Race has always been a provocative subject when the needs of science and statistics intersect with politics. Now that debate is once again heating up in France, as the planned introduction of “ethnic statistics” has caused a fierce dispute that touches the very heart of French republicanism.
According to a law that dates back to the French Revolution, and reconfirmed in 1978, French government officials are forbidden to collect information about a citizen’s ethnic or racial origins, whether real or alleged, when conducting a census or other efforts to gathering statistical information on the population.
There are two main reasons for this. The first is the republican principle, enshrined in the Constitution, that recognizes only citizens and does not accept any distinction among them due to origin, race, or religion. The second reason is historic: the painful and still vivid memories of the Vichy regime of WWII, when citizens’ “racial” and religious origin was stamped on national identification documents and was used as a key tool in rounding up French Jews for delivery to the death camps.
Today, the issue has returned to the forefront because of a new fight against racial discrimination, which appears to require more accurate measures of social inequality. Existing public statistics, it is believed, do not provide enough relevant information to analyze the possibility of discrimination in employment or housing. After all, without appropriate statistics, it is difficult to prove discrimination.
Indeed, many argue that by refusing to take into account distinctions linked to ethnic and religious origin, these distinctions are legitimized. One British social critic, quoted by the sociologist Dominique Schnapper, compares the behavior of the French, unwilling to mention ethnic discrimination, with English people of the Victorian era, who refused to talk about sex.
Supporters of gathering statistics on race and religion also look to the experiences of the United States, Britain, or the Netherlands, where census takers are free to inquire about ethnic origins and a citizen’s sense of belonging. Since 1990, the US has collected data on ethnic origins. Although the First Amendment of the US Constitution bans any religious test for citizenship or political office, ruling out questions on religious beliefs, it is possible to gather information on ethnicity, even in certain cases of multiple ethnic origins, such as “White,” “Black,” “Asian,” and “native American.”
In Britain, the concern about social promotion of minorities led to the introduction in 1991 of statistics that indicate ethnic status. As for the Netherlands, companies were obliged to report the ethnic composition of their workforces until the law was repealed in 2003.
But today’s French legislation is less rigorous than it seems. It distinguishes between anonymous files from random samples and established for scientific purposes, which may contain data about a person’s origins, and files that are not anonymous, which have direct consequences for the people concerned – and for which it is strictly forbidden to register any information about ethnic origins. The 1978 law allows government statisticians to ask “delicate” questions only if these questions are relevant for the survey and with the consent of the person polled.
But government statisticians have long been studying the national origins of immigrants, and are permitted to indicate the previous nationality of people who have acquired French citizenship. Thus, there is a distinction between mentioning the original nationality, which is allowed, and mentioning ethnic and racial origins, which is not.
Is it necessary to go further just because the indicators linked to national origin are not enough to identify discrimination – especially indirect discrimination – based on ethnic grounds? Some surveys show that the groups involved are doubtful about this.
Statistics are not only a reflection of reality: they help to shape it. Statistical categories often tend to become social categories. Not only are racial indications (White, Black, Arabic, Asian) very inaccurate in a world where racial mixing is now common, but, as François Héran, the head of France’s National Institute of Demographic Studies, argues, it is also necessary to prove that difference means inequality and that inequality necessarily means discrimination. Indeed, ethnic counting could merely reinforce the logic of community separation.
Given the desire to penalize ethnic discrimination, the push for government to gain this knowledge is understandable. But the state has other means to encourage equality on the basis of national, social, or economic criteria. In view of the risk of inciting fresh antagonism, gathering racial, religious, and ethnic statistics may not be worth it. The ban against ethnic and racial data is a taboo that should not be overthrown easily, and not without carefully weighing the risk to social peace.