Racial Classification in America

Benjamin Klutsey and David Bernstein discuss how the U.S. government categorizes race and ethnicity, and why that system is flawed

Confusing categories. The U.S. government’s racial classifications were designed for bureaucratic convenience. Image Credit: blackwaterimages/Getty Images

BENJAMIN KLUTSEY: Today’s conversation is with David Bernstein. David Bernstein is a university professor and the executive director of the Liberty & Law Center at the Antonin Scalia Law School at George Mason University in Arlington, Virginia, where he has been teaching since 1995. He has been a visiting professor at Brooklyn Law School, Georgetown University Law Center, the University of Michigan School of Law and William & Mary Law School. He’s the author of numerous articles and several books, including his latest, “Classified: The Untold Story of Racial Classification in America,” which is the subject of our conversation today. Thanks for joining us, David.

DAVID BERNSTEIN: Thanks for having me.

Researching Racial Classifications

KLUTSEY: David, I’m very curious about what inspired you to write about racial classification.

BERNSTEIN: I have done a lot of work on the history of race in the United States, on constitutional law, and some very interesting work on the history of racial classification, and what you might call the battle days before the civil rights movement, and how people determined whether someone was Black or Asian for immigration or work purposes and so forth. That’s why I've considered antediluvian stuff; no one’s trying to measure whether someone is a mulatto these days like they used to.

It occurred to me that we have all these boxes that we need to check whenever we fill out a mortgage form or the census, or apply to college or apply for employment and so forth. I had in the back of my mind, “I wonder if anyone ever really enforces these and what the policies are behind them and how they’re defined.”

There are a couple of incidents that led me to think about it further. One was I had a nanny from Peru, and I was helping her get a green card. I’m sitting with her in the Immigration Services office, and she has to fill out the form, and it says, “What’s your ethnicity?” She had no trouble putting Hispanic for that, but then the second question is, “What’s your race?” She was puzzled, and she said, “What do I put down?” In Spanish, I said, “Are you white?” She said, “No, I’m not white.” I said, “Well, you’re not Black.” She was, “No, I’m not Black. I’m Mestizo,” which is, in Latin America, what they call someone who’s of mixed Spanish and Indian heritage. And of course, we don’t have any classification like this. I was like, “Huh.” Interesting that the classifications we have are arbitrary. She doesn’t identify herself this way, but there’s no way for her to identify herself by what she would naturally do.

Another thing is actually the president of our own university, President Cabrera, I noticed he was an immigrant from Spain. And he was on the cover of something like Diversity in Higher Education Magazine. I thought that was odd, because okay, he’s from Spain, but what makes him any more diverse than any other Mediterranean immigrant from Portugal or Italy or Greece or wherever, all of whom I knew were considered white? Are people from Spain actually officially considered Hispanic?

I started looking into this, and the standard take in the legal academic literature is that race is basically self-identified, that there’s only one case ever which any court in the modern times has ever determined whether someone really meets a criterion or not. I said, “First of all, are there official classifications?” It turns out, it’s almost never mentioned in the law review literature, oddly enough. But there is something called Statistical Directive Number 15 that was enacted by the Office of Management and Budget in 1997, put in the Federal Register in 1978.

It actually lists the standard familiar classifications that we all know: African American, Black, Hispanic/Latino, Native American, white and Asian American. That’s continued on with very little change since the 1970s. States have their own classification systems that often match the federal one, but often have slightly different definitions. So there are actual official definitions. There aren’t that many. And I would say that racial classification, ethnic classification is mostly a matter of self-identification.

But not really surprisingly, if you think about it, that one case that everyone cited as the only case in which racial identity in modern times was, in fact, disputed—there are actually several dozen. They involve Asian heritage, who is Asian, who is African American, who is Native American, but especially who is Hispanic. The Hispanic classification, maybe because it’s the least logical in some way, actually winds up with a fair amount of litigation, in the context especially of affirmative action. Well, you put down, for example, that your name is Joseph Smith, but you claim to be Hispanic. You don’t really look Hispanic, you don’t have a Hispanic-sounding last name, you don’t speak Spanish, by your own admission. What makes you Hispanic?

Then the question is, do you go with a cultural definition of, “If you meet the definition of the kind of person we think might be discriminated against because they’re Hispanic, then you meet the definition for Hispanic”? Or do we go with the literal federal definition, which just basically says you have to have a Spanish-speaking ancestor?

Creating Racial Categories

KLUTSEY: Wow. It’s a very interesting topic. In many ways, it’s bizarre because of the nature of the classifications. As you mentioned, there are these five standard, official classifications that were established in the late 1970s: American Indian or Alaskan native, Asian and Pacific Islander, Black, Hispanic and white. You know that there was no logic to these categories. But why did they pick these, and can you give us some background on how these came to be?

BERNSTEIN: Sure. The illogic in the sense is that basically, there’s no consistency among the classifications. African American or Black is defined as someone with ancestry in one of the Black racial groups of Africa. That’s a “racial classification.” Asian is that you are descended from one of the original peoples of Asia. It’s more of a geographic classification—but not all of Asia, only Asia from western Pakistan on. White is defined by having European, Middle Eastern or North African heritage. And Hispanic is a cultural classification. Native American is defined by this part of federal law as basically a cultural affiliation. They’re not consistent with one another.

How they came about, basically . . . well, let’s go back a little bit. There were always racial classifications in American law, especially in the Jim Crow South, of course. But the federal government classified people for census purposes as white, Japanese or Chinese (as there were more immigrants from those groups) or what would have been known as Negro or colored—what we now call Black or African American. But there were very few legal consequences because the vast majority of what was going on, racially speaking, was happening at the state level.

After World War II and the Holocaust, and the civil rights movement was really taking off, and African Americans were moving north and getting more political power, there was a real push to say, “No, we really shouldn’t be giving people legal privileges or not, based on their racial status.” I think the intellectual climate at the time was that part of the civil rights movement’s objective was, “We’re going to abolish racial classifications.” It was really well-intentioned and not malevolent at all and not meant to divide up people in any bad way by race.

The federal government started saying, “We have a problem with federal contractors. We’re giving government money to defense contractors and highway builders, and a lot of them are engaging in pretty gross discrimination based on race, and we want to ban federal contractors from engaging in discrimination.” Well, how do we do that? Well, we make them sign a document saying they’re not discriminating. Okay, that’s very nice. How do we enforce that? Well, the only way to enforce that is to know how many people of each group they’re hiring.

Obviously, statistical discrimination won’t always work as telling you exactly what’s going on. But if someone, say, is a government contractor in Atlanta building highways in the 1950s, and they have no African American highway workers, that would be a pretty good indication that they refuse to hire Black people. No one really thought about this too much. The original forms said, “How many African American”— or they would use the word Negro in those days—“Negro employees do you have? How many Spanish Americans?”—which was a code word for Mexicans. And sometimes they would say “Jews,” and then they would say “others.” It was really meant to prohibit discrimination extremely broadly.

Visible Differences

BERNSTEIN: One thing I really didn’t appreciate when I wrote “Classified,” but I’ve come to appreciate it more now that I think about it, is that the difficulty with the classifications—why did they choose the classifications they did? The difficulty with these classifications was that the policy of civil rights bureaucracies and civil rights activists at the time was that you should never ask someone about their ethnic or religious identity, that this is an invitation to discrimination. Again, we wanted to really abolish these classifications; we didn’t want to force people to self-identify and potentially subject themselves to discrimination.

The natural way of enforcing these laws, then, it has to be the only so-called visible minorities are going to be classified for civil rights purposes. Because if you can’t ask people their ethnicity—well, how do you know whether someone’s Catholic unless you ask them? Catholics were subjected to a lot of discrimination at the time, but there’s no way of knowing who’s Catholic unless you ask. And with Jews, to a large extent, it’s the same thing.

To some extent, there were two things going on. One is, there was an impetus. We know that the classic example, the most significant example of discrimination in the United States is against African Americans. Which groups are most analogous to African Americans? Because Asians were a racial minority and Mexican Americans, to some extent, with their indigenous heritage, often were a somewhat racial minority. Let’s choose those. Because we’re not asking people if they’re Hispanic or Mexican—we’re just looking at them—it’s only really the Mexicans who look somehow different who are going to be included. It was a combination of what is bureaucratically convenient.

There’s a book, “Seeing Like a State,” that says the government will count things that they can count. If you’re not asking people their identities, they’re going to wind up counting the people who they could see are visible minorities—plus the ideological impetus that the classic form of discrimination in the United States is on the racial basis. It’s people who look different who we’re most concerned about.

Basically, by the early 1960s, the classifications had ossified into the Black—which we, again, call Negro at the time—Black category; Japanese and Chinese and maybe Filipinos, which became the Asian classification; Spanish Americans as sometimes Spanish Americans or Puerto Ricans or something, Mexican Americans or Puerto Ricans, which became the Hispanic category; and the Native American category. Without anyone really thinking about it, we had really re-created the classic racially discriminatory classifications, which makes some sense, as the classifications we also use for civil rights enforcement purposes.

By the time we get to the 1970s, what’s happening is the Hispanic category is actually the most controversial. Because we have dozens of different federal agencies that are trying to, on the one hand, enforce civil rights legislation and, on the other hand, keep track of different minority groups to see what progress they’re making, to see if there’s anything the federal government could do to step in.

Hispanics, Mexican Americans had traditionally been classified by the census as white. There was no real history of having a Hispanic classification. Some agencies were using Mexican American or Puerto Rican. Some agencies were using Spanish American, some of which was quite vague. Some were using Spanish-speaking household, some were using Spanish-language minorities. Some added Cuban American in.

And this led to the educational subset of what was then the Department of Health, Education and Welfare writing to Secretary of Health, Education and Welfare Caspar Weinberger and saying, “We have all different agencies presenting us with data and it’s all—we can’t compare. It’s apples and oranges because some are included, and some also have an ‘other’ category to include like Cajuns and Portuguese immigrants, local minorities, basically. We need to regularize this.” And Caspar Weinberger said, “We really do. Let’s do that.” And they created a committee to regularize it.

Surprisingly, when you go back and look at newspapers at the time, there was no real controversy, no real discussion. It was done not secretly, but without anyone really paying much attention. Because really the only purpose was, we just wanted regularized statistics within the bureaucracy so we know what we’re measuring. Like “Seeing Like a State,” we have to know what we’re measuring.

And they basically, in a very haphazard way, just gathered the Hispanic subcommittee. They just had one Cuban American, one Puerto Rican American and one Mexican American volunteer, who had no particular expertise in the matter, and said, “Okay, figure out what to call the classification.” They could have called it anything. They decided on Hispanic. Once you decide on Hispanic, for example, it includes people from Spain but not Brazilians, just without anyone really discussing it. That’s what we did.

When these criteria were announced in the Federal Register, they said, “By the way, these are only meant for statistical purposes. They’re not meant to be the basis for any eligibility for any government program. They’re not meant to be scientific or anthropological. We weren’t looking at that.” Nevertheless, as these things had to happen immediately, you are a government—or you’re, let’s say, Harvard, and you want to engage in affirmative action. The Department of Education is now requiring you to gather these statistics. The default is we’ll use those for other purposes as well. They quickly became ingrained in American society, and not just ingrained in the sense that we check these boxes, but that at least some people came to identify with these classifications who wouldn’t have before.

No one would’ve called themselves Hispanic in 1970. We talk about Hispanic literature, Cervantes, but no one thought of people as being Hispanic. But now people at least who have a secondary identity often have Hispanic as an identity. The Asian American category is just made up. No one from China thinks to themselves, “I’m an Asian,” right? They’re Chinese.

I was just visiting a college with my daughter who’s about to attend college next year. Someone was proudly recounting how they belong to the Asian American Student Organization and how this has been a big part of their campus life. Really, the whole classification—there were some activists who try to promote it—but mostly comes directly from the federal government. It’s a very long answer.

Liberal or Anti-Liberal?

KLUTSEY: No, it’s very interesting. On page 244 of your book you write that “the internal American struggle between the desire to maintain official racial classifications to redress harm from racism and wanting to eliminate them as unconstitutional and anti-liberal continues.”

Now, as this series of conversations focuses mostly on liberalism and pluralism, basically fostering a society where we can live peacefully amid deep divides, I’m interested in hearing more about this internal struggle. Could you unpack the tension between maintaining a constitutional liberal republic and the desire to redress past harms through official racial classifications schemes?

BERNSTEIN: Ultimately, in the book, I argue that as a general matter, at least the default position should be that we should have a separation of race and state. The government should just not have official classifications because of all the obvious—at least to me—dangers of people organizing around race and the government classifying people by race that we’ve seen in many societies around the world over time.

However, the one place where I draw the line is, I don’t think that—as long as we’re going to have civil rights legislation and as long as the public thinks civil rights legislation is necessary, it really is impossible to enforce many civil rights laws without gathering some data based on race.

Now, we could argue about things like whether statistical discrimination should be the basis for employment litigation in 2022, and maybe we don’t need that, but there are other things like the Voting Rights Act. How are you going to prove that the voting rights of, say, Mexican Americans in the border town in Texas were suppressed if you don’t know how many Hispanics there are, how many of them are registered to vote and how many of them you’d expect to vote based on past patterns? There’s no way of doing that.

In that sense, some classification is inevitable. That, one has to concede even if one supports a separation of race and state. Where it gets more complicated is, what about trying to redress harms based on race more generally and not just through things like affirmative action, but simply studying how is Native American health going, or how much academic progress are African American boys making, or things like that that we see in the media all the time.

What I argue in the book is that I think there’s nothing wrong with trying to gather that data. But that the way the government has skewed the system by requiring everyone to use these classifications—which means that these classifications are by far the easiest dataset to get if you’re a researcher, but also if you get money from the government, you may be required to use these criteria—has actually impoverished our understanding of what’s really going on within minority groups and nonminority groups in the United States.

Limited Usefulness of Classifications

BERNSTEIN: For example, if you’re going to study how Hispanics are doing, the Hispanic population is obviously quite diverse. And just to give an easy example, Cuban Americans living in the Miami metro region are going to have different indicia of all sorts of things and different challenges and successes than, say, recent immigrants from Puerto Rico to central Florida who work in agriculture as manual laborers. Putting those two groups together and averaging them out doesn’t really tell you anything meaningful. If you’re actually doing this research, you want to be able to separate them.

In the book I give the example of not just the ridiculousness in a scientific sense or anthropological sense of combining all Native American populations as one group. But even within one state, I think it was Arizona that I have in the book, different tribal groups within Arizona have quite different indicia on various measures of public health. If you just average, say, Indians in Arizona are doing well or not doing well. Let’s say on average they’re doing okay, but it turns out that on one particular reservation they are having huge rates of infant mortality or rates of alcoholism or early death or whatever, you’re not going to find that out. I think that we need to have a more nuanced understanding. The government has forced all the statistical researchers into this box. There are people, of course, who go outside the box, but again, it’s difficult.

I just saw something in the news where they had this study, “Hispanics are doing the following.” Well, that’s very nice. “Hispanics are doing slightly worse than average on the following educational indicia.” But does that apply to Cuban Americans in Florida? What about Argentines? Or how are the people who fled Venezuela since the Chavistas took over doing? No way of knowing, but this is what you really want to know.

For that matter, one thing that bothers me in retrospect is that I’ve become so inured to the classifications that we all live with, that it didn’t occur to me until toward the end of the book that the white category doesn’t really make that much sense either. It’s not really a natural category. As someone I quote in the book says, “It’s a government-created pseudo-race, including everyone who has origins anywhere from Iceland to Turkey to Morocco.” What do these different groups have in common?

As most of us are aware, one of the groups with the worst socioeconomic indicators in the U.S. is Appalachian Americans, who have very high rates of substance abuse, high mortality rates, low average age of death and so forth. They have been really hit hard by the opioid epidemic. And as we probably know, historically speaking, one of the main reasons, or at least one of the main ways that Lyndon Johnson sold the war on poverty, was even in the early ’60s, it was recognized that rural Appalachian whites were really having a lot of economic and other troubles. We don’t really get statistics about Appalachian whites unless you’re some researcher who’s found the grant specifically to study that, because again, they’re all just folded into the general white classification.

We know that there are different white groups who have different levels of educational, socioeconomic and other kinds of success. But we don’t really know very well groups that back in the ’70s were in this “other” classification because they had been subject to historical discrimination locally, like Portuguese immigrants to New England, or Cajuns in Louisiana or even Basque ranchers in the mountain West. No one knows how these groups are doing or if they’ve completely assimilated or anything else because no one has that data. Because the government, again, has forced everyone into these boxes. I think some statistics are needed. Some much more nuanced statistics should be used in social science research.

And the one thing I utterly, utterly reject—and I think is crazy really—is that on the other side of the ideological spectrum, there are people who believe, number one, that race is an immutable factor. They don’t believe race is immutable—they believe it’s a social contract—but they believe it’s immutable in the United States. Therefore the government must classify everyone by whatever groups that they find most convenient, then divvy out everything for equity purposes. That’s not itself crazy. That’s wrong, in my opinion, not crazy.

The crazy aspect is that there’s a whole group of people who write about race and the sociology, anthropology and law and other fields who, if you tell them, “Isn’t there a danger that by encouraging racial consciousness, you’re encouraging division and potential conflict?” they will tell you, “Quite the opposite.” And they don’t differentiate among white people. “We need to make white people more racially conscious so that way they recognize their white privilege. And once they recognize their white privilege, they can then become allies in anti-racism.”

Now, admittedly, there are people out there who are what we classify as white who are anti-racism activists. But if you look at social psychology, anthropology, history, basically any field you want, and say, “How often is it that you get people to strongly identify as members of a racial group, and they use that identity not to promote their own group but to promote the interest of minority or other groups?” I would say that’s a vanishingly small percentage of human history.

It’s not just optimistic to think that. It’s, in fact, quite crazy to think that this is what white people will do if you make them more conscious of their whiteness. Indeed, to the extent there’s literature out there in sociology and political science, not surprisingly, people who identify more with white being their primary identity are not surprisingly more likely to be racist and support racist policies and politicians.

Using Classifications for Political Gain

KLUTSEY: Yes. Sometimes these classifications are used for political gain. You mentioned Nixon in the book as one example of exploiting these categorizations for political benefit. Could you explain that further and perhaps highlight the most egregious example of this as you were researching the book?

BERNSTEIN: Sure. The Nixon case is interesting. I mentioned the Hispanic classification already. One of the issues was, as I noted, originally the classification was Spanish Americans as a euphemism for Mexican Americans because they were the largest Spanish-speaking ancestry group. Once there was a larger Puerto Rican immigration after World War II, many of them had African origins and therefore the discrimination they faced was partly racial, so they were added in.

Then we had this group of Cubans who fled Castro. Now, Cubans were not historically considered in the United States to be nonwhite. Of course, you could always find examples of people saying that, but we have a very good sociological example that we know it wasn’t, which is that Desi Arnaz, who is obviously Cuban American, was the star of the most popular TV show of the 1950s with a white wife. And the studio actually was worried about that, but in fact no one actually raised an eyebrow.

It turns out the American public, at least if you didn’t have obvious nonwhite ancestry, we’re happy to consider Cubans to be white. Most Cuban Americans themselves identify as white. And then we have DNA studies, I could tell you as a matter of scientific fact, that show of the major Latino subgroups in the United States, Cuban Americans have the largest percentage of European DNA. All indicators were this is just another white immigrant group.

They were treated as such primarily by the government bureaucracy, but then Nixon said, “Wait a second. Now that we are starting affirmative action programs, I want two things to happen. I want there to be a cross-national classification. Because you know one problem we’re having? We’re having these radical Chicano activists or Mexican nationalists. I want to clamp down on that.”

We’ve sort of forgotten about this, but there was a wave of Puerto Rican terrorism in the United States in the ’60s and even the ’50s, Puerto Rican nationalists who wanted independence. “I want the Mexicans and Puerto Ricans who come to the United States”—it’s actually a brilliant political strategy—“to think of themselves as American.” It’s a way to come up with a new American category that’s pan-national. And that turned out to be Hispanic, but it could have been a lot of other things.

“Also, once we do that, I want to include Cubans. Why do I want to include them? Well, they vote Republican. They vote for us,” he said. “Why should we be rewarding Mexicans and Puerto Ricans with affirmative action when they vote for Democrats, but not also including a hardcore Republican constituency?” Once you include Cubans who, again, are mostly white-identified in the classification, that means it’s no longer a racial category at all. It could include anyone white as long as they happen to have Spanish-speaking ancestry. That’s fairly egregious there.

BERNSTEIN: The most egregious example though, although understandable, of just a classification being changed for purely political reasons, was that as we noted, one of the five major classifications that were invented was the Asian American and Pacific Islander classification. One group that was subsumed within that classification were Native Hawaiians, people who were descended from the population of Hawaii before Americans and other Westerners started settling there.

What they found—by the 1980s, schools on the West Coast decided, universities decided, they had “too many Asians,” and it started becoming more difficult for Asian Americans to be admitted to places like Berkeley or Stanford or whatnot. Native Hawaiians were in the same classification. Native Hawaiians have relatively low socioeconomic indicators. They’re not one of the quote, unquote, “model minority” Asian groups. So they said, “Wait a second. We are the indigenous people of Hawaii. We were oppressed by the white settlers. We lost our culture. We have relatively low standards, low achievement in education and economics. And we’re treated like a group to be discriminated against instead of getting affirmative action like other minority groups. We want there to be a separate Native Hawaiian classification.”

The government—they approached OMB [the Office of Management and Budget], and OMB said, “You’re too small. We can’t have this few hundred thousand people in a separate classification.” They’re like, “Okay, we’re happy to be part of the Native American and Alaska Native classification. We may not be exactly like the Native American tribes but we’re pretty close. Alaska Natives—that seems like a pretty good analogy.”

The American Indian groups went ballistic. They said, “There’s no way we’re sharing the resources that we get from the Bureau of Indian Affairs with a bunch of latecomer Hawaiians who have nothing to do with us.” Their opposition killed that. Eventually, the compromise that was reached was, we’ll have a new classification called Hawaiian Native and Pacific Islander. We still have AAPI groups out there that still use the old nomenclature, but officially, for the government’s purpose, Pacific Islanders and Native Hawaiians are in a separate classification.

Again, as with all these things, Pacific Islander is a term of art referring to people from Samoa and so forth but not to the Philippines, which is odd because the Philippines are literally Pacific islands. We know now—again, from DNA and so forth—that they are more closely genetically related to Pacific Islanders than to East Asians. Nevertheless, they’re still in the Asian classification because they weren’t, I guess, clever enough to join the Native American lobbying for a separate classification.

Effects on Medical Research

KLUTSEY: That’s really fascinating. You also have a whole chapter devoted to the consequences the system has had on medical and other important areas of research. Quoting from page 196 of your book, you say, “The standardized racial and ethnic categories developed in Statistical Directive Number 15 in 1977 came with an explicit warning: ‘These classifications should not be interpreted as being scientific or anthropological in nature,’” as you said before. “And indeed, the classifications have no valid scientific or anthropological basis, yet the FDA and the NIH require medical researchers to classify study participants by Directive 15 categories.” What effect do these have on the quality of research?

BERNSTEIN: This is maybe the most shocking thing in the entire book. I had no idea before I started writing. In fact, as I was researching this book, it would not have come up in my research because it’s relatively obscure—I’m not in the medical field—but for the fact that someone else who has written about these issues alerted me to an article he had written mentioning this stuff. I started looking into it.

What happened was that, back in the early 1990s, women’s groups started lobbying for Congress to require medical researchers to include women in their research studies. Medical researchers preferred men because you won’t have all the complications of pregnancy and menopause, all the other things that might be confounding factors. So they tend to use men. But women, with at least some potential justification, point out that women actually are physically different from men. They have different reactions to certain pharmaceuticals, and not including women means you’re neglecting women’s health. Fine.

At the last minute, some civil rights groups, like NAACP and so forth, jumped on the bandwagon and said, “We want you to include minorities as well.” The justification there is not nearly as persuasive. There is generally no reason to think that what we call racial classifications in the U.S. will have any impact on scientific studies. Now, people will say, “Well, wait, there are some studies showing some differences between African Americans and whites.” Those are not racial differences; those are genetic differences. Not all African Americans, obviously, have the same genetic background. Africa is an extremely diverse continent.

If you just limit your study to African Americans, you don’t know whether this is going to be something that was a genetic difference that was concentrated in one regional group that happens to be somewhat represented, and it’s very high in that group, so it comes off as prevalent in the group of African Americans as a whole. Or whether it’s something that’s more generally genetically distributed among people of African descent.

People from Ethiopia and Somalia were increasingly large immigrant groups in the U.S. They are quite ethnically diverse within those groups, but some Somalis and Ethiopians are, for example, more closely related to Arabs and Jews genetically than they are to sub-Saharan Africans. In fact, part of my point in the book is that we should be looking to genetic differences. It could be that someone left Africa 30,000 years ago, brought some gene with them and that gene wound up concentrated in Iceland. You might wind up having, just for that particular disease or whatever, some genetic commonalities.

In any event, what happened was, nothing much happened until the late 1990s, when Congress then explicitly ordered the FDA and NIH [National Institutes of Health] to require everyone under their jurisdiction—which is a huge percentage of medical, biomedical, pharmaceutical research—to classify their subjects by race. The claim was not so much there are necessarily these ethnic differences, but that if you don’t have enough people who are Black and Hispanic and so forth, that those groups will lack confidence in the viability of the studies for them.

To me, the answer to that is, well, it’s circular. I’m an Ashkenazic Jew, personally. Ashkenazic Jews have a lot of weird genetic anomalies because of the fact that we have a dramatic founder’s effect, where the whole population of about 15 million Ashkenazic Jews in the world is descended from the same 400 people who lived in 1300. I don’t really worry about whether a study on vaccines or pharmaceuticals is applicable because no one’s told me to worry about it. It’s generally not something to worry about. But because the government says, “Oh, if you don’t have enough,” then of course you’re going to say, “Well, if the government says you have to have minority groups, then you must have to have them.”

In any event, what Congress did not do was tell the FDA or NIH how they measure what groups they’re going to use. The FDA or NIH could have done something that may have been at least somewhat scientifically plausible, which is actually gathered geneticists and anthropologists and so forth and say, “Okay, how should we define the groups that we want to be represented?” You can imagine what a political nightmare that would have been for the agency to start having hearings on: “How do we define race for medical purposes?” They didn’t do that, and what they did instead was use the default, ridiculous classifications.

There’s two different points I make in the book. One is that race itself, even if you could somehow define it in some coherent way—the way we normally define race, by different continents people came from. Even if you want to go with that, that is very problematic because of a lot of genetic diversity within the different groups and mixing that. Basically, geographic proximity is what makes people similar, not physiognomy. That’s one problem.

Even if you thought that that problem could be massaged somehow, the groups we have in the United States don’t match that either. African Americans, again, not only are they geographically diverse, they offer substantial admixtures. Not everybody, but on average, of American Indian heritage, European heritage—you could have someone who self-identifies as African American who could be 80% European. By origin, he checks off that box. Well, what good is that? It’s not ideal.

I’m going to go into levels of ridiculousness. The Asian classification is even more problematic and ridiculous because it includes everyone from people from Pakistan to people from the Philippines. People from Pakistan are Caucasians. Even if you use racial terms, they’re geographically quite far away from, say, Vietnam. You can have a study that says, “Well, we studied this on Asians,” but all of the people could have been, say, from Bangladesh. Does that tell you anything about Chinese people? Absolutely not. There’s no reason to think it would. When I see studies now that mention Asians, if it’s relevant to my research, I write to the office to say, “Well, which group of Asians did you use?” Because who knows? Filipinos are more related to Austronesians. This classification is ridiculous scientifically.

Then we get to Hispanic, which isn’t even a racial classification. People who are Hispanic could be any combination of white, European, African descent, indigenous descent, even Asian descent. I learned this living in Peru. The largest supermarket chain, for example, is Wong, founded by Chinese immigrants. Something like 6% of the population of Peru is of Asian origin. It’s like when you say American, it doesn’t tell you anything genetically whatsoever.

It’s just crazy to use this as a criterion for medical research. Of course, the white classification is also very problematic because it has groups like Icelanders, Ashkenazic Jews and Hungarians, who are somewhat distinct genetically. There’s like eight different subgroups. It’s all crazy. This would be funny but for the fact that, first of all, you have to spend a lot of resources as medical researchers trying to gather enough of these different groups to satisfy the government. That’s the least of the problems. Then it also discourages you because there’s limited resources for actually looking at real genetics, which could actually be helpful.

Also, the real-world consequences. Just to give you an example of a real-world consequence, when Moderna was testing its COVID vaccine, it announced that “we are committed to diversity”—as opposed to science, I guess—“and therefore, we are delaying our trials until we get more members of minority groups.” I thought that this was just, at the time, them just being afraid that they didn’t have enough members of the different groups, and they want to satisfy the government.

It turns out, though, that the head of NIH months later bragged about the fact that he made Moderna delay its trials until they have enough members of these different groups. I was like, “Really?” Because I don’t know how many people died, but the quicker we got people vaccinated, the fewer people died. People were dying of COVID because the Moderna vaccine was delayed because you had to satisfy a government official—who, oddly enough, has written in his academic work against using race in medicine—to satisfy a government official who insists that you have enough members of scientifically irrelevant classifications to satisfy the bureaucracy, which is just really nuts.

How To Encourage Tolerance

KLUTSEY: That is bizarre and fascinating at the same time. You had talked earlier about the unintended consequences that might result from these types of classifications, including, for example, the attempts to increase white consciousness for the sake of getting people to acknowledge white privilege, that backfiring. Instead of more tolerance, increased white racial consciousness encourages ethnonationalism, nativism and other manifestations of intolerance. You talked about why that might be the case.

But what might we do instead, if we wanted a more tolerant society? How do we talk about this topic and the unintended consequences of the racialized approach with academics and analysts who use the race-based system for their research and activism?

BERNSTEIN: I think one thing that I try to emphasize—and I mentioned earlier that some activists seem to think that race is just this immutable thing: We’re never going to get beyond race, or if we are, it’s only because people become woke after they get their right racial consciousness. I think if you look historically, there were divisions in American society that were more prominent, in terms of how people actually live their lives and how people would respond even to polling, than racial divisions or ethnic divisions are in the United States today.

Those of us who are historically minded will recall that in the 1920s, the most popular political organization other than the political parties in the United States was the Ku Klux Klan. It had a lot of power regionally in wide swaths of the United States. It wasn’t because of racism, although, of course, that was always part of the platform. But their main issue in the 1920s was being anti-Catholic. As late as 1960, when John F. Kennedy ran for president, there were doubts whether he could become president because people wouldn’t vote for a Catholic for president.

It turns out a lot of people would not. There were a lot of Stevenson voters that switched to Nixon because they wouldn’t vote for a Catholic. He was redeemed by the fact that there are a lot of Catholic voters who voted for Eisenhower, who in turn decided to vote for one of their own running for president. Now, this is only 62 years later, a blink of the eye for historical purposes. We have a Catholic president, a Catholic speaker of the house, a Jewish Senate majority leader. Six, I think, justices on the Supreme Court are Catholic. I’ve lost track now. It’s either six or seven.

Once in a while, when abortion comes up, you hear some expressions of anti-Catholic animus, but for the most part, no one cares. You couldn’t imagine this. If you have told someone in 1960 that 62 years from now, there’s going to be what I just described, and it’s not going to be controversial, people would have thought that’s just crazy. Most Americans back in the ’30s said they would not vote for a Jewish president. It’s probably like 5% now. Intermarriage among different religious groups was extremely low.

One statistic that never ceases to shock me—every time I mention it, I get shocked again—is that in 1958, Gallup for the first time asked, “Do you approve or disapprove interracial marriage?” Four percent—not 40%, not 14%—4% said that they approve of interracial marriage. By the late 1980s, it was only 50% or so, maybe not quite a majority yet. Now it’s over 90%. We’ve had this immense progress. Asian Americans and Hispanic Americans intermarry at high rates. African Americans, I think in the book I say the intermarriage rate was at 16%. I think in the latest data, 2020 census data, now it’s 22%. This was unthinkable even in my childhood.

You think about historical violence against Mormons, against Jehovah’s Witnesses. All the white ethnic groups that used to face discrimination, no one even thinks of these people as being anything other than white today. There is no particular reason other than ideology, I think, to believe that the racial divisions that we still have in the United States can’t also be overcome. And the future of American ethnic identity wouldn’t be that everyone could have their own individual identity, but that we basically identify as Americans at least.

Instead of seeing an American as you might have in the 1950s as inherently being a white person, it would just be someone of indistinct ethnic and racial origin. If you ever want to see what it used to be like, go to the Coca-Cola Museum in Atlanta. Look at Coca-Cola ads until the early 1960s, and you literally think you’re looking at Nazi propaganda, who they depict as your average American. That would, of course, not be the case today. I think that’s the trend, and most of that trend was done just by ignoring it. There’s no movement to get Catholics accepted. That just happened.

Redressing Historical Wrongs

BERNSTEIN: What I argue in the book is how do you get around, if you do recognize the legacy of slavery and Jim Crow, and you do recognize that, for example, Native Americans or reservations have had all these problems—what you’d expect when the government kicks them off their land and puts them in desolate places, and so forth and so on—what do we do? I point out that there are ways of looking at those issues that are not racial.

Someone who’s a descendant of slaves, that’s a political-legal dissent. The person who is the son of the Nigerian ambassador to the United States, he decides to stay here after high school, is not necessarily similarly situated as someone whose grandparents were sharecroppers in Mississippi and his great-grandparents were slaves. We could potentially make allowances for such individuals without it being on a racial basis. It’s not just that you are Black, but that you have this legacy of your family facing legal discrimination in United States.

With American Indians, I would probably limit the classification to people who actually live on reservations. I think there’s three or four million people in United States who identify as American Indians. Most of them are indistinguishable, socially and culturally, otherwise, from the general so-called white population. They’re far less than 100% Native American ancestry. They don’t speak Native American languages; they don’t live on reservations. They don’t really have many ties to the historic culture. They don’t look especially different from other European Americans.

There’s no particular reason to include them in whatever programs you have. There might be exceptions here and there, but for the most part, the separation of race and state could include a recognition of certain legal or geographic classifications. You could, for example, say, as a university, we want to help students who live on American Indian reservations, or who are descendants of slaves and are concentrated in an area where they have gone to school with and live with other people like that, who have had fewer resources because of legal constraints over time.

The Supreme Court has already, in 1974, in the context of American Indians, endorsed, for example, tribal membership as a nonracial classification. I think tribal membership is too broad, frankly, because some groups like the Cherokee have very, very liberal tribal descent rules. There is someone in the United States—I don’t know his name, but there’s somebody in United States who’s been identified who has something like 1/5,238th Cherokee. His great-great-great-great-great-grandfather was a member of the Cherokee tribe. He inherited it. He is officially a member of the Cherokee tribe, but in practice, he is indistinguishable from anyone who didn’t have Cherokee ancestry. There’s nothing specifically Cherokee about him.

A lot of federal programs would include him. I’m not sure why that makes any sense. That does seem to be, at that point, a racial criterion only because he has Cherokee blood, so to speak. I do think that a general public policy—and, of course, this will be resisted on both ideological and pragmatic grounds by organizations that owe their existence to there being divisions in society—could be to try to encourage a common American, nonracial, nonethnic identity.

Certainly, I’m not saying that people can’t enjoy the local Black student union or the Mexican American club or the South Asian students club. On the other hand, people have lots of other identities. They may be Catholic or evangelical or Mormon, or they may be part of an LGBT community. All these are identities, and the federal government really only officially recognizes racial identity or ethnic identity. We think that that is not a long-term good way of dividing society. Encouraging people to be checking these boxes constantly and thinking of themselves as member of these classifications is, in my opinion, not a good idea.

Following France’s Lead?

KLUTSEY: Great. Interesting. Now, should we be like France?

BERNSTEIN: France, that’s a good question. France, for two historical reasons, will not classify, will not gather any kind of ethnic or sociological data. It’s not just that the government doesn’t do it. There’s a taboo against doing it. The ideological reason is liberté, égalité, fraternité; we’re all just Frenchmen, and we shouldn’t differentiate ourselves. The historical reason is that they’re embarrassed that they had this data about Jews in the 1940s, and instead of burning the information when the Nazis came, they actually turned it over to them and helped them round up Jews—which actually goes to one of the dangers of classifying people to begin with.

It’s understandable, but I think people who are against the French system and against complete separation of race and state in that regard would point out correctly that fairly, at least, it also inhibits the government from recognizing certain things. I could give two examples. One of which is that the North African population in France—they usually say the Arab immigrant population, but actually not all North Africans are ethnically Arab, another classification issue—have not fully integrated into French society. A lot of them have not. They live in suburbs, and suburbs apparently in Paris are the bad neighborhoods, not like in the United States.

One reason that their lack of integration, and whatever could have been done to improve integration, was neglected was that no one was and no one still is keeping statistics about this. To the extent they’re facing a lot of discrimination based on ethnic criteria or they’re just having trouble culturally integrating, whatever it might be, there’s very little information about this, both because the government doesn’t keep the data and because, again, there’s this cultural taboo against studying it.

The second issue that’s arisen in France is that when there was a rise in antisemitism, primarily coming from these same North African immigrants—both the Jews and Arabs and Muslims in France are primarily North Africans. When there was Islamist or nationalistic-based antisemitism coming from the North African Muslim community, the French government was quite slow to respond. They say, well, rates of antisemitism overall are still quite low, and this and that, because we’re only going to look at the French population as a whole. We’re not going to say there’s a particular problem with this particular subgroup of the population. It took them quite a while, a lot longer ago they should have, to say there’s a real problem of extremism that we should be addressing.

That’s why I think, to the extent the United States remains committed to the social and economic progress of various groups, it’s probably not a good idea to completely ignore cultural, ethnic differences. Again, the categories we use are way too crude. For enforcement of civil rights law, the traditional classifications may be good enough.

The Asian one’s a little crazy because you could easily imagine someone discriminating against Chinese people but not South Asians. In fact, you could imagine a South Asian person discriminating against Chinese people but not South Asians if he’s a business owner, right? That classification probably should be improved. But the other classifications, because of historical powers of discrimination, are far from perfect, but they’re probably good enough purely for keeping track of discrimination.

For every other purpose, I think the answer is contrary to the French answer of don’t ever look at ethnicity ever in any way. Look at what you are trying to achieve. If you’re trying to achieve better medication, better pharmaceuticals, look to genetics, which may be to some extent correlated with race. You’re not going to use the racial categories we use for that purpose because they’re scientifically absurd.

If you’re trying to look at sociological, anthropological data to see how different groups are doing, actually divide people into legitimate subgroups that match their own ethnic, anthropological heritage and associations, rather than just lumping people together. How much in common do I have, as a Jewish guy from New York who graduated from Yale Law School, with an Appalachian high school dropout? His family’s been in the United States for 400 years, and we’re both white. If you met us on the street, you may not be able to easily say who’s who, but beyond that, if you’re measuring us sociologically, there’s very little in common.

For affirmative action purposes, again, if you’re looking to redress historical discrimination, you should be using much narrower, nonexplicitly racial classifications. If you’re looking for diversity purposes, I think that the whole diversity rationale is a crock. It was made up by Justice Powell; it’s not the real reason universities engage in it. But if we want to take the diversity rationale seriously for a moment, the way it actually works in practice is that, let’s say you’re University of Texas, and you have admitted 2,000 Mexican American undergrads under your Hispanic classification. You get your first Hmong from Minnesota, a Hmong child of refugees who wants to go to University of Texas. The University of Texas will say, “Well, we have you, our 2,501st Asian, or we could take in our 2,001st Hispanic, another Mexican American.”

You will have more diversity if you just look at these crude Hispanic versus Asian classifications. But in practice, you actually really have your 2,001st Mexican American, as opposed to your first Hmong. That’s not considered in practice at all because of, again, the crude classifications we have to use. If you truly do believe in diversity as a university, you should be asking people for much more nuanced versions of their identity, rather than the civil rights-based classifications that were not only not invented to give people a leg up in higher education, but were explicitly stated as they were not meant to determine eligibility for any program.

What Should We Do?

KLUTSEY: As we wrap up, I was thinking perhaps you could reflect or provide some advice or call to action to readers or academics and those who are generally interested in this topic. What do you want readers to take away from this?

BERNSTEIN: The first thing they should do, of course, is go directly to Amazon.com and type in “Classified: The Untold Story.”

KLUTSEY: That’s important.

BERNSTEIN: Then buy the book, so they have an even deeper understanding of what they’re talking about than reading our conversation.

KLUTSEY: Absolutely.

BERNSTEIN: The second thing is, at least for me, it’s both liberating and problematic, but I can’t read the news the same way anymore. I was listening to NPR during the election in 2020, and they said, “We’re going to go to talk to Asian American political activists about their views.” Then in practice, when they went, “We were here at the Korean American Association of Atlanta. We’re here at the Chinese Democrats of Hollywood office in California.” It’s like they’re calling it Asian American, but actually in fact, these are subgroups.

Now whenever I see these phrases, these are completely meaningless, bureaucratic nonsense they somehow adopted. I want to actually encourage people to think that way because I think one of the only ways we’re going to get around this is to recognize that these are just made-up classifications and often have very little meaning in the real world.

I think the other thing that’s going on is, the Supreme Court is likely poised to declare this in the current term, that the current system of affirmative action is unconstitutional. There’ll be questions of what are we going to do to replace this. I don’t think that the answer is going to be that we’re just going to look solely at people’s SAT scores and so forth. I don’t think that Harvard University is in the position to say, culturally speaking and given their status in society, that if we only have to admit like 2% African Americans, that’s okay. That’s not acceptable, whether for various historical and cultural reasons.

We want to maintain some level of diversity. We want to ensure that groups that were historically underrepresented are not completely closed out for their own success. I also think there’s a case to be made that when people don’t have a seat at the table they become hostile to existing institutions.

I don’t want to compare, obviously, the historical experience of African Americans with the present experience of conservatives. I will just make this very limited point, that one reason that Trumpism and hostile institutions have taken off among certain segments of the American right—as they look at universities, they look at the media, they look at cultural and arts institutions, they see themselves as not represented at all. And they feel like these are not institutions that we trust because we have no representation. I think there’s that case to be made.

But we have to figure out, if we’re going to get beyond the standard racial classifications and we want to have a multicultural, tolerant and cosmopolitan society, how do we allow people to both respect if they want their cultural, ethnic heritages classified as they like without having these official, crude and somewhat bizarre government classifications?

KLUTSEY: Well, David, this has been a learning experience. Thank you very much for joining us.

BERNSTEIN: Thank you for having me, again.