and some might view the violation as nonetheless offensive, but it mischaracterizes the nature of the legal rights that we have to suggest that saying "it's my right!" answers the question of whether something is legally inappropriate.

I would also note, in response to Adam, that there are all kinds of circumstances when a person is required to give information and will receive no associated service. If I have evidence of a crime I can be compelled to testify about it, and in so testifying can be cross-examined about other information, such as my criminal record. If I am sued by someone I can be required to disclose all sorts of information I would rather keep private.

Alice, the argument I was making was basically that courts are allowed to compel testimony exactly because everyone, including the person being compelled, benefits from the government service of enforcing law and order. The person giving testimony may benefit indirectly of course but everyone who lives in the society can be taken to be consumers of the law and order that the government provides. (In the case of my being sued, all of society, including the person who is currently being sued, benefits (presumably) from the possibility of courts being able to forcibly extract the information necessary to decide legal disputes).

The point really is that the argument I was trying to make is exactly the one you're making in your first paragraph, so we are in agreement.

I was glad to see the end of the long form. Every fifth home got a long form, and it was checked by your neighbour who was completing the census. I complained about the lack of privacy and was told that I could send my form into the regional centre, but it turns out that the regional centre simply sent the completed form back to my neighbour, for checking.

This may be a non-issue in urban areas but in rural areas, where everyone knows everyone, it was much more than just an issue of having to provide too much information to the government. It was having to provide too much information to your neighbour.

(off on somewhat of a tangent.. I've been sick, so blame the meds if you want...)

What I find ironic about all this is that it involves many of the same people who introduced Ontario's 'sunshine laws' on public sector salaries. Talk about providing a lot of information to your neighbour!

Ironically, the sunshine law is probably the best thing that ever happened to B-school academics, as it made getting a big raise easier at negotiation time, since it's really easy to make a 'well, if you're paying Jones 120, I must be worth at least 130!". Baseball *fought* public disclosure of salaries for the longest time, for that very reason.

As this is a minority government, perhaps the Opposition could gang up on the Tories on this. The case would be that lack of detailed information, particularly economic information, impedes social policy, particularly government spending.

Provinces should cry foul over any attempt to alter their fair share of federal transfers.

Interestingly there was a "West Wing" episode early in the second season in this issue.

Also last time I got the long form and answered that I was a Type I Diabetic. I was selected for supplementary questions in a phone interview. They only annoying thing was that I was at university and they kept calling my home. Drove my mother batty. On the other hand, their questions on diabetes were intelligent, thoughtful and concise. There is so much about diabetes that isn't intelligently understood by most people that I was duly impressed.

Patrick, you might benefit from reading some of the jurisprudence associated with those sections, rather than just reciting the text of the charter. In your defense, it's my fault, I failed to mention section 7, in which the court has found a generalized right to privacy (section 7 provides that persons shall not be deprived of life, liberty and security of the person, except in accordance with principles of fundamental justice - it's sufficently broad as to enable the courts to read a great deal into it). Section 8, 11(c) and 13 are considered to be mere specific examples of that generalized right (this is obvious in the context of search and seizure - what use would section 8 be, for example, if the government was permitted to compel Canadians to tell them, for example, the contents of their house?). As for section 2(d), you should know that freedom of expression includes the freedom not to say anything.

There is, in fact, a fairly significant body of jurisprudence dealing with situations where the government can compel Canadians to produce information about themselves (in the criminal context, but also in the regulatory context - i.e., the tax act, driving license, securities act, environmental areas, etc.). While the courts have generally permitted the government to infringe on the privacy rights of Canadian in those circumstance (and, just for the record, I suspect that they would do with respect to the census too, though I don't know if there's any actual caselaw, the point is that they recognize it as a infringement of that right.


You're close. Typically, the heavy lifting on justifying the violation of one's privacy right is done on the basis that, because of the nature of the expectation, a person has only a minimal expectation of privacy. So, for example, the police can ask to see your driver's license when they pull you over because, by virtue of having engaged in a regulated activity you have implicitly accepted a reduced expectation of privacy . Similarly, the courts have concluded that people have minimal expectations of privacy in their tax and business record. The information sought in the long form census is, of course, far more intrusive and goes far more deeply into the core of the information that a right to privacy is intended to protect, so the justification would have to be greater.

In any event, the point isn't whether the requirement to complete the long-form census form violates the charter, I don't think it does because it's probably either a violation "in accordance with the principles of fundamental justice" (under section 7) or a violation which is "reasonable and demonstrably justifiable in a free and democratic society" (the test under section 1), but that it does nevertheless violate the right to privacy which is one of our fundamental rights. That's a right that we should take seriously and shouldn't dismiss simply because it intrudes in our ability to come up with accurate datasets for socio-economic research. I mean, we're in a world of trouble if only people who take that seriously are just people like me and "census conspiracy" nuts! Certainly, it's a weird world where the Tories are standing up for fundamental rights and the country's social scientists are criticizing them for doing so.

And I'd think you might be surprised at who doesn't like filling out the long-form census. Jack identified a real problem in rural areas. I would imagine in urban areas like Toronto, Vancouver or Montreal, which have significant populations of recent immigrants, there's probably a significant subset of people who fled from countries where governments do all sorts of nasty things with the information they collect about their citizens and who aren't all that keen to provide such information here (you might say, "sure, but Canada's not Iran", to which the answer would be "exactly"). And then there are just people who are worried that the government might use their census data to cross-reference it against other information they've given the government (income being the obvious example - anyone know someone who plays fast and lose with their taxes?).

Bob: You brought it up. I mainly wanted to make a point about sound argumentation, not about law. IMO, all but one of the charter sections you mentioned were irrelevant. But I am not a lawyer, and pretending to be on usually just makes one look stupid. I defer to Prof. Woolley's comment.

Clearly a census is required for the functioning of representative democracy, and in the context of the modern welfare state I think the more detailed long form census sampling is also reasonable and necessary. I'm not at all worried about privacy, and I think the privacy issue has become ridiculous in Canada. People have conflated privacy with anonymity. There is no right to anonymity, and for good reason.

In any case, if for whatever reason the Gov't and/or the public doesn't want to do the long form, then fine; don't do it. But don't pretend that the option web survey thingy is just as good. Do the basic census required for seat apportionment and then be done with it.

I can think of at least 2 ideological & partisan reasons the Cons are doing this, quite apart from appeasing their libertarian base's own personal resentment of being req'd to do anything for their country:

1) One of the most important uses of this more detailed info. is in social policy, particularly in population health issues, to combat the effects of poverty, etc. But arming those pesky Social Planning Councils & the like with good data just leads them to asking for better social programs like subsdized day care etc. etc., all the time, which req's more money -- read: more taxes. So this will cut them off at the knees, & the Con's can go on merrily trumpeting their own anecdotal "data" (crime up, gun registry useless, plenty of jobs for anyone, only reason for anyone to be hungry or a troublemaker is laziness). Similarly on environmental & food safety issues (data? we don't need no stinkin' data). The Cons have been quietly & systematically going about dismantling policy research capacity both within gov't & in nonprofits as part of their overall goal of whittling gov't down to nothing but military & police forces (& pork-barreling to reward friends & stay in power).

2) Another important use is for fine-grained demographic information at the postal code level that is extremely useful for fundraising, which they no longer need since they've exploited it to the full & now have a great database of their own, and which they want to deprive those pesky nonprofits & especially their rival political parties of.

If the long form census is so great, tell me how many of the 33% getting it you think will respond? I bet it will be close to zero. You don't need to explain to me why that's not an airtight argument against it, but I think it's worth considering.

Add me to the list of people who are happy about this. I do pretty much nothing illegal, and my antipathy to the long-form census is based only on this: I don't know you, I don't trust you, and the vast majority of my life is none of your business. Where "you" refers to the government of Canada acting on behalf of its citizens, as well as the individual actors in the government who will have access to this data.

Good news: further to my 2nd point about that data being a goldmine for fundraising & campaigning purposes(which is why the Cons want to blow it up, along with the vote subsidy, so they can turn this into a one-party state), it seems the business community has finally gotten wind of this & will be opposing it, because they use the same data for marketing purposes. See:


The US census bureau did the same thing: the census long form was dumped and the information is now being collected from the American Community Survey. On the other hand, the ACS only tries to get 1% of the population to respond, so they can put more effort into follow-up.

"One of the most pressing policy questions facing Canada is how best to develop our skills and improve upon education attainment levels. We need the best possible data to develop policy, and this measure would provide us with data that are at best badly flawed, and are likely to be completely useless." -SG

Maybe the Conservatives reckon that a resource-based economy only needs on-the-job acquired skill sets?

I started a petition opposing the Canadian government's move to scrap the census. You can sign here: http://www.gopetition.com/online/37527.html


I am a seniors' housing market analyst in BC. We rely heavily on Census data (for which we pay thousands of dollars) for our market studies and for our community seniors' housing needs assessments. I have already written to Minister Clement, the PM, my MP, the Globe, and all of my colleagues who are in any way dependent on Census data for their work. What is happening is completely appalling. Kate

"People have conflated privacy with anonymity. There is no right to anonymity, and for good reason."

That may be true, but that's really beside the point. Disclosing detailed information about yourself and your family goes to the core of privacy, not anonymity.

You're missing the point. Everyone knows this census data is very useful. BUT, the bureaucrats at the Canada Revenue Agency have similar privacy requirements, employees are carefully vetted, etc.

And yet Access to Information requests regular show that CRA employees are routinely accessing information that they shouldn't. For personal gain, for their own amusement, to gather information on creditors, or their spouse/ex-spouse.

There was one case in Oct 2009 where a CRA employee accessed 37500 emails and 776 documents about ordinary Canadians.

How many of these bureaucrats were convicted & sent to prison? None.

And the innocent Canadians whose privacy was violated? Sorry, too bad.

Why should we believe Statscan is any different?

Because there aren't any similar cases involving StatsCan employees?

Because there aren't any similar cases involving StatsCan employees?

Cases that you know of. We only learned of the CRA cases due to Access to Information requests by some tenacious journalists.

Given the propensity of many levels of government to stonewall & cover up wrongdoing, the fact that misuse of Statcan data hasn't been in the news doesn't mean much.

In fact, I recall during the 2002 census there was some bizarre mix-up and hundreds of completed census forms were mailed to other people's residences.

Now, if Statscan's (and other government departments that routinely deal with very private information) security procedures were audited in detail, and if bureaucrats who abuse their authority are fined, lose their jobs, and go to jail for a long time, I might agree with you.

Lemme see if I've got this straight. You're so worried about a hypothetical situation for which their is no concrete evidence that you'd just toss out the whole thing despite the obvious benefits? Uh huh.

Lemme see if I've got this straight. You're so worried about a hypothetical situation for which their is no concrete evidence that you'd just toss out the whole thing despite the obvious benefits? Uh huh.

Given that this sort of thing regularly goes on in other government departments, even then, it is rarely caught, and even when it is caught, they get a slap on the wrist. So I am concerned about misuse, fraud & theft. Any reasonable person should be.

If you owned a business that regularly handles large amounts of cash, you would probably take precautions. You would have regular cash counts by different people, strong locks, safes, alarms, security cameras & audits.

You would likely do these prudent things even if you have never been robbed (either by your employees or by outsiders). If there was a theft, you would want them prosecuted and sent to jail, pour encourager les autres.

You might even do something like reducing the amount of cash you keep on hand, to mitigate the size of a theft.

These are the same government bureaucrats who couldn't account for $15 million of gold at the Royal Canadian Mint. Oddly enough, one of the first things Mint officials did was conduct some opinion polls and hire a PR firm. Eventually, the RCMP determined that no theft occurred, but that the record-keeping practices & internal controls of the Mint had been terrible for many years.

So yes, I do have difficulty trusting in the competence & honesty of government bureaucrats.

Given the available evidence, and the lack of consequences for those who misuse government data, I can't see how any reasonable person wouldn't be concerned.

You're becoming incoherent.

What does an administrative snafu at the mint have to do with the claim the StatsCan employees are incompetent and corrupt? Totally irrelevant.

And if you think the private sector is any better than the public on privacy, you're living in a dream world.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad