What colourblindness hath wrought

One of the more questionable trends in social justice thinking over the past decade has been the suggestion that there is something wrong with “colourblindness” as a strategy for achieving racial and ethnic integration. Indeed, for a while it became quite popular to suggest that the demand for strictly colourblind (or “identity-blind”) institutional procedures is itself a form of racism. It is, of course, possible to come up with some edge cases in which strict colourblindness winds up running counter to some widely shared intuitions about justice. This unfortunately makes it easy to forget that in the overwhelming majority of cases, the best way to achieve equality among persons is to refrain from drawing invidious comparisons based on unchosen characteristics, and the best way to do that is to deny decision-makers access to information that could be used as the basis of invidious comparison. Because of this, colourblindness remains the bread-and-butter solution to most problems of racial injustice.

I am reminded, in this context, of the complaint that one used to hear, all the time, about racism in America being manifest in the inability of a black man to hail a taxi in New York City. This was once a huge part of American popular culture (Michael Moore did a bit on it, Danny Glover launched a highly-publicized lawsuit), and was cited constantly in academic discourse. And yet about ten years ago the issue simply disappeared from public consciousness. Why? Because of ride-hailing apps (and the subsequent introduction of taxi-hailing apps). Specifically, because these apps do not provide the driver with information about the passenger’s race. In other words, Uber and Lyft (and Curb) impose colourblind decision-making on drivers, which basically solves the problem.

Remember though what I said about edge cases? It turns out that the ride-hailing system is not foolproof. There is still the potential for discrimination based on the customer’s name, which in some cases might reveal race. It turns out that Lyft provides the driver with the customer’s name at the time the ride is requested, whereas Uber conceals the name, only providing it after the driver has accepted the ride. One study found that this generates a higher level of driver cancellations for black passengers on Uber, but not Lyft. One could then, I suppose, make the case that the Lyft system is better because it is less colourblind (although its primary effect is not to eliminate discrimination, but rather to make discrimination invisible to the customer). But it is important to maintain a sense of perspective here. Overall, the introduction of a colourblind system has taken what was, by all accounts, a major source of racial grievance in America and pared it down to a source of minor annoyance. The fact that one might still want to tweak that remedy around the edges, possibly in a race-conscious manner, should not cause us to lose sight of how that primary outcome was achieved.

Proponents of colourblindness have often been able to point to Canadian institutions to defend their position, in contrast to the United States, which historically has been more inclined to adopt race-conscious measures. There are several reasons for this, the most important being that the Canadian federal government does not use the census to impose a racial classification system upon the population (historically it has used a somewhat complex and shifting ethnic-origins classification system), as a result of which most institutional decision-makers do not have access to official information on race. This has notable effects on universities, in particular, because most are incapable of adopting race-conscious admissions policies, even if they wanted to, because they don’t have any information about the race of applicants, or even the racial composition of their student body.

I should note that this has been changing, in recent years, in part due to the influence of American social justice thinking and policies. In Ontario, an optional “Applicant Diversity Census” was recently added to the province’s centralized university application system. It is unclear, however, how much of that information is actually transmitted to universities, because the University of Toronto, as part of the general enthusiasm for DEI initiatives, decided in 2020 to collect its own data on the race of students. This involved the creation of the (mandatory) UofT Student Equity Census. The motivation for the collection of this data was stated as follows:

This initiative aligns with broader institutional efforts that support anti-discrimination and enhanced equity. Such efforts include Answering the Call: Wecheehetowin, U of T’s response to the Truth and Reconciliation Commission, recent efforts towards Anti-Semitism and Anti-Islamophobia, and the Anti-Black Racism Task Force Report, which emphasizes the importance of collecting demographic data.

When the numbers came in, however, the results were somewhat awkward for proponents of race-conscious admissions policies. Here is what a strictly meritocratic admissions process has produced (based on the 2023 UofT Equity Survey, compared against 2021 census data for the Canadian population):

I’m not an expert in statistical analysis, but just eyeballing this graph, one could certainly get the impression that there aren’t any big problems of access to higher education for minorities. Again, it is worth emphasizing that these demographics are the outcome of a colourblind admissions process. (It it also perhaps worth mentioning that, not only do these data represent the demographic profile of the domestic undergraduate population at the best university in Canada, UofT is approximately the same size as the entire U.S. Ivy League, in terms of undergraduate enrollment. This suggests that the Canadian higher education system remains a very powerful engine for upward social mobility.)

Anyhow, the university invested an enormous amount of time, energy and money, to discover that the only racial group underrepresented in our student body are white Canadians, especially white men. Which is something that anyone who teaches at the university, or who has spent 10 minutes walking around campus, could have told you. (It’s also unclear what we should make of this, since this sort of demographic analysis doesn’t say much about what’s going on in Canadian society, it primarily reflects the operations of the Canadian immigration system. But that’s just another reason to refrain from importing ideas from U.S. race politics into Canada, since the situation in that country is quite different.)

Of course, the numbers that are most likely to surprise people are not the blue ones on the right-hand side, but rather the red ones on the left-hand side. That’s because most people in Canada have no idea what the demographics of Canadian society are. Because we consume so much American media, Canadians tend also to imagine that the numbers are about the same as in the United States. Specifically, Canadians vastly overestimate the percentage of the population that is black, and so most are surprised to discover that in the 2021 census it was only 4.3%, and that in the previous census it was 2.9%. This of course has important effects on people’s perceptions of how “representative” various institutions are. For example, it might lead a casual observer to think that, at 6.4% of the student population, blacks are underrepresented at UofT, when in fact they are overrepresented.

Something like this seems to have afflicted the UofT Task Force on Anti-Black Racism, which included the following recommendation:

A.5. Recommendation that every faculty, college and campus work with EDI to conduct a review of its admissions processes and consider a data-driven and holistic application process for Black students, who remain one of the least represented groups of students on campus. Each division/program to collect data about Black Students upon entry into program.

To the extent that the word “representation” has meaning, black students are not the least represented group of students on campus, and personally I don’t see how it benefits anyone to describe them as such. Now that we have the data, presumably we can put this harmful stereotype to rest. (I am assuming, by the way, that as the best university in the country, UofT should be drawing students from across the country – the way that Harvard does in the U.S. or Oxford does in the U.K. – and so the Canadian population should serve as the relevant baseline for determining representation.)

Interestingly enough, even if one sets aside the specific bias that afflicts Canadians, there is a more universal cognitive bias, which causes members of all groups in multicultural societies to overestimate the size of minority groups. For example:

It is a common finding that Americans of all descriptions vastly overestimate the demographic size of minority groups in the population. According to one 2022 study, “Black Americans estimate that, on average, Black people make up 52% of the U.S. adult population; non-Black Americans estimate the proportion is roughly 39%, closer to the real figure of 12%. First-generation immigrants... estimate that first-generation immigrants account for 40% of U.S. adults, while non-immigrants guess it is around 31%, closer to the actual figure of 14%.”

(For academic literature on this bias, see here and here.)

Strategically, this is something that the left should be more attentive to, as it sometimes leads minorities to overplay their hand. One must wonder how much of the political madness of the past decade – with members of minority groups vociferously attacking traditional liberal rights and institutions – was caused by a failure, on the part of these groups, to realize the extent to which these liberal rights were the only thing standing between them and overwhelmingly larger democratic majorities. (As David Brooks observed, of the sequala: “It turns out that identity politics is more effective when your group is in the majority.”) The attack on colourblindness, it seems to me, was part of this madness. It led people to forget that the beneficiaries of strict procedural neutrality are, first and foremost, those most likely to be otherwise subject to discrimination.

This is something that, in my experience, immigrants are quite aware of, which is one of the reasons that colourblindness tends to be quite popular with recently-established ethnic groups. This is based on the plausible expectation that deviation from procedural neutrality will tend to favor insiders and incumbents, which they are not. One should avoid drawing the conclusion from this, however, that all minorities favor colourblindness. There is a very important and glaring exception, which involves the case of nationalist movements, involving non-immigrant minority groups who are attempting to achieve some form of sovereignty or self-determination. In Canada, this means the Quebecois and various Indigenous groups.

The distinctive features of nationalists is that they are not seeking integration into the institutions of the majority society, but rather are seeking to opt out, in order to produce an alternate set of institutions in which they will form the majority. For example, the UofT Equity Census shows absolutely no interest in the major axis of diversity in Canadian society, which is the linguistic divide between French and English. Indeed, I have yet to meet one person at my university who cares even a tiny bit about the services and accommodations that are provided to francophone students, or who thinks that there should be some representation of francophones on faculty. (Our student orientation in the fall, for example, provides a special session for “Latinx” students and absolutely nothing for francophones.) Partly this is just Toronto being Toronto (i.e. insufferable). Partly it reflects the cognitive capture of the left in Canada by American social justice politics (exacerbated by the number of Americans on faculty). But partly it is because Canada has both French and English universities, and UofT is an English university. If French students are unsatisfied with the level of service provided at an English university, they can opt out and go to a French university in Canada – something that is not an option for native speakers of Spanish, Mandarin, etc.

It is noteworthy, in this context, that the attack on colourblindness that features centrally in DEI thinking originated in arguments made by African-Americans, focused on the specific circumstances and aspirations of blacks in America. An important feature of African-Americans, as a group, is that they are not immigrants, but rather are descendants of a group that was involuntarily incorporated into the American polity. In this respect, they have a lot more in common with French Canadians than with African-Canadians (or any of the other ethnic groups in Canada produced through immigration). This is reflected in the fact that African-Americans also have a strong nationalist sensibility, which is not focused on achieving integration into majority institutions. It is not surprising, therefore, that many African-Americans were critical of colourblindness and sympathetic to identity politics, because all nationalists are committed to identity politics. The mistake is to generalize this beyond the specific case of nationalist movements, and to imagine that, because black nationalists in the United States are opposed to colourblindness, that all visible minorities everywhere in the world should be opposed to it as well.

Here we arrive, finally, at the deep reason why it is inappropriate to apply American social justice politics in the Canadian context. Racial politics in America involve a complex blend of nationalist and integrationist demands, not all of which are simultaneously satisfiable. This is why racial conflict is so intractable in that country. The situation with Indigenous people in Canada is similar, which is why those problems are so intractable in this country. (I have omitted any discussion of Indigenous representation at UofT for that reason — it raises a whole other set of issues.) But the situation with immigration and ethnic diversity in Canada is not nearly as complicated, and as one can see by looking at the results of UofT admissions procedures, in some cases is not even a problem, much less an intractable one. The suggestion that we should abandon the policies that have been successful in Canada, in order to adopt a set of policies that have failed in the U.S., strikes me as positively perverse. We should instead learn to take a win when it is offered.