Procedure and House Affairs Committee on Nov. 17th, 2011

[The Chair, Conservative, Joe Preston]

I call our meeting to order, and we'll get started today.

We have two sets of witnesses and a little bit of committee business to do at the end of the meeting today, so we're going to just see if we can find a little bit of time out of each of the witnesses.

We're here today, still pursuant to the order of reference of Thursday, November 3, 2011, on Bill C-20, An Act to amend the Constitution Act, 1867, the Electoral Boundaries Readjustment Act and the Canada Elections Act.

We have as our first witness today Mr. Wayne Smith, our chief statistician.

Mr. Smith, I understand you have a set of opening remarks. I'd ask you to introduce your colleagues today, make your opening remarks, and then we'll get to questions.

Mr. Lukiwski.

[Tom Lukiwski, Conservative, Regina—Lumsden—Lake Centre, SK]

Perhaps I just haven't seen it, Chair, but I was wondering if Mr. Smith has a written copy of his remarks and if they've been distributed.

[The Chair, Conservative, Joe Preston]

Have we distributed opening remarks?

No?

[Wayne Smith, Chief StatisticianStatistics Canada]

No.

[The Chair, Conservative, Joe Preston]

Okay.

It's a great move. You'll make the members pay attention, rather than reading ahead.

Mr. Smith, it's yours. Thank you.

[Wayne Smith]

Thank you, Mr. Chairman, for this opportunity to address the committee in relation to its study of Bill C-20.

Today, I am accompanied by two officials from Statistics Canada who are experts in census and population estimate methodologies, namely Mr. David Dolson, Director of Social Survey Methods, and Ms. Johanne Denis, Director of the Demography Branch.

As I think you know, Statistics Canada's role in the readjustment of electoral boundaries is in the supply of population data to support, first, the application of the formula for allocation of seats to the provinces and territories, and, second, the delineation of electoral districts within provinces and territories. For the purpose of delineation of electoral districts within provinces and territories, there is only one source of population data that provides the necessary detailed geographic breakdowns, and that is the census of population, which is conducted every five years.

For the purpose of allocation of seats between provinces and territories, there are two alternative sources of population data that could be employed. The first source, and the one that has been used in the past, is the unadjusted population counts from the decennial census of population. Statistics Canada will publish counts from the 2011 census of population on February 8, 2012. The second alternative source is Statistics Canada's population estimates program. This program produces annual and quarterly estimates of the populations of the provinces and territories. Estimates in this program reflect at any given point in time all of the information that Statistics Canada possesses in order to provide the best possible evaluation of those populations.

Bill C-20 proposes, in a departure from previous practice, to use the currently available estimates of provincial and territorial populations at July 1, 2011, for purposes of calculating the allocation of seats between provinces and territories. These estimates reflect results of the 2006 census adjusted for net undercoverage, augmented by births and immigration since the census date and reduced by deaths and emigration.

Given that the objective of Bill C-20 is to launch the readjustment process at this time, the relevant statistical issue for consideration by the committee is which of the two alternative measures of the populations of the provinces and territories is likeliest to be the closest to the true value: the currently available population estimates or the unadjusted 2011 census of population counts that will be released in February. To answer this question, the census counts and the current population estimates need to be compared to the definitive estimates of the 2011 population that Statistics Canada will produce in 2013. These will reflect estimates of net undercoverage of provincial and territorial populations from the 2011 census of population to be generated by studies that are currently under way but not available.

Let me explain briefly the key notion of net census undercoverage. Official statisticians in all countries know that a census of population, however well conducted, will miss some people while counting some others twice. Statistics Canada, after each census, conducts a statistical study of these two effects.

Estimates from the 1996, 2001, and 2006 censuses indicate that net undercoverage, because we miss more people than we double count, is typically on the order of 2% to 3% of the population counts in the Canadian census. We cannot know at this time what the level of net undercoverage will be for the 2011 census of population—the necessary study, as I said, has not yet been completed—nor can we definitively know whether estimates of natural increases and migration that underlie the population estimates will be confirmed.

The best guide, therefore, to answer the question of which of the currently available population estimates or the unadjusted 2011 census population counts will be closest to our definitive estimates is to look at what has happened in previous censuses. Having done this work, I can inform the committee that the population estimates for the provinces and territories available at the time of the release of the census counts have typically been substantially closer to the definitive estimates than the unadjusted census counts themselves.

To demonstrate this, I have prepared a table, which I think you have in front of you, based on the 2006 census, that looks essentially at the situation as it unfolded for the 2006 census. The table compares the unadjusted 2006 census population counts and the population estimates published in September 2006, which is essentially the same generation of estimates that we're talking about right now for 2011, to the adjusted 2006 census population counts that were published in September 2008.

At the Canada level, the population estimates published in September 2006 were 0.3% higher than the definitive population counts, while the unadjusted counts were 2.8% lower. As at the Canada level, at the provincial and territorial levels, the population estimates were invariably significantly closer to the definitive population counts than the unadjusted counts were.

In summary, even with the release of the 2011 census unadjusted population counts on February 8, 2012, it is Statistics Canada's view that the currently available estimates of population at July 1 represent the best available evaluation of the population of the provinces and territories that is available at this time or that will be available on February 8. It is therefore appropriate, in our view, that they should be used for the purposes of Bill C-20.

Thank you.

[The Chair, Conservative, Joe Preston]

Thank you for that. We'll go to rounds of questions.

We'll start a seven-minute round with Mr. Lukiwski.

[Tom Lukiwski, Conservative, Regina—Lumsden—Lake Centre, SK]

Thank you very much, Mr. Chair.

Thank you for being here, Mr. Smith. Your presentation was informative for me because I think the largest question all of us have is based on population estimates and their accuracy.

The formula the minister explained to all of us is based on the population estimates. If we are to do a relatively accurate redistribution, or in this case increase, of seats by population, we have to be assured that the numbers we're using are accurate.

I have some confidence based on your presentation, but I would like to get a little more information and perhaps probe a little deeper, if I can, to give not only me but other members of this committee some confidence that the population estimates, as opposed to the census data, are the figures we should be using for the purposes of Bill C-20.

You've given us a chart based on the 2006 population from the census perspective in the population estimates. You've stated, and it shows on the graph, that the population estimates are, I think in your words, invariably more accurate.

Why are the population estimates more accurate? Is it because you've designed a better formula, or is it because perhaps the census data taken only comes out once every few years?

I think it's important for us to know why we can count on population estimates and should be counting on them in terms of accuracy of population counts both in Canada and across the provinces.

[Wayne Smith]

There are two elements to the response.

One is that for the purposes of the estimates of population, we have very rich information sources in Canada. We have very accurate information from vital statistics on births and deaths and on immigration to the country. We have extremely useful administrative files that allow us to capture interprovincial migration. The issue ultimately comes down to whether or not the data has been adjusted for net undercoverage and the normal size of that adjustment.

Given the precision of the measures that go into the population estimates, and given the fact that the population estimates were adjusted for the undercoverage in the previous census, that almost guarantees in Canada, and certainly has for multiple censuses now, that the estimates are going to be closer because they reflect an adjustment for undercoverage, rather than the unadjusted counts that we publish at the beginning of February. The difference of the people that we missed on the net basis in the census is on the order of 2% to 3%. That difference is sufficiently large that any possible error in the estimation process is exceeded by that error.

It's a very technical explanation, but that's the essence of it.

[Tom Lukiwski, Conservative, Regina—Lumsden—Lake Centre, SK]

I appreciate that.

It seems fairly obvious, but just to make sure we're crystal clear on this one, the essence of what you're saying is you are convinced that if our government used population estimates in calculating any possible allocation of new seats, that would be a more accurate assessment of the population from province to province than if we used the census data.

[Wayne Smith]

That is absolutely our view.

[Tom Lukiwski, Conservative, Regina—Lumsden—Lake Centre, SK]

Thank you.

Let's talk for a moment on the undercoverage. I'm a little unsure of some of the things you talked about.

Before that, let me ask you, with respect to the information you have examined over the years, because you said you've been examining this for several years now, are there any specific demographic groups or any particular areas of Canada where there seem to be more problems with undercoverage than others? Why would that be if there is such a pattern?

[Wayne Smith]

The groups that tend to be subject to undercoverage particularly are young people, young adults particularly, notably young males. Recent immigrants are another group.

[Tom Lukiwski, Conservative, Regina—Lumsden—Lake Centre, SK]

Could you define young males?

[Wayne Smith]

The 20 to 35 age range is where we see we have.... The issue comes down to really one of mobility, the fact that they are very difficult to find, and uncertainty in households where Johnny is at university or at a labour camp in northern Ontario—whether he is being counted where he is or whether they should put him on the form. We tend to miss some, significantly.

Because recent immigrants are often living in unusual living circumstances, that makes it more difficult for us to find them, and also because of a certain degree of discomfort some recent immigrants have, given their personal experiences in dealing with governments. Language issues as well cause a problem for us with recent immigrants.

People who obviously don't have a well-established place of residence are a problem. Certainly we see problems with people who are in low-income groups. Urban aboriginals, for example, can be a problem. Homeless people are obviously a problem for us.

Essentially, it's a question of mobility. People who are very mobile and don't have a strong attachment to a fixed residence are the ones we are most likely to miss.

[Tom Lukiwski, Conservative, Regina—Lumsden—Lake Centre, SK]

Based on what you're saying, am I to conclude that there is no real pattern, no real specific either demographic group or province that you'd point to and say this may be an area or a group where there is more undercoverage than the average, or less?

[Wayne Smith]

Well, I guess there is a demographic group—young males of 20 to 35 years—and certainly recent immigrants. Those are demographic groups where we do see a higher incidence.

Geographically, the problem—because of what I just said—tends to be focused in urban cores, large urban areas. Provinces, therefore, that are very urbanized are potentially more subject to....

I'd like to ask my colleague to comment on whether we see that provinces that are more urbanized tend to have a higher rate of net undercoverage, but geographically, that's the usual relationship.

[David Dolson, Director, Social Survey Methods, Statistics Canada]

Consistently, census after census, we observe higher undercoverage rates in large urban centres across the country.

[The Chair, Conservative, Joe Preston]

Thank you, Mr. Lukiwski.

Mr. Christopherson, you have seven minutes.

[David Christopherson, NDP, Hamilton Centre, ON]

Thank you very much, Chair.

Thank you very much, Mr. Smith and your colleagues.

Please don't take this question the wrong way, and I say that to the government members opposite and to yourself. I am merely asking this question for clarification; I'm not setting anything up.

My question is this. Who do you report to?

[Wayne Smith]

I report to the Minister of Industry, who is also the minister responsible for Statistics Canada, Minister Paradis.

[David Christopherson, NDP, Hamilton Centre, ON]

Right, so you're part of the bureaucracy, as opposed to any kind of an independent like the Auditor General.

[Wayne Smith]

Statistics Canada has always been set up as a department of government reporting through a minister to Parliament.

[David Christopherson, NDP, Hamilton Centre, ON]

Okay. That's fine, thank you. I appreciate that. I just wanted to be clear.

Did you make the recommendation, first of all? Did this idea come purely from an analysis of numbers? Was it your recommendation to the minister that he look at changing the formula?

[Wayne Smith]

No, it was not our recommendation. We were asked simply to provide information about the relative merits of the different sources.