Collecting ethnic statistics in Europe: a review
In its statement reproduced above and formulated six years after the adoption of EuropeanDirective 2000/43/EC implementing the principle
of equal treatment, also known as the Race Equality Directive, the European Commission emphasizes the crucial role played by statistics in activating anti-discrimination policies and increasing their capacity to ensure social cohesion and promote diversity and equality. It also mentions the persistent misunderstandings, and indeed strategic manoeuvres, which dog relations between data protection and the production of statistics on discrimination. This opinion was earlier voiced by the European Commission against Racism and Intolerance (ECRI) in its first General Policy Recommendation in 1996.1 Believing that ‘it is difficult to develop and effectively implement policies [...] without good data’, it recommends collecting ‘in accordance with European laws, regulations and recommendations on data-protection and protection of privacy, where and when appropriate, data which will assist in assessing and evaluating the situation and experiences of groups which are particularly vulnerable to racism, xenophobia, antiSemitism and intolerance’. This position has also been endorsed by the United Nations Committee on the Elimination of Racial Discrimination (CERD; see Banton 2001). The strong impetus given by international human rights bodies to
the collection of ethnic statistics sounds like a paradox. How is it that those agencies designed to combat racism are arguing for data collection on identifications that are linked to ethnicity and race? These data have been used historically to exclude, discriminate, and sometimes eliminate by mass murder, groups and individuals who have been identified and perceived as undesirable minorities (Seltzer 1998; Anderson and Seltzer 2001; Kertzer and Arel 2002). Moreover, ethnic and racial statistics have not only been used for evil purposes; their existence as such reflects a misconception of human nature and social life and fosters a reification of identities (Zuberi 2001). In this context, one could not expect a positive use of statistics which have been understood as the main vehicle for the essentialization of populations and the perpetuation of ethnic and racial hierarchies. The new generation of anti-discrimination laws and policies set by two equality directives enacted in 20002 has drastically challenged this framing of ‘ethnic statistics’ by fostering the need to monitor unfair treatment (Fredman 2008). Data collection has then become a major issue for EU equality bodies (Mannila 2005; Simon 2005; Makkonen 2007). The rise of a ‘multicultural Europe’ (Modood, Triandafyllidou and
Zapata-Barrero 2006) and the increasing diversity of most of European metropolis (Vertovec 2007) add to the pressing need to reconsider the way ethnic diversity is reflected in statistics. So if a move from colour-blindness to ‘ethnic consciousness’ may be anticipated in countries adopting positive action measures to combat discrimination, and/or in countries eager to reflect the ethnic diversity of their population in population statistics, the question of the different
methodologies for collecting ethnic data and the constraints created by the legal framework of data protection will be of crucial importance. If ‘ethnic statistics’ are to be collected, one way or another, what should they look like? There is no easy answer to this complex issue. Not only can no
agreement be found in social science upon a definition of ethnicity, but its conversion into a simple question on a census form is a challenge. It is even more so if official statistics, such as social surveys (Burton, Nandi and Platt 2010), take on board a constructivist approach to ethnicity (Brubaker, Loveman and Stamatov 2004; Wimmer 2008; Brady and Kaplan 2009). Rather than developing a ready-made solution which does not exist, this article aims at providing a state of the art on the collection of ‘ethnic data’ in official statistics in western Europe. It then points out some of the difficulties that are bound to occur when trying to cope with the complexity of ethnicity in the crude and simplistic world of statistics. Beginning with the fuzziness and heterogeneity of definitions and vocabulary used about ‘ethnicity’ in censuses, it stresses the different options chosen to collect data and their consequences for the contents of the categories. The use of alternative strategies, such as classification by the ‘second generation’ or third-party identification, will be discussed since they appear to represent politically acceptable proxies if explicit ethnic questions have to be avoided.