How The Census Bureau Socially Constructs The Next America(s)

What Is Your Race?: The Census And Our Flawed Efforts To Classify Americans I’m frequently accused of being overly interested in race and ethnicity, to which I reply: “Didn’t you fill in your Census questionnaire?”

Now, Kenneth Prewitt, whom Bill Clinton appointed head of the Census Bureau in 1998, has published an informative book, What Is Your Race?: The Census and Our Flawed Efforts to Classify Americansdocumenting the federal government’s dysfunctional combination of near-monomania over counting by race and lack of coherent thought about the long-run effects of how racial boundaries are drawn.

Despite his half-decade in charge of the Census, Prewitt shares with the average American a certain perplexity over his old department’s fixation upon race and ethnicity:

What perhaps puzzles the reader is why [Census] race statistics are so terribly important that they are announced simultaneously with the population figures mandated for reapportionment. You may also be puzzled that the census form dedicates so much of its space to the race and Hispanic question but has no space for education, health, employment, or marital status questions.

(Not to mention the absence of a citizenship question, which would ask about a simple yes-or-no legal distinction far less murky than race or ethnicity.)

Census And Race

An old-fashioned nice white Protestant liberal, Prewitt, who is now Carnegie professor of public affairs at Columbia University, expresses befuddlement at how a job he apparently assumed would be suitable for a technocratic good government Progressive like himself wound up plunging him into the maelstrom of modern racial politics. Thus his proposals for technical improvements in the Census quite unexpectedly (to him) degenerated into a donnybrook over race, complete with angry charges of, guess what, “racism.”

Prewitt points out that, from the disinterested perspective of promoting the commonweal, the federal government’s racial preoccupation synchronizes poorly with the lack of informed public discussion over the purposes of all this categorizing of people. Instead, the crucial process of drawing official racial and ethnic boundaries tends to be either hijacked by interest groups or is the remnant of bureaucratic inertia and lack of foresight.

Prewitt is particularly concerned about the on-going racial mobilization of immigrants that the federal statistical machinery fosters.

As a good Democrat, Prewitt understands that counting by race is necessary to provide racial preferences to the descendants of American slaves and to American Indians, whom he calls (reasonably enough) the “uniquely disadvantaged” victims of America’s past.

But, somewhat heretically, he asks: Do we really want immigrant groups to be viewed through the lens of race?

…if immigrants are racialized–the future sadly repeats America’s past.

Judging from the footnotes, Prewitt was influenced by the late Hugh Graham Davis’s 2002 book Collision Course: The Strange Convergence of Affirmative Action and Immigration Policy in America (which I wrote about here).

Paralleling Milton Friedman’s observation that a country can have a welfare state or open borders, but not both, Davis pointed out that racial preferences for blacks and American Indians are one thing–but for America both to allow massive immigration and offer racial preferences to most of the world’s peoples is such a paradoxical policy that few people can contemplate it.

Prewitt argues that if race and ethnicity are social constructs, as all good liberals like he himself assume, then our democracy ought to be prudent about how we socially construct them.

Prewitt takes the scientifically discredited “race does not exist” conventional wisdom seriously, but you don’t have to accept that to see the good sense in his advice. He documents a long history of bureaucrats and politicians maladroitly seizing upon existing categories for ill-suited new purposes, such as Affirmative Action; or making up categories for short-term advantages, only to unleash unexpected consequences.

We always wind up with what Prewitt calls “statistical races”–Census agglomerations that take on lives of their own because the government, the Main Stream Media, and Big Business bribe and browbeat individuals into pledging conformity to their categories.

To cite an example that Prewitt only dances around: the government’s decision, which he traces back to the Office of Management and Budget’s 1977 Directive 15, to move South Asians from the Caucasian to the Asian / Pacific Islander race, was not a good example of “carving nature at its joints,” as Plato would say. Genetics and physical anthropology suggest that Indians and Pakistanis have more in common with Middle Easterners than with the Chinese.

But this regulatory ploy freed Subcontinental immigrants up to get in on the minority business development action. So they didn’t complain.

Prewitt wisely recommends that instead we should:

“…start with agreement on public purposes and then design suitable statistics to meet policy challenges. Without clarity on why the nation should measure race, clarity on what to measure is impossible.”

Despite Prewitt’s Democrat loyalties, this is especially good advice for Republicans.

In recent years, the GOP hasn’t shown much interest in thinking through the implications of Census redefinitions (or in anything else that makes its operatives’ heads ache). Thus the Census Bureau is scheduled to deliver a report to the Office of Management and Budget by 2015 on how to change the race and ethnicity questions on the 2020 Census. Yet I’ve never heard of any Republican strategists paying attention to this dull but potentially devastating argument.

Perhaps the last time prominent Republicans got heavily involved was back in the 1990s, when Newt Gingrich briefly pushed for official recognition of a “multiracial” category.

Some individuals with a mother and father of different races had protested that the government shouldn’t order them to, in effect, choose which parent they love most when filling out their Census form. (Tiger Woods, for instance, refused to pick among his ancestors when Nike pressured him to identify as black. In contrast, Barack Obama chose to identify solely as “black” on the 2010 Census–whether out of loyalty to the one-drop rule or because he always resented his white mother and grandmother.)

Others lobbyists wanted government endorsement of a new pan-“Multiracial” identity lumping together, say, Eurasians with black-Cherokee hybrids.

But this proposal for one all-purpose Multiracial category failed to “carve nature at its joints” (as Plato would say). And eventually the Clinton Administration announced that, on the 2000 Census, respondents could check as many different race boxes as they felt like.

The NAACP objected to that because, under most principled interpretations, divvying individuals up among their ancestors would, ultimately, reduce the size of quotas for blacks, who currently can claim anyone with any black blood. After mulling this quandary over for years, the Clinton Administration brilliantly ju-jitsued Newt and placated the NAACP by declaring that, for the purpose of increasing racial preferences, mixed race people would be counted 100% as belonging to whichever of their nonwhite groups are most legally privileged.

I’m tempted to say that the lesson the Republicans learned from Newt’s gambit getting trumped is that they can’t win at playing these games, so why bother?

But, in truth, I haven’t seen much evidence that they learned anything at all.

For example, what if a push develops to allow not just “multiracial” but “multiethnic” answers on the 2020 Census? After all, it could be argued, why is the government forcing individuals with one Hispanic and one non-Hispanic parent to choose between them?

What would be the partisan impact, in the short run and the long, of accepting a “multiethnic” Census category?

It’s too much anymore to expect Republican operatives to think about whether that change would be good for Republican voters. But is nobody on Team GOP capable of at least modeling whether that would be good for Republican politicians?

Do the Republicans even have any quants anymore who will take on questions of race? The Democrats have been hiring scores of Nate Silver-like political moneyballers, while the Republicans have been firing Jason Richwine.

American politics, as Dr. Prewitt has learned to his distress, is fundamentally entwined with race and ethnicity. Thus, the firing of Richwine for thinking statistically about race was part of the Respectable Right’s ongoing unilateral intellectual disarmament. How does this process end well for the GOP?

Prewitt, who treads delicately and follows the contemporary requirement that you notice nothing remotely critical about anyone other than non-Hispanic whites, doesn’t offer many current examples of how obscure bureaucratic decisions of the past shaped today’s reigning mental models of the world.

But readers no doubt can come up with their own.

For instance, Senator Marco Rubio (Cuban) and Representative Luis Gutierrez (Puerto Rican) are treated by the MSM as natural leaders of the push for Amnesty for illegal Mexicans. In turn, Rubio, at least according to the magazine covers so prevalent earlier in 2013, is the natural GOP presidential nominee in 2016.

Underlying this logic–if that’s the word for it–is the MSM’s presumption of Cuban-Puerto Rican-Mexican solidarity. But why do pundits believe something for which little evidence has been put forward?

Largely, it’s because the government’s Hispanic/Latino category lumps all three groups together for convenience in reporting statistics.

Whether or not that much Latino cohesion exists in the real world, you can see it right there on the page in government statistics–and in all the journalism and marketing research that builds off federal data.

And why are all “Hispanics” grouped together? Originally, according to Prewitt, the Johnson Administration excluded Cubans because they tended to be prosperous Republicans who weren’t going to need special benefits. Richard Nixon, whose best buddy was Cuban banker Bebe Rebozo, tossed them in. Instead of divide and conquer, Nixon launched a new Republican strategy: unite and submit.

In fact, of course, below the Rubio/ Gutierrez level of the nationally ambitious, both Cubans and Puerto Ricans care little about Mexicans’ immigration troubles–because Cubans and Puerto Ricans enjoy special, extremely favorable immigration deals already.

However, both Cubans and Puerto Ricans tend to produce more charismatic personalities than the kind of Mexicans who don’t have any better options in life than to immigrate to the U.S. So we’ve seen the promotion of what would be a peculiar Cuban putsch within the GOP: a member of this tiny and declining fraction of Republicans might garner the 2016 nomination aided by Census-certified Hispanicity.

Both Rubio (pro-“path to citizenship”) and Ted Cruz (anti-Amnesty) are widely seen among Washington sophisticates as potential Republican presidential nominees–because they can, presumably, get the Mexican vote.

Whether Mexicans even like Cubans remains uninvestigated, except on (Here’s a recent Onion piece gingerly joking about this.)

But all these people whose names end in a vowel or a Z seem alike to Washington Insiders. After all, the Census groups them together. So why should Washington Pooh Bahs bother to check for themselves?

Another irony only vaguely touched upon by Prewitt: the federal government bends over backwards to accommodate the anti-black and anti-Indian racial prejudices of Latin Americans. It created a nonracial category called “ethnicity” in which the only choices are “Hispanic, Latino or Spanish origin” or “No, not of Hispanic, Latino or Spanish origin.” This allows immigrants from Latin America (or even Spain!) to declare themselves to be both blue-blooded whites and deserving of the unequal protection of the laws to make up for… the crimes los Anglos committed against their pure Castilian ancestors during the War of Jenkins’ Ear.

Or something. I don’t know. In fact, nobody seems to know why various white hidalgos get to make a nice living for themselves in America as Leaders of the Brown Oppressed Masses. But then nobody asks, either. That’s just the way it’s been since the 1970s.

Prewitt argues, rather as I proposed in 2009, that the Census just ask everybody’s ethnicity. He suggests that the Census should ask about ancestral nationality–which countries you, your father, and your mother were born in—and phase out race.

This would help distinguish between, say, Michelle Obama (who is descended from American slaves through all four grandparents) and Barack Obama (who isn’t descended from slaves–but probably is descended from slave traders).

Be warned, though, that that’s my example. Prewitt’s are duller.

Another of Prewitt’s major themes: the obvious contradiction between the conventional wisdom that “Race does not exist” and the Census’s use of five racial categories that coincide closely with the five proposed by the 18th Century German physical anthropologist Johann Friedrich Blumenbach. He took the four races proposed by master taxonomer Linnaeus and split Malay/Pacific Islanders from the Asiatics to create five. According to Wikipedia, Blumenbach proposed:

  • the Caucasian or white race
  • the Mongolian or yellow race
  • the Malayan or brown race
  • the Ethiopian, or black race
  • the American or red race.

Contemporary American Race Denier Stephen Jay Gould tut-tutted over Blumenbach’s influence:

An extraordinary thing happened two hundred years after Blumenbach announced that the world’s population should be divided into five race groups distinguished by skin color. The United States government agreed.[The Geometer of Race, By Stephen Jay Gould, Discover, November 1, 1994]

Prewitt is a true believer in the Gouldian conventional wisdom and is wholly unaware of Gould’s later discrediting because of his projection of his own biases onto better scientists who came before him. (This June 14, 2011 New York Times’ editorial acutely delineates Gould’s fundamental flaw.)  So Prewitt seems genuinely surprised the Census still uses Blumenbach’s much-denounced five races.

But why does it? Because they are good enough for government work.

Seriously. There are always multiple ways to classify people, with the famous splitter v. lumper dichotomy being the most obvious. Should you have just a few big boxes or many small ones? Good question. It depends on your purpose.

Moreover, there are other racial groups that are missing from Blumenbach’s Big Five. For example, Australian Aborigines were only discovered by Captain James Cook in 1770, shortly before Blumenbach wrote. Also, there are isolated groups, such as the pygmy negritos of the Andaman Islands, who don’t fit in well into Big Picture frameworks.

And, of course, there are lots of individuals and peoples who are mixes.

A simple all-purpose definition is that a racial group is a partly inbred extended family. But in thinking about your own relatives, you’ll notice that it’s not completely self-evident how to group them.

Overall, though, Blumenbach was a great scientist. Considering how hard it was to accumulate evidence about distant groups in the 18th Century, he did an impressive job of carving nature more or less reasonably close to its joints.

In contrast, Prewitt attributes the Clinton Administration’s 1997 splitting of Pacific Islanders from Asians, which completed the Blumenbachanization of Census racial categories, simply to crude political pressure from Senator Daniel Akaka (D-HI), who is half Native Hawaiian.

Akaka essentially argued that Pacific Islanders don’t act like Asians. Prewitt summarizes

“…Native Hawaiians and Pacific Islanders had lower incomes, fewer educational opportunities, and less access to health care compared to the larger Asian population. He urged that these disparities be separately tabulated and that government programs be targeted accordingly.”

In other words, overachieving Asians were cutting down on underachieving Pacific Islanders’ opportunity to get themselves some Affirmative Action.

Akaka’s second argument, according to Prewitt, was “expressive.” In other words, Pacific Islanders don’t want to be officially classified as some sort of portly afterthought to Asians. Native Hawaiians like being Native Hawaiian, like expressing their Native Hawaianess, and want their government to salute them for it.

As a white technocrat who wouldn’t dream of expressing pride in his own race, Prewitt had trouble reconciling how much of his Census job turned out to be facilitating racial/ethnic pride displays, which often degenerate into a competition by ethnic politicians to bring home the racial bacon.

But, such is the way of the world … one that white people like poor Professor Prewitt, much less white Republicans, are unlikely to change.

In summary, Prewitt, despite his attempts to be as Politically Correct as possible, has opened up an important can of worms.

He hints at how racial / ethnic preferences interact with racial / ethnic categories on the Census: in essence, there can’t be disparate impact discrimination lawsuits about a lack of “statistical proportionality” in hiring and firing unless the government counts how many minorities comprise the denominator. So, if the government either stops counting a group or vastly expands the number of groups counted, the practicality of de facto quotas plummets.

Conversely, if the government isn’t offering money and prizes to individuals for identifying as nonwhite, the global tendency for people to want to be seen as white as possible (see, for example, retired Dominican slugger Sammy Sosa’s use of skin-lighteners) will mean that the percentage of Latin Americans and South Asians wanting to be seen as nonwhite will decline.

Thus Prewitt’s proposal to drop the race category in favor of a Nativity category–everybody listing their national ancestry. This would so complicate the ancestry statistics that quotas driven by employer fears of disparate impact lawsuits would become exponentially less feasible.

My guess, however, is that Affirmative Action for the descendants of black slaves in America is here to stay. The pre-MLK suffering of blacks in America has become too sacralized to challenge.

But still, Prewitt’s suggestion of adding a Nativity category is an excellent idea. For example, this proposed Nativity category would provide the statistical groundwork for denying Affirmative Action to the descendants of African black immigrants who didn’t suffer from slavery or even Jim Crow–such as, oh, Barack Obama.

Another way to achieve some of the advantages of adding a Nationality category is to open the Ethnicity category up to everybody. Instead of Ethnicity being only for Hispanics, let everybody list whatever ethnicities they feel like (e.g., 1/4th Amish, 1/4th Scots-Irish, 1/4th Peruvian, and 1/4th Armenian).

Just like Senator Akaka, what American doesn’t deserve to express his or her identity? Why should the government cold-shoulder everybody whose ethnicity isn’t Hispanic?

Granted, this would make a hash of attempts to enforce ethnic preferences. But I see that as a feature, not a bug.

Alternatively, we could abolish the Ethnicity category altogether.

If it’s politically necessary to continue preferences for some established Americans of Latin American background, let’s at least use the new Nativity category to deny them to people who chose, or their parents chose, to come to America, warts and all.

At the same time, the racial category could be expanded by making it more multiculturally sensitive. In 2010, over 18 million Hispanics (about 3/8ths of the total) checked the “Some Other Race” box on the Census form, in large part because the Census Bureau fails to recognize Latin American racial identities such as mestizo, mulatto, pardo, and Indio.

But above all, let’s not give preferences to the Spanish-surnamed whites who provide so much of the leadership and brains of the ongoing Katharine of Aragon’s Revenge.

Finally, Republican strategists should consider this: the GOP is inextricably linked with being the White Party. Thus, it should socially constructing the Census questions to elicit either a White Big Tent or a White Little Tent solution.

A Big Tent strategy, which I’ve argued for in the past, would try to incorporate as many nonblacks as possible into a sense of whiteness. Both Latin American, South Asian, and even East Asian cultures tend to be biased in favor of fairness. If we stop rewarding them for claiming to be nonwhite, many will tend to find they have more in common with whites than with blacks, and perhaps vote accordingly.

Alternatively, the GOP could pursue a Small Tent strategy: rewriting Census categories to exclude from whiteness Arabs, Armenians, fair-skinned Latin Americans etc. to speed up the process by which whites achieve that modern promised land of minorityhood.

But this is naïve. It’s not being a minority per se that makes you above criticism–it’s being Not White (or Not Straight, Not Christian, Not Male, Not Cis-Gendered, etcetera etcetera). Deep down, the zeitgeist is driven less by love for minorities than by hate for traditional majorities.

But, at least, racial realities must be understood well enough so that whites (formerly known as Americans) can get to play the Census game too.


Steve Sailer (email him) is movie critic for The American Conservative and writes regularly for Takimag. His website features his daily blog. His book, AMERICA’S HALF-BLOOD PRINCE: BARACK OBAMA’S “STORY OF RACE AND INHERITANCE”, is available here and here (Kindle)