If Canada used American racial categories...

The US Census Bureau counts people by race: white, black, Asian, and so on. Statistics Canada counts people by "visible minority status" - we don't like to mention words like "race" or "white" up here.

The other day I found myself wondering - if Canada used American racial categories, how would the demographics of the two countries compare? 

The above table shows how Canadian visible minority categorizations match to US racial groups. The mapping from one to the other is not precise. In the US, information on race is gathered from the question shown on the right.

There is no category for Arab or West Asian, and it's not clear how people of these origins answered the Census question. In 2010, a group of Arab and Persian Americans launched a check it right campaign, urging people to "check the "other" box on the census form and write in their true ancestry." Unfortunately, this was a somewhat futile gesture, as according to the US census documentation write-in entries such as "Lebanese, Arab, Moroccan, or Caucasian" are coded as white. As far as I can make out, unless an Iranian or Afghani identifies themselves as "Asian" (which some might have done), they are included in the white count.

Latin Americans provide another challenge. Many have some mix of European, Aboriginal and African ancestry. The best that I could do was to assume that Canada's Latin American population has the same mix of whites, Aboriginal people and blacks as the US Hispanic population, and allocate Latin Americans across the appropriate categories.

The final result of these calculations is shown in the table below. The "white" figure for Canada is most comparable to the US "single race" figure, whereas the figures for American Indians, Asians and Blacks are best compared to the US "including multiple races" figure, as they include people with European ancestry. It's important to bear in mind that the National Household Survey was a voluntary survey, and its coverage of various demographic groups appears to be somewhat uneven, the numbers should be interpreted with care. [Updated] 

There are no big surprises here. Canada is a whiter country than the US. We have a much larger Asian population and a much smaller Black population. Interestingly, Aboriginal Canadians are a substantially larger share of Canada's population than American Indians are of the US population.

These facts have a couple of policy implications. First, relative to the US, Canadians need to focus more on policies to support Aboriginal people's economic success, in part because they are a relatively more important part of our labour market, especially in Saskatchewan, Manitoba, and the North, and in part because they face so many struggles.

Second, an employment equity or affirmative action policy targeting "minorities" will work out quite differently in Canada and the US. In the US, the minority population is largely made up of Blacks and non-white Hispanics. In Canada, it is predominated by Asians. Research by Mikal Skuterud (here) and Krishna and Ravi Pendakur (here) has documented the enormous heterogeneity in visible minority experiences in Canada. Policies that are appropriate for one group, such as language training and credential recognition for recent immigrants, may be less helpful or necessary for others.

Basically there's nothing really profound or deep here. I just wanted to know.

Comments

[SimonC]

Hmm!
Just two quick comments.

1) Metis represent a large percentage of persons self-declaring an aboriginal identity in Canada. When someone self-declares an aboriginal identity, that person is not asked if she is a visible minority.

Would metis answer "american indian" or "white" when asked about their *race*?

2) Census (2006) does use the word white :

[SimonC]

(because you are assuming 100%)
cheers!

[Frances Woolley]

Simon, thanks for those comments. A couple of responses.

- I said "we don't like to use the word white." We have to, because it's in the employment equity act "people.... who are non-white in colour...", but contrast the wording of the Canadian and US question - the word race isn't in the Canadian question at all. Statistics Canada doesn't use the word "white" any more than it has to - type "white" into the Statistics Canada search engine and you'll get a whole bunch of results for White Rock, BC (a beautiful place, but not what the typical searcher is looking for). The white+aboriginal numbers are reported together as the "not a visible minority" category, which many people mistakenly believe is equivalent to white.

- on Aboriginal people - someone who self-identified as Metis would probably tick both the white and the Aboriginal box. if you look at the US numbers, almost half of Aboriginal people report more than one race. This means the Canadian number I've given for "white" is more comparable to the "single race" figure for the US than the "inc. multiple races", whereas the Aboriginal (American Indian) number is more comparable to the "inc. multiple races" figure. Totally irrelevant aside: did you know that people of mixed African-American ancestry are sometimes called "Red Blacks", which is the name of the new Ottawa football team?

[Frances Woolley]

Simon -

On the Aboriginal question, there is bias coming from two directions. Yes, there might be some Metis who would answer "white" to question #19 if asked.

However there are probably a whole load of people with some Aboriginal heritage, who don't tick the Aboriginal box. I'm thinking, for example, of a certain economist who will happily tell anyone that he's part Cherokee. He wouldn't identify himself as Aboriginal, but he's proud that he has some American Indian heritage. Since anyone who has long deep roots in this country can probably claim some Aboriginal ancestry, I would think that there's considerable scope for underreporting of Aboriginal ancestry as a result of the exclusion of "Aboriginal" from the list of vis min categories.

[Wonks Anonymous]

In the U.S context, most whites who claim some Amerindian ancestry turn out not to when given genetic tests.

[Frances Woolley]

Wonks - that might be true, but remember the relationship between the Europeans and the Aboriginal population was very different in Canada and the US. Basically native Americans were sitting on quality agricultural land, and this led to a certain amount of conflict between the Europeans and native americans.

A good number of Aboriginal Canadians, on the other hand, were sitting on land that was not useful for agriculture, but was valuable for the fur trade. The Europeans didn't have the ability or the population to be able to trap, so had long-standing trading relationships with the Aboriginal people.

And when people meet and interact regularly, stuff happens.

[Simonc]

"I would think that there's considerable scope for underreporting of Aboriginal ancestry as a result of the exclusion of "Aboriginal" from the list of vis min categories."

I dont think the the way the form is built you will see underrporting of *ancestry*.

There are three aboriginal type of questions:
Ancestry #17 (Do you have Indian blood?)
Identity #18 (Do you consider yourself an Indian?)
Band membership #20 and Registered Indian status #21 (Does the Government consider you as an Indian?)

The ancestry question (#17) is asked before the identity question (#18) and the visible minority question (#19) is only skipped if the persons reports an aboriginal identity.

Your economist would say "I'm part Cherokee" at #17
Then would say "I dont consider myself an aboriginal" #18
Then would say "I'm white" at #19.

So I dont think the ancestry part would go underreported. And I don't think that economist with a fraction of Indian blood would have reported "American Indian" as his race in the US either.

Re: the US Indians ticking more than 1 box. Did they have any other question where they were able to claim their "aboriginalness" ? If not, it may be a result of people having nowhere else to do it (and then your economist friend might have checked American indian as one of his race).

[Simonc]

Could it just be that Aboriginals don't want to be called "Visible minorities" and that the question #19 skip is just a way of protecting these "sensibilities" ?

I mean, look at the way this part of the form is designed: "white", "black" , etc.. are mentionned.

[Frances Woolley]

Simon - all of this is in aid of implementing the Employment Equity Act. The EEA aims to address barriers to employment for members of four designated groups: Aboriginal Canadians, persons with disabilities, women, and members of visible minorities. Aboriginals aren't in the visible minority list because they're Aboriginal.

That 4.3% of the population number comes from the Aboriginal Identity question (#s 20 and 21) not the ancestry question (#17).

[Simonc]

EEA exmplanation makes sense.

The identity numbers come from #18, not 17/20/21.

If we look at the aboriginal thematic tables for census 2006 (http://bit.ly/oFIRrH)

You will find
1 678 000 persons with Aboriginal Ancestry (http://bit.ly/18RJYeU) (q17)
1 172 000 persons with Aboriginal Identity (http://bit.ly/11v1vrf) (q18)
623 780 persons with Registered Indian Status (http://bit.ly/18RJYeU) (q21)

--

[Frances Woolley]

Simon - I just checked in the 2006 PUMF. The percentage of people responding that they had some Aboriginal ancestry in the ethnicity question is quite a bit higher than the percentage of people identifying themselves as Aboriginal on the Aboriginal identity question (3.8% of the pop identify on the Aboriginal identity question, as opposed to more than 5% on the ancestry question - but a lot of that 5% have a mix of Aboriginal, British, French etc heritage.

Basically the bottom line is that we don't know how Canadians would respond to the US census questions - but I still think the broad trends revealed by this exercise are quite interesting.

[marcel]

We have a much larger Asian population and a much smaller Black population.

Canada has a larger Asian population or a higher proportion of its population is Asian? Given that the ratio of the 2 countries' population is about 10:1, I have difficulty believing that the former is true.

[asp]

You can't be serious that latinos are considered more then 50% white and less then 2% aboriginal! That's almost exactly backwards.

[Frances Woolley]

asp - the US Census asks two questions. One is basically Hispanic yes/no. The other is race: white/black/American Indian/Asian/ etc / some other race.

Of all the people who identify themselves as Hispanic on the Hispanic question, 53% then go on to identify themselves as white on the race question.

The way census defines race, it's a matter of identity, not some objective measurable fact.