Will Counting Arabs Make Them Terrorists?
A Look at Whether the Census Impacts Our Daily Lives
The Census Bureau is proposing to change how they count race and ethnicity. Hispanic will now be counted as a separate race rather than an overlapping ethnic group, and a new race, for the Middle East and North Africa (MENA) will be added. There are a lot of reasons why various Census data users might like or dislike these changes. I’m not here today to tell you these changes are good, or to hash out the argument about how to define the MENA region, or quibble over how to define “Hispanic” or “Latino” or “Latinx” for Census purposes.
No, today I just want to shame one very embarrassing argument from the Heritage Foundation. Now, to be clear, I’m a pretty darn conservative dude, so Heritage is generally kinda my jam. Disappointingly, however, for today Heritage has handed over their coverage of the Census data change topic to a foreign policy and national security writer, who’s gone ahead and argued that we should never track race because tracking race creates special-interest voters and terrorists and, besides, the Founding Fathers don’t care about race.
What happened exactly? Here, let me show you:
I want to respond on three fronts here.
- How do other countries do this and what has the result been?
- What was the Census-taking policy of the early republic, which allegedly promoted assimilation?
- Is assimilation actually aided by not counting racial, ethnic, or ancestry categories?
*Note about the title: the article I’m criticizing did not say counting MENA people as being from the MENA would make them terrorists. It’s a click-baity title. So shoot me. But it does, I think, reflect the concern I assume is implied here by a national security writer talking about Census classifications: that poor assimilation will lead to not just uncomfortable but actually dangerous outcomes.
Most Countries Count Race
Here’s a map of countries with race counted in at least 1 Census since 1991:
Now, this map is a bit misleading, because not all countries are counting anything like what we call “race.” Take Russia, for example. Under Tsarist rule in the 19th Century, population surveys or censuses often did not ask about language or ethnic identity, until 1897 when, as a concession, they asked about language (but they asked about language in Russian, so, not exactly an unbiased sample). The Tsarist government suppressed local ethnic identity, and they thought limiting the presentation of these groups in official data might help. The Bolsheviks, however, managed to gain extra support from minority ethnic groups by promising to recognize them: and upon the first Soviet Censuses, dozens or even hundreds of ethnic groups were recognized. However, these were ethno-linguistic groups, not racial groups like in the US. Of course, the Soviets then tried to game their own system to get people to identify with state-provided ethnic groups. If you weren’t on the list somewhere, you were forced to identify as something on the list, because Soviets were really big about lists.
No but, seriously, this is a great test case. A government tried to use ethnicity-blind-ish censuses to promote a single national identity. This (allegedly, and I’ll return to the plausibility of this story later) gave extremists an “in” with minority groups that felt unrecognized. Those extremists took over and gave recognition to their client groups, but then used control of that recognition to try and push society towards a preferred political state exactly as the Tsarists had done. In the first case, non-racial census-taking was the tool for trying to manipulate ethnic identity. In the second, ethnically-specific census-taking was the tool. Today, Russia recognizes over 160 ethnic groups.
We can also look at at modern authoritarian states, such as Uzbekistan. Uzbekistan has not had any Census since 1989. They do a periodic “10% sample” of debatable quality, but even that hasn’t tracked ethnicity since 1996. Why no full Census? Well there are many reasons, but one suggestion by some observers is that the government doesn’t want to acknowledge how large the ethnic minorities in the country, especially Tajiks, have become. In this case, the authoritarian government avoids information provision in order to reduce the prevalence of ethnic minorities.
Similarly, Rwanda’s first post-genocide Census did not track Hutus and Tutsis, although it did track national-origin still, even aside from citizenship. The French Census meanwhile, along with much of Europe, is devotedly non-racial. The exception to the European norm of non-racial Censuses is in the British Isles, where Ireland and the UK do track race or ethnicity in the Census. Here we see a difference from the norm, in that Ireland and the UK tend to be the most accepting of other races. Ireland and the UK are generally regarded as being far better at integrating immigrants than, say, Germany, France, or Belgium (hence why France and Belgium in particular have high rates of radicalization and ISIS-membership among their Arab and Muslim populations).
You’re probably getting an idea of what my point is here. Governments that try to use the Census as a policy tool generally either fail, or succeed, and the result is a humanitarian catastrophe. You can’t use the Census to change people. It does not work. Ethnic identities are not fixed, but they are unlikely to be changed by government bureaucrats. It is amazing for Heritage, which usually has such a low opinion of government’s power to get things done, to here assert that something as mundane as counting people who identify as from the Middle East would destroy assimilation efforts. More to the point, of those countries that ignore race and ethnicity, the actual experience is very mixed. In Rwanda this choice has come after over a decade of healing and reconciliation, so it’s hard to say exactly what the impact has been, but it seems plausible that avoiding topics of ethnicity have been helpful for them. Meanwhile in Uzbekistan and France, avoidance of race and ethnicity serves to aid the interests of authoritarians and alienate ethnic minorities further.
What should we make of this? Perhaps this: in the context of a huge government-led effort to force embittered people to get along and mandate desegregation and affirmative action, as we see in Rwanda, a choice to avoid a Census count of ethnicity probably makes sense. Some ethnic groups may fear being counted, and for good reason! And when your government is laser-focused on a huge social engineering project like post-genocide reconciliation, it makes sense to deploy the Census for some ancillary backup.
But most American conservatives dislike the idea of social engineering. Few support aggressive affirmative action campaigns. It is rare to see conservatives calling for radical housing programs to encourage residential and educational integration. The Rwandan experience is not the USA.
Rather, the case of the UK or France is more apt. Integration programs are politically fraught. Ethnic minorities often feel alienated and persecuted when excluded from official recognition, especially given that we have always provided official recognition to major groups in the past. Comparing the UK and France, it is difficult to believe that a person could prefer the racial politics of France to the racial politics of the UK. Certainly no national security writer and foreign policy expert at the Heritage Foundation should be holding up France’s policy towards its racial minorities as exemplary.
(Sidenote: India doesn’t count race…. and has been identified in various sources as one of the most racist countries on earth.)
How Did the Founding Fathers Do It?
In the US, meanwhile, we ask all sorts of stuff in our Census. Last year, Pew produced a wonderful little tool letting you see how the Census has asked about race from 1790 to today.
On every Census since the first one, the US has asked about race.
Let’s be clear. This idea that somehow racial categories are a product of recent Census progressivism mind-numbingly dumb. George Washington filled out a race on his Census form. And you know who presided over that first Census? Thomas Jefferson. And you know what he did? He asked about race.
For terrible reasons, to be sure! Those earliest questions were really aimed at the question of slavery, since enslaved people were counted differently than free people. But even this focus began to change in 1850, when a question about mixed-race people was added. In 1860, “Indians” not living on reservations were counted and classed as such. So to be clear, at least since 1860 something like the modern system for tracking race has been in place. The Heritage article says that things were somehow different for the first two centuries of American history.
Maybe if you begin your count in the 1600s.
But that’s not all! Since 1850, we’ve also asked, at various times, about ancestry, place of birth, language, ethnic identification, origins, and other topics. So we didn’t just know how many black people there were, but how many Czechs, and Irish, and Germans. And how many German-speakers. And how many German-born. This great broadening of the Census occurred because a Census Office was established for the first time, led by the all-but-unknown J.C.G. Kennedy and J.D.B. DeBow, who in my opinion should be viewed as the fathers of the modern Census. By 1930, detailed questions about parentage were being asked.
In the American Community Survey, which is the successor to the Long Form Census, we ask about race, ethnic identity, place of birth, languages used, and ancestry! So, for example, we already ask the questions that Heritage is fuming about: we ask about ancestry. I can add up the MENA ancestries easily enough. Turns out, people claiming MENA ancestry are about 1% of the population. Now, I suspect the number we’d get from a MENA race would be a tad bit higher due to better coverage, but probably not tons higher.
To be clear: during the period of high immigration and aggressive assimilation, the Census already asked questions far more detailed than many of the questions we ask today. The Census used to ask about churches, and detailed questions on literacy and disability, and other topics we no longer include, such as direct parentage. The claim that somehow the currently detailed questions are the tiniest bit causal in regards to the pace of assimilation is just silly. Census forms have very little impact on integration one way or another, but whatever impact they do have is likely to be complex. Yes, they allow groups to form a sub-national identity; but they also allow existing sub-national identities to feel included and recognized in the larger national story. It’s no surprise we started asked detailed questions about immigration and origins in 1850, just 15 years after the major 19th century migration surge began.
How Do We Promote Assimilation?
I’m on record as sharing many of the Heritage piece’s author’s concerns about immigrant integration. Nor am I comfortable with a political culture that prioritizes ethnic solidarity over national identity and patriotic community. But there’s literally no way to know if assimilation efforts work without effective data collection, and, as noted above, there’s just no serious evidence that this data collection alters immigrant integration.
Countries with abysmal immigrant integration sometimes don’t collect racial or ethnic data (France, Japan) and sometimes do (China, Russia). Countries that do a good job integrating immigrants (Canada, Australia) pretty well always track race or ethnicity, but not all countries tracking race or ethnicity do a good job assimilating immigrants. In other words: there’s just no indicator that how you count people in the Census has any meaningful impact, but what minimal association there is would seem to suggest that tracking race or ethnicity is associated with better integration, rather than worse. But, again, this association is extremely weak and the more plausible takeaway is that Census forms really have no impact whatsoever on their own.
Assimilation can be promoted many other ways: mandatory public service or a draft, strong emphasis on English-language adoption, cultural norms friendly to intermarriage, understandable and fair immigration laws, natives willing to try and adopt the best pieces of immigrant cultures (ex. American love of foreign food), good educational systems, and many opportunities for social mixing all go a long way to promote assimilation. Likewise, extreme economic, climatic, or political conditions can sometimes accelerate immigrant integration (though sometimes they can retard it as well).
But Census forms? They just don’t matter that much.
The reason is very simple: racial and ethnic identification arises from a wide range of factors. For example, politicians saying they would ban immigration for one group of people will tend to make people within that group pull together a bit more and form a stronger group identity. On the other hand, public programs that offer special rewards and enticements for certain groups can also foment such group identity. But more usually, the strongest force for ethnic formation is just daily life. Accent and language formation, intermarriage, geography, employment, patterns of churchgoing and socialization, these all tend to gravitate together over time and create stable communities of commonality which we can call “ethnic groups.” Pressure from outside can accelerate this process, but it’s inevitable one way or another without some kind of shake-up. The crucial thing though is that most kids have come to understand some basic facts about their identity long before they ever have to check a box on a form. We do not live in a post-ethnic society where government is forcing ethnic identities on us. The categories in the Census reflect some very broad categories of people that are in fact widely acknowledged to exist. The Census should offer options that reflect how people actually identify, not simply how we think they ought to identify.
What Good Are Census Forms, Then?
The data they provide does matter. With better data, we can find groups who are assimilating better or worse than others. Likewise, we can ensure that minority groups feel included in the American project. And this is the key to American assimilation. Unlike the French, or the Russians, or the Chinese, we haven’t assimilated incoming groups by totally destroying their ethnic background. We’ve assimilated by co-opting their background so that we own it more than their homeland does. Hot dogs and hamburgers. Pizza. Kindergarten. Virtually every cultural meme about New York. Barn designs in Pennsylvania. Any number of other “immigrant institutions” can be listed. We didn’t integrate by eliminating, but by assimilating, i.e. absorbing or adopting. This absorption is neither hindered nor hastened by more precise Census figures, but with more precise figures we can at least see when we’re making progress.
Plus, many users of government data aren’t part of the government at all. We don’t just provide the Census for the government, but for the public’s benefit. The Census is an exercise in national introspection, a time when we look back on the decade past (or year past for the ACS) and see how things have changed, how we’ve grown and developed, how the American project has borne fruit. This data is used by local governments, refugee relocation organizations, schools, researchers, hobbyists (hi guys!), and, perhaps most importantly of all, government data is used extensively by private businesses that could never undertake research of such scope and neutrality on their own.
So You’re Saying We Should Encourage Ethnic Labels?
Eh, no.
See, there’s a red herring at the core of the argument to which I’m responding. It presumes to be saying we should oppose a society in which everybody’s first identification is racial and we interpret everything through a highly racialized lens. That view seems agreeable enough. We all want to live in that “I Have a Dream” end-state, right?
But adding or removing the MENA race, or shifting Hispanic identification from the “ethnicity” line to the “race” line has literally zero impact on that wider social issue.
Meanwhile, the actual argument being made here is that any racial categorization is inherently “fractionalizing,” and therefore that the government should not track race at all. This is a radical view that the article in question does not explicitly espouse, but it’s impossible not to see this as the obvious implication. Without race-specific data, we’d never know if one race was systematically impoverished. We’d never know if one race didn’t have adequate ballot-access. We’d never know if previously-measured inequalities were fixed!
But more to the point: if we stop measuring at least some broad metrics on race, following the identifications that actually prevail in society on the whole, we will be breaking with centuries of established American tradition.
So If You Oppose Adding MENA and Hispanic As Races, You’re Racist, Right?
No.
There are many good reasons to oppose these changes. If Heritage really wants to oppose them, they can at least come up with better reasons. For example: the MENA definition is clearly a haphazard and kinda insulting conflation of extremely different ethnic groups. The four that come to mind right away are Iranians, Arabs, Israelis, and Turks. These groups are very different. Lumping them together could be seen as a wee bit odd. Then again, lumping together Asians is just as odd.
Speaking of Asians, you could argue that, as denizens of “West Asia,” Middle Easterners should be classed as Asian instead of the current Caucasian.
You could also argue that while Arabs, for example, are seen as a group, “Middle Easterners” are not an identifiable group, so Census here is propagating a definition at odds with actual societal definitions. So maybe “Arab” should be its own race, but others shouldn’t be. But then that comes across as pretty heavy-handed too.
Or maybe you think MENA should be more widely defined. Maybe you think it should include all Turkic, Semitic, or Iranian-speaking peoples, so should stretch all the way to the Uighur populations of western China and some of the people of Afghanistan.
Or maybe you just want to argue that at just barely 1% of the population and lacking the historical claims to special recognition made by Native Americans and Pacific Islanders, MENA populations aren’t large enough to justify a special Census category. That opinion is defensible on its face.
These are all viewpoints that can be justified. What’s not justifiable, however, is the view that we shouldn’t track racial data because the Founding Fathers wanted to promote assimilation by eliding racial or ethnic differences. That view is entirely unsupportable.
Check out my Podcast about the history of American migration.
If you like this post and want to see more research like it, I’d love for you to share it on Twitter or Facebook. Or, just as valuable for me, you can click the recommend button at the bottom of the page. Thanks!
Follow me on Twitter to keep up with what I’m writing and reading. Follow my Medium Collection at In a State of Migration if you want updates when I write new posts. And if you’re writing about migration too, feel free to submit a post to the collection!
I’m a graduate of the George Washington University’s Elliott School with an MA in International Trade and Investment Policy, and an economist at USDA’s Foreign Agricultural Service. I like to learn about migration, the cotton industry, airplanes, trade policy, space, Africa, and faith. I’m married to a kickass Kentucky woman named Ruth.
My posts are not endorsed by and do not in any way represent the opinions of the United States government or any branch, department, agency, or division of it. My writing represents exclusively my own opinions. I did not receive any financial support or remuneration from any party for this research. More’s the pity.
