Evidence of meeting #10 for Procedure and House Affairs in the 41st Parliament, 1st Session. (The original version is on Parliament’s site, as are the minutes.) The winning word was estimates.
A recording is available from Parliament.
11:25 a.m.
Chief StatisticianStatistics Canada
I do not distribute or use the data, this is done by the committees responsible for establishing electoral boundaries. I provide the data.
I agree that within a province, we are going to be using a data set that is different from the data used to—
11:25 a.m.
Liberal
11:25 a.m.
Chief StatisticianStatistics Canada
In my opinion, if the objective is to do the best distribution possible between the provinces and the territories, it is preferable to use the most accurate data. If such data does not exist at a lower level, I do not see why it is problematic to use the available data.
11:30 a.m.
Liberal
Stéphane Dion Liberal Saint-Laurent—Cartierville, QC
You come with statistics on the entire Canadian population, and it seems to me that this is what needs to be done. However, in its count, the government excludes the three territories. It is therefore basing itself on the population of the provinces in order to distribute the seats amongst the provinces. The only statistical effect of this methodology is that it deprives Quebec of one seat.
Do you think that there is a statistical reason—and not a political one—justifying the fact that the government, in its own count, establishes the proportion for the provinces without considering the three seats from the three territories?
11:30 a.m.
Liberal
Stéphane Dion Liberal Saint-Laurent—Cartierville, QC
It could be done based on the entire population.
From a statistical point of view, it would be possible to allocate the seats to provinces and take into account the three territories that exist and that have one seat each. You don't see any statistical problem with that?
11:30 a.m.
Chief StatisticianStatistics Canada
It's not a statistical question. The government, in deciding how to go about allocation of seats, establishes a formula. The nature of that formula is not a statistical question. I really honestly can't take a position to answer you.
11:30 a.m.
Liberal
11:30 a.m.
Liberal
Stéphane Dion Liberal Saint-Laurent—Cartierville, QC
No, I think I have my answers. They're worrying me, the answers I received, but thank you very much for that.
November 17th, 2011 / 11:30 a.m.
Conservative
Scott Reid Conservative Lanark—Frontenac—Lennox and Addington, ON
Four minutes. You could give me Mr. Dion's minute and that would be fine.
11:30 a.m.
Conservative
11:30 a.m.
Conservative
Scott Reid Conservative Lanark—Frontenac—Lennox and Addington, ON
I think Mr. Dion had an interesting point. I've often thought it odd that the formula involves working with the number of seats for the provinces rather than all the seats. That's a peculiarity. I don't know if it crept into the system with the 1985 formula or if it's simply an artifact of the original much-changed and ever-changing formula that existed before. It is an oddity that we've inherited. It's a legislative oddity, as opposed to a statistical one.
Mr. Smith, I went through the numbers you had there. I took column A and column B and did a quick comparison. I see that they produce fairly minor shifts when we look at the raw numbers for the provinces, most provinces. Then there are significant shifts for Quebec, Ontario, B.C., and Alberta. This is just me doing it by hand.
The population difference in Newfoundland between column A and column B is only 4,000; P.E.I., 3,000; Nova Scotia, 21,000; New Brunswick, 20,000; Manitoba, 29,000; Saskatchewan, 13,000. There is a 105,000 difference in Quebec; 526,000 in Ontario. Column A is always bigger than column B. Alberta is 85,000; British Columbia, 197,000.
Does that mean, effectively, that if we use the numbers in column A as our basis, which is what I think the legislation proposes, versus column B, the status quo, which is what the 1985 formula would propose, and all other things are equal, we would get more or less one extra seat for Quebec, four or five seats for Ontario that wouldn't otherwise be given, and then a seat each for B.C. and Alberta? Is that more or less the practical impact of this?
11:30 a.m.
Chief StatisticianStatistics Canada
I'm not comfortable answering, because the application of the formula is not a responsibility of Statistics Canada. I'm not sufficiently conversant with it to give you an expert opinion in that regard.
11:30 a.m.
Conservative
Scott Reid Conservative Lanark—Frontenac—Lennox and Addington, ON
I realize it's not a statistical question, but I guess that's the practical impact of using this. if you are assuming about 110,000 people per seat, you'd wind up with something like that. It's not a statistical question, but am I wrong...?
11:30 a.m.
Chief StatisticianStatistics Canada
If I can take as given that the divisor that creates the seat is 100,000, then given those discrepancies I would arrive at the same conclusion.
11:35 a.m.
Conservative
Scott Reid Conservative Lanark—Frontenac—Lennox and Addington, ON
So using the new formula reduces the degree to which the larger provinces are underrepresented and the smaller provinces are overrepresented. That would essentially be the practical result once column A is chosen, the preliminary post-census estimates based upon the population counts.
11:35 a.m.
Chief StatisticianStatistics Canada
My way of putting it is that it would advantage provinces where we have a more significant undercoverage problem, by increasing their population; therefore, using the formula that we were just discussing would lead to greater seat numbers. We missed more people in those provinces, and it's compensating for that fact.
11:35 a.m.
Conservative
Scott Reid Conservative Lanark—Frontenac—Lennox and Addington, ON
Is that because people are more mobile in those provinces?
11:35 a.m.
Chief StatisticianStatistics Canada
This is where the major urban centres are. The downtown core, large urban centres—that's where we miss people. That's where the problems are. These are the ones that are disadvantaged if you use the unadjusted counts. It's compensated if you use the adjusted counts.
11:35 a.m.
Conservative
11:35 a.m.
NDP
11:35 a.m.
NDP
Chris Charlton NDP Hamilton Mountain, ON
Thank you very much.
Thank you very much for your presentation. I'm particularly interested in the comment you made about underrepresentation and undercoverage. I may simply not be understanding well enough, so let me pose a couple of questions, and I may go elsewhere, depending on your answer.
When you talked about why there is undercoverage or underrepresentation, you pointed to very specific factors. Underrepresentation would not, for example, take into account economic trends. There was a point in time when Saskatchewan was losing population at a pretty dramatic rate, and then the economy turned around and it gained population at a very dramatic rate. Those kinds of factors aren't at all part of your estimates for the future. Is that right?
11:35 a.m.
Chief StatisticianStatistics Canada
We're not projecting here; we're looking backwards. They are taken into consideration. We do measure interprovincial migration, and the population estimates reflect everything we know about interprovincial migration. For example, one of the techniques we employ is to take different generations of the tax files, find somebody in one file, and then see where they were in the next taxation year. Have they moved? That information is used to help us estimate interprovincial migration.
In fact, the estimates that are currently available do reflect the best information we have on interprovincial migrations, which are usually a result of economic considerations.
