Employment Immigration Media Population Race and ethnicity

Why misleading, misinformed screeds calling for the elimination of race-based stats ought to be retired

Chart 1 Share of Canadian population reporting ‘Canadian’ ethnic origin vs share reporting* ‘visible minority’ status, by census year
Chart 1 Share of Canadian population reporting 'Canadian' ethnic origin vs share reporting 'visible minority' status, by census year

Source(s): Census of population public-use microdata files (PUMFs), Statistics Canada

The recent Canadian Economics Association (CEA) conference has been criticised for its lack of diversity, as have its precedents. It wasn’t clear at first why one particular economist was singled out for criticism. Apparently it had to do with an article  this economist, an older white female, wrote a couple years ago calling for the elimination of “visible minority” status from employment equity and, along with it, the race question altogether from the Canadian census. She argued race is an arbitrary, antiquated and irrelevant concept. That her screed was poorly written and argued is understandable, albeit unfortunate, given it’s not her area of expertise (‘feminomics’). That the same economist is supposedly set to chair the 2017 CEA conference is a legitimate concern.

To address the arguments the article makes in order:

That different national statistical agencies inquire about race differently and/or use different terminology in their respective questionnaires neither makes the question nor issue “arbitrary”. Not only do household survey questions differ between countries at any particular point in time, but they evolve differently over time as well, based on each country’s particular historical context. For example, the first decennial US census had only two categories for race, free whites and slaves; It wasn’t until the 1960 US census that Hispanics were distinguished as a racial minority.The Canadian census only started directly asking about race in 1996.

The different treatment of Arabic North Africans in the Canadian and US censuses – distinguished as a racial minority in Canada, but not the US – was cited as anecdotal evidence of the arbitrary nature of racial classification. In fact, the US Census Bureau had for some time been considering revising its questionnaire to include a separate race category for Americans of Arab descent.2,3 The additional category was tested in 2010, is being further tested this year and is tentatively scheduled for inclusion in the upcoming (2020) US census, pending  (2018) congressional approval.4

The suggestion that the United Nations has called for Canada to stop collecting data on race based on its concern over the term “visible minority” is disingenuous at best. Especially so when the UN reports routinely note a litany of socio-economic challenges facing racial minorities in Canada, findings substantiated by that same “visible minority” data.

The only issue the UN raises is that race by itself is sometimes insufficient for context, and that cross-tabulating it with ethnic origin in such cases may help.5 In no way can the agency’s reports reasonably be construed as suggesting Canada stop asking about race.

Citing poor results of employment equity programs in addressing racial diversity as a reason to stop asking about race is even more disingenuous. For example,  the federal public service’s spectacular failure to raise (and actually somehow manage to lower) racial minority representation cannot reasonably be ascribed to confusion over semantics.

Suggesting that rare exceptions of racialised ethnics groups that faired better and white ethnic groups that didn’t fair as well in one research paper somehow proves race isn’t really the issue is probably the most disingenuous. Especially so when the abstract of that same paper notes, “… for most visible minority groups earnings gaps are identified even among third and-higher generation Canadians.”6

That paper found Japanese and Korean Canadians third generation plus did relatively better than other racialised ethnic subgroups – just two of nineteen such subgroups distinguished in the paper. Likewise, Greek Canadians third generation plus didn’t do as well as other white ethnic subgroups – the only one of ten such subgroups distinguished. These rare exception may have been influenced by the sample restrictions, and may in part be attributed to their respective populations’ historical migration patterns, as discussed in the paper.7

Significantly, the paper provides both the results for the broader racial minority groupings (tables 2, 3) as well as for the distinguished racialised ethnic subgroups (table 4). Given the rare exceptions just noted, it should come as no surprise that the findings for the broader racial minority groupings mirror those of the more detailed subgroups: Blacks have the worst outcomes, followed by South Asians, Chinese and ‘Other’ visible minorities. In other words, race / “visible minority” status – not specific ethnic identity – is the main discriminating factor, as clearly indicated in the paper. That the referenced op-ed attempts to suggest otherwise is troubling.8

As such, the suggestion to do away with asking about race altogether, and instead exclusively inquire about ethnic origin, is ill-advised and ill-informed.

Oddly, perhaps the best argument against such a suggestion can be found in Statscan’s historical treatment of these questions. A question specifically inquiring about race, or “visible minority” status, was first included with the 1996 Canadian census. Prior to that, race was imprecisely imputed using ethnic origin, place of birth, mother tongue and religion questions. Quite controversially (worth a separate write-up), Statscan also explicitly added “Canadian” as an ethnic origin response for the first time on the same 1996 questionnaire. This made ethnic origin an even less precise proxy for racial identity, as clearly demonstrated in Chart 1 — now that is arbitrary. More to the point, this modification made asking about race explicitly from thereon even more critical to the Canadian census, not less.

For context, the US Census Bureau’s 2010 census test found that asking about both race and ethnic origin resulted in more precise identification. Given changes introduced to the Canadian census in 1996, it’s now even necessary for Statscan to directly query race for precision. This approach would also be consistent with the UN position on minority issues in Canada.

That said, why is an older white female economist advocating for the removal of race from employment equity and the Canadian census, and why is Statscan seemingly agreeing with this ill-advised/informed suggestion, contrary to its (prior) better judgement? That’s a complicated question worth trying to address separately. (Tentative title: Why a ‘diversity of views’ is no substitute for actual diversity.)

As it stands, those wishing to see greater diversity at CEA going forward will likely be disappointed, if the unfortunate views held by the supposed 2017 conference chair are indicative.


1. Chapter 1: Counting minorities: A brief history and a look at the future , Report on the American Workforce, US Bureau of Labor Statistics,2001.

2. We the People of Arabic Ancestry in the United States, US Census Bureau, March 2005.

3. The history of how those of North African / Middle Eastern descent are categorised in the US census is complex. A 2012 post on succinctly summarises the issue: “Ironically, those of Arab and Middle Eastern descent fought to be legally and officially designated as “White”, which was formalized by the Census in 1970. However, in the context of recent demographic, political, and cultural changes, Arab and Middle Eastern Americans now are more inclined to identify as a separate racial or ethnic group…”

4. Preparing for the 2020 Census: Measuring Race and Ethnicity in America, Director’s Blog, US Census Bureau, October 6, 2014.

5. Report of the independent expert on minority issues: Mission to Canada (October 13-23, 2009), United Nations Human Rights Council, March 8, 2010.

6. The Visible Minority Earnings Gap Across Generations of Canadians, Mikal Skuterud, Department of Economics, University of Waterloo, November 2009.

7. Ibid, tables 2 to 4, pp. 21-24. See table footnotes and related notes on methodology and historical context.

8. Suggesting that ‘Other’ visible minority – comprising Hispanic/Latino, West Asian/Middle Eastern/North African, Japanese and Korean – is too broad and could be further articulated in employment equity is a legitimate critique, but not one made in the referenced op-ed. It’s worth clarifying that the visible minority census question does distinguish between each of these ‘Other’ visible minority groups, so the data is available.

Leave a Reply

Your email address will not be published. Required fields are marked *