The Challenge of Measuring Immigrant Origin and Ethnicity in Europe

Paper prepared for “Changing Cultures: European Perspectives” Conference of the ESA Research Network for the Sociology of Culture Ghent, 15-17 Novem ber 20 0 6 Th e ch alle n ge o f m e as u rin g im m igran t o rigin a n d e th n icity in Eu ro pe . Principal authors: Dirk J acobs (Université Libre de Bruxelles) Marc Swyngedouw (Katholieke Universiteit Leuven) Laurie Hanquinet (Université Libre de Bruxelles) Véronique Vandezande (Katholieke Universiteit Leuven) Prepared with the valuable aid of: Ana Paula Beja Horta (Universidade Aberta) Maria Berger (University of Am sterdam ) Mario Diani (University of Trento) Amparo Gonzalez Ferrer (Fundacion J uan March) Marco Guigni (University of Geneve) Miruna Morariu (University of Geneve) Katia Pilati (University of Trento) Paul Statham (Leeds University) (for the m om ent the sole responsibility lies with the principal authors – draft not for citation) In tro d u ctio n Counting and classifying inhabitants of foreign origin and/ or identifying them as being part of ethnic m inorities is often a sensitive issue. Categorization of certain social and ethnic groups when m onitoring the population is an im portant tool in developing adequate policy, am ong other things to be able to tackle discrim ination and unequal opportunities. Although often m eant to improve the situation of those being counted, census m aterial and statistical data have at repeated tim es, however, been m isused to single out ‘foreign’ elem ents in order to contain, m istreat or even deport and – in the worst case scenario - to exterminate them . As a result, discussions of the measurement of ethnicity or foreign background often provoke strong feelings. Furtherm ore, although seem ingly a technically neutral tool, statistics are often the product of a particular social and political context. In the context of ethnicity, they m ay reflect ad hoc dom inant views on insider- outsider relations and thus become contested over time. Moreover, when used as a basis for affirmative action, specific dem arcations can have im portant consequences for vested interests of particular groups. There is no uniform European system for ethnic categorization. Different European nation- states use the m ost diverse statistical constructions of foreign origin or ethnic minor ity populations . Several countries traditionally even shun from producing such data. This m akes international com parison a very difficult endeavour. Anyone wanting to perform com parative research on im m igrants or ethnic m inorities in Europe is unavoidably confronted with the m ost diverse types of nation al statistical data and has to opt for ad hoc solutions. Attem pts at international comparison can thus be very tricky due to data characteristics. It is im portant that researchers are aware of these problem s and do not sim ply accept data (especially in com parisons) at face value. In this article we em bark on a com parative explorative study of the way in which im m igrant background 1 and ethnicity is taken stock of by national statistical institutes in a set of European nation- states. We restrict the scope of the article to an exploratory study of the existence (and non existence) of official definitions and related operationalisations. We are m ainly interested in looking at potential applications of statistical data on foreign origin and ethnic background for academ ic research. We will then reflect on the question what kind of data we would need in the future. D iffe re n t trad itio n s acro s s Eu ro p e In all European states the classification and counting of nationals and foreigners is regarded to be a legitim ate endeavour. When exam ining m igration and international m obility, nationality is often the m ost readily available criterion for distinction. In the m em ber states of the European Union the category of ‘EU citizen’ has recently becom e sort of an interm ediary category in between the ‘national citizen’ on the one hand and the ‘genuine foreigner’ on the other hand (J acobs & Rea, 20 0 5). Indeed, in a growing num ber of policy m atters the process of Europeanisation has lead to equal rights for residents from other EU mem ber states, to which other foreigners are not necessarily entitled. As a result, in all kinds of official statistics increasingly the distinction is being m ade between ‘EU citizens’ on the one hand and ‘third country nationals’ (inhabitants who do not hold the nationality of one of the EU Mem ber States) on the other hand. The latter are often presented as being m ore ‘foreign’ than the EU- citizens. Take for instance Italy which distinguishes foreigners as being either EU- citizens or “extracom unitari” (non- EU- citizens). Tellingly, the first law focusing on working conditions of im m igrants referred to extracom unitari to pinpoint workers of non- EU nationality 2 , whereas the 1998 im m igration law and its 20 0 2 modification 3 identify ‘foreigners’ as citizens of st ates who are not members of the EU and stateless people. Most countries in Europe lack an official (legal) definition of people of foreign origin or of ethnic m inorities (of foreign origin). They only use the distinction between nationals and non- nationals (or ‘foreigners’). As a side effect some forms of ethnic disadvantage and discrim ination can often not be m easured by m aking use of official population statistics since im m igrants disappear as a specifically identifiable group once they have acquired stat e citizenship. At first sight, the existence of uniform statistics on non- nationals does seem to offer some possibilities for comparative work between nation- states. The fact that all countries produce statistics distinguishing nationals and non- nationals within their population does, however, not guarantee us a reliable basis for international com parison. Nationality legislations apply naturalisation conditions so differen tly, and the logics of ius sanguinis (blood links as the basis for nationality attrib ution) and of ius soli (residence as the basis for nationality attribution) in such varying m anners, that one would be com paring apples and pears. It is m uch m ore likely for a 1 When we refer to figures on im m igrants in this paper, our focus is on nationals with an immigrant background and on foreigners with a legal residence status, but not on undocum ented m igrants. 2 Italian law 943/ 1986. 3 Italian law 286/ 1998 and Italian law 189/ 20 0 2. person with a foreign background to have state citizenship in Belgium or the Netherlands than it is in Austria or Switzerland, to give but one exam ple. This m akes a correct com parison between these sets of countries very difficult. Across all EU m em ber states im portant differences in the m odes, requirem ents and procedures for the attr ibution and acquisition of citizenship persist (Cantisani & Greco, 20 0 6a). Counting nationals and non- nationals has nothing controversial and little critical questions are raised with regard to the growing im portance of distinguishing EU- citizens and third country nationals. Counting and classifying individuals on the basis of their ethnic origin or even ‘race’ is, however, to a far lesser degree seen to be acceptable in continental Europe, while it is a standard operating procedure in the United States, t he UK, Canada and Brazil. In the latter countries interethnic relations are judged to have as m uch im portance and relevance as gender or class relations. Official statistics routinely distinguish races and this does not seem to provoke large scale criticism . Indeed, in recent years, the statistical construction of ethnic and racial groups in the US has mainly provoked a debate with the possible introduction of the category of a ‘mixed race’ in the 20 0 0 Census (Amaro et Zambrana, 20 0 0 ; Krieger, 20 0 0 ; Lee, 1993; Nobles, 20 0 0 ; Riche, 1999; Sondik et ali, 20 0 0 ; Waters, 20 0 0 ; William s et J ackson, 20 0 0 ), without fundam entally questioning the whole racialising fram ework to start with (but see Kertzer & Arel, 20 0 2). In continental Europe, ethnic classifications often have no com parable institutional or statistical translation, although they are very currently being used in day to day life. One could argue in favour of such classifications as tools to m easure ethnic and racial discrim ination. Moreover, targeted groups m ight m obilise them in order to defend their interests in policies of redistribution of social goods (jobs, housing, etc.). However, the possibility of using such categorisations to tackle discrim ination apparently does not provide sufficient justificatio n for the construction of ethnic categories in official statistics. The m isfit between the vastness of the debate, particularly in France, which is triggered by the m ere possibility of constructing ethnic categories in statistics (see for an overview: Spir e et Merllié, 1998 ; Blum, 20 0 2) on the one hand and the limited number of studies pertaining to the actual possibilities of operationalisation of ethnicity on the other hand (Bulm er, 1996 ; Sim on, 1997, 1998; Aspinall, 20 0 2; Lie, 20 0 2) attests to the strong political dim ension of the m atter. The political passions which feed the scientific debate strongly dem onstrate that the definition of statistical categories on ethnicity and race is not m erely a technical m atter. The construction of these categories is influenced by ideologies, visions about nations and visions about interrelations between social groups. An additional elem ent which further com plicates the debate is that they are also perform ative: the use of ethnic categories reinforces the ethnicisation an d racialisation of society. Once they are socially constructed, these categories gain their own life. Schem atically one can distinguish two traditions related to ‘ethnic statistics’ in Europe. In France, Germ any and m ost southern European countries, the dom inant statistical categorisations m erely distinguish individuals on the basis of their nationality. It basically boils down to a lim itation to two categories: the national and the foreigner. Often an additional distinction is m ade am ong the foreign population between those com ing from other EU- m em ber states and those who don’t. With the introduction of the category ‘im m igrant population’ – including all people born as a foreigner abroad - France has nevertheless tried to m ake the dem ographic con tribution of immigration to its population a bit m ore visible without however distinguishing ethnic groups. In contrast, m ost northern European countries have been producing data on the ethnic and/ or foreign origin of their populations in a m ore detailed m anner. The UK has for instance a system of self- identification of ethn icity. Norway keeps track since 1994 of its ‘im m igrant population’ by counting the num ber of persons who neither have parents nor grandparents born in Norway (Vassenden, 20 0 5). National backgr ound is kept track of by looking at the person’s own, their m other’s or possibly their father’s country of birth. Separate statistics are produced for different regions of origin 4 . Although in strict legal term s ‘foreigners’ are non- nationals, Denm ark also uses the term ‘foreigners’ (Udlaendinge) as a statistical concept, regardless of citizenship, to refer to im m igrants (people born outside of the country whose parents are either foreign citizens or both were born outside of Denmark) and descendants (people born in Denm ark by parents whom neither of which is Danish citizen born in Denmark). Specific num bers are produced for ‘lesser developed nations’ (countries outside of Scandinavia, the EU and North Am erica) 5 . The Dutch equally try to count their population of foreign origin (regardless whether they hold Dutch nationality or not) since 1995 on the basis of country of birth of the parents of its residents. The Dutch have adopted the category of “allochthones” to label the ethnic or foreign origin of segment s of its population and can m ake distinctions with regard to countries of origin. This category, at first m ainly statistical, has gradually becom e adopted in ordinary language as a particular social category with a num ber of specific – often negative - connotations. In terestin gly, Belgium is caught somewhere in between these two traditions (J acobs & Rea, 20 0 5) and finds itself in a stalemate position. The French speaking part of Be lgium tends to follow the French tradition of refusing ethnic categorisation, while the Flemish (the Dutch speaking part) try to copy the Dutch model in distinguishing “allochthones” and “autochthones”. In Flanders, as in the Netherlands, the term “allochthone” is widely used in academ ic, political and institutional circles to refer to immigrant (mainly non- EU origin) inhabitants. This difference in conceptualisation within one and the sam e state has, however, not lead the federal Belgian state – which is still in charge of population statistics – to produce any official data on the num ber of “allochthones” on the national level. Federal law stipulates that the national statistical office, l’Institut N ational de Statistique (INS), does not have the authorisation to produce any statistics relating to ethnic origin: “In no case what soever can the investigations and statistical studies of the national institute for statistics be related to the private life, the political, philosophical or religious opinions or activities, race or ethnic origin” 6 . The INS seems to follow this guideline in a strict m anner and thus refrains from producing statistics on place of birth of parents and grand- parents (one of the criteria of the Flem ish definition of “allochthone”). In an internal note, the national statistical office INS comments the article in the following way: 4 For instance Nordic countries, Non- western countries (Asia including Turkey, Africa, South and Central America and Eastern Eur ope) and Third world countries (Asia including Turkey, Africa, South and Central Am erica). 5 Statistical Yearbook of Foreigners in Denm ark 20 0 2. Annotated edition. (Danish Ministry of Refugee, Im m igration and Integration Affairs). 6 Article 24 quinquies of the law of 4 J uly 1962 with regard to(?) to public statistics, m odified by the law of 1 August 1985. “Exluding all research into political, philosophical or religious opinions or activities, into race or ethnic origin and into sexual life, article 24 quinquies allows the national institute for statistics to rem ain sheltered from all controversy. The national institute for statistics has to be a neutral and independent organism and a trustworthy and credible instrum ent for the adm inistration of the country” (INS, 1986). The result is that the Flem ish adm inistration has a detailed definition of “allochtones” but does, at the same time, not have the appropriate instruments to count how many there are (J acobs & Rea, 20 0 5). Eu ro p e an e ffo rts to w ard s h arm o n izatio n o f s tatis tics Is the deadlock situation in Belgium predictive of the way in which the issue will be handled on the European level? If Belgium up till now has failed to find a com prom ise between the two traditions in the m atter, will Europe do better? The European com m ission indeed hopes it will be able to push EU m em ber states in the direction of a uniform statistical apparatus for counting immigrants by going b eyond the criterion of nationality. The Com m ission has recently (Septem ber 20 0 5) proposed a regulation on com m unity statistics on m igration, attem pting to harm onize data collection on this topic (COM 20 0 5, 375 final). The regulation, currently (end of 20 0 6) lined up for discussion in the European Parliament 7 , proposes the production of annual statistics disaggregating the population according to country of birth8 . It equally calls for (annual) statistics on the num ber of persons acquir ing citizenship and form erly holding citizenship of another state (or being stateless). This would indeed already constitute an im portant step forward in facilitating transnational comparison . There are, however, still quite som e technical issues to be resolved. Coverage of data on nationality acquisition currently varies significantly across countries – som e counting all changes in nationality, others lim iting them selves to certain procedures9 . Tailoring will be necessary to allow for genuine comparability (Cantisani & Greco, 20 0 6b). Even if technical m atters are sorted out, the produced figures will still not be com pletely satisfactory for all research purposes (Cantisani & Poulain, 20 0 6). One of the fundam ental problem s will be that new acquisitions of citizenship can be counted but that the exact m agnitude of the existing stock of foreign origin citizens will remain unknown if this kind of data is not available for earlier periods 1 0 . 7 Codecision procedure reference COD/ 20 0 5/ 0 156 Currently annual information on country of birth (combined with age and sex) is not available for the following EU member states: Czech republic, Germany, Estonia, Greece, Italy, Cyprus, Lithuania, Luxem burg, Hungary, Malta, Poland and Portugal (Cantisani & Poulain, 20 0 6: 20 0 - 20 0 1). 9 Tricky issues are (a) incom plete registration m echanism s for sem i- automatic types of acquisition and (b) inclusion (or exclusion) of atypical forms of acquisition (for instance naturalisation of people living abroad, acquisition through adoption, reacquisitions of citizenship by form er citizens, etc.) (Cantisani & Greco, 20 0 6b ; Perrin, 20 0 6). 10 In cases where the stock cannot be determined using inform ation of population registers, an alternative ad hoc solution m ight be provided by a cohort approach using retrospective data from the census (Perrin, 20 0 6). 8 Th e so- called second and third generations, an intensively studied group in social sciences and the focus of quite som e political debate and policy m aking, will furtherm ore still rem ain out of sight since they were born in the country to which their parents or grand- parents m igrated. Moreover, in a num ber of countries (Belgium , France, Spain, the Netherlands, Germ any, Ireland, Portugal and the UK) they have been attributed citizenship at birth, according to the place of birth of the parents or certain residence requirem ents (Cantisani & Greco, 20 0 6a: 174) and thus will not be made visible in statistics on acquisition of citizenship. A classification based on country of birth of parents or self- identification, which would be able to resolve this problem, is not proposed in the Regulation. The Dutch notion of “allochtone” (based on country of birth of parents) does allow to pinpoint this segment of the population. The question should, however, be raised to what exten t that logic should be exten ded to the level of grandparents – which the Dutch are increasingly doing. Perhaps even m ore im portantly, the issue should be resolved what to do with the ‘offspring’ of m ixed couples in your classification system . Some would argue a self- identification system is better, as it ex ists in the UK. Such a system is, however, equally not devoid of problem s and pitfalls. Decisions have to be m ade on the categories am ong which respondents can choose. Mem bers of visible m inorities can furtherm ore deliberately choose to classify them selves as being part of the m ajority group. This would then, of course, be their legitim ate choice but would at the sam e tim e frustrate correct analysis of discrim inatory practices against so- called ‘visible m inorities’. Comparing the situation of immigrant orig in (ethnic m inority) groups in different European countries m ight in our view profit from using data which does not rem ain lim ited to the current citizenship status of inhabitants. If all national institutes of statistics would gather data – and make them available - on indicators such as nationality at birth, place of birth, nationality of the parents, place of birth of the parents or nationality at birth of the parents, this would open up possibilities for com parisons which are not (or at least less) blur red by differences in nationality legislations. Most European countries today, however, do generally not possess a lot of readily available data on these ‘alternative’ indicators. Moreover, when state adm inistrations do collect this kind of data, they are often not publicly accessible for research purposes and often not even available to the national statistical offices. Som e countries have annual data, others only have (ten yearly) census data. Particular data m ight be collected for adm inistrative purposes but this does not guarantee that they are available for statistical analysis – let alone that they would allow for international com parative work. Let us just address the situation in a selected num ber of countries to clarify this point. In Germ any , the national institute of statistics (Bundesam bt für Statistik ) has only data based on current nationality (annual data) and place of birth (census data). Sw itzerland has information on current nationality, nationality at birth and place of birth 1 1 . Info on nationality of the parents is present for children born out of wedlock and for offspring of m ixed couples (in order to be able to apply particular regulations on ius sanguinis). In Spain current nationality and place of birth are equally available. Data on nationality of the parents and place of birth of the parents is only 11 This inform ation being, however, presented as “born in Switzerland and nationality” and “born in a foreign country and nationality”. available for individuals who (still) live in the sam e household as their parents. In Portugal the National Statistics Institute basically only has inform ation on current nationality (annual data), although the Census equally delivers inform ation on the place of birth. In Italy information is available on current citizenship (annual data) and place of birth (census). Inform ation on citizenship at birth is lim ited to data whether one had Italian citizenship at birth or not (through a Census question). Technically speaking, data on nationality of the parents is available for people who were born and reside in Italy. Belgium , to give a last exam ple, is technically capable of trackin g place of birth, nationality of parents and place of birth of parents in its detailed Population Register but due to legal obstacles the National Institute for Statistics can only present inform ation on current nationality. The National Institute for Statistics has data on nationality at birth for the 1991 Census but no longer for the 20 0 1 Census (when that question was struck). Table 1. Annual statistics across the 25 EU- m em ber states, Norway and Switzerland Population register Annual statistics on citizen sh ip Annu al statistics on coun try of bir th Annual statistics on cou n tr y of birth of paren ts Austria Y Y Y N (dou ble ch eck n eed ed ) Belgium Y Y N N Cyp r u s Y N N N Czech Republic Y Y N N Den m ar k Y Y Y Y Eston ia Y N N N Fin lan d Y Y Y N Fr an ce N Y* Y* N Ger m any Y Y N N Gr eece N N N N Hungary Y Y N N Irelan d N Y* Y* N Italy Y Y N N Latvia Y Y Y N Lithuan ia Y Y Y N Luxem bourg Y Y N N Malta N N N N Neth erlan ds Y Y Y Y Nor way** Y Y Y Y Poland Y N N N Portugal N Y N N Slovak Republic Y Y Y N Sloven ia Y Y Y N (dou ble ch eck n eed ed ) Spain Y Y Y N Swed en Y Y Y Y Switzerlan d*** Y Y Y N (d ou ble ch eck n eed ed ) Un ited Kin gdom N Y* Y* N * as an estim ate (based on survey) ** not an EU m em ber state Table 1 provides an overview of available annual statistics in the 25 EU m em ber- states plus Norway and Switzerland. All countries have Census inform ation on citizenship and country of birth1 2 , but not all countries can provide annual data. It should equally be stressed that the methods used to produce annual statis tics are quite diverse. Som e countries rely directly on population registers (Belgium , Denm ark, Latvia, Finland, Sweden, the Netherlands, Spain, Austria, Norway), while other countries m ake estim ates (m ostly based on Census data and other inform ation) (Cantisani & Poulain, 20 0 6). As a result of all this divergence in available data, international com parisons are doomed to stick to the lowest common denominator, which is in practice often the sim ple distinction between nationals and non- nationals. It is for this reason that the European Com m ission in its regulation wants to develop a new European wide system including data on country of birth. The OECD has already tried to establish a data base with detailed inform ation on the foreign born population in alm os t all OECD- countries (Dum on t & Lem aitre, 20 0 5). The problem with this database 1 3 , however, is that it is m ainly based on (10 yearly) Census inform ation and thus will rapidly be outdated. One could say it is at least a start, but it is clear this does not solve com parative problem s and sheds no light at all on the presence and num erical im portance of second generation m igrants. As we have stated before, some countries in the EU have tried to conceptually classify their population according to (pseudo- ) ethnic criteria by clearly moving beyond the sim ple distinction between nationals and foreigners and going further than country of birth. In the rem ainder of this article we will discuss the (proto)typical cases of the Netherlands (which has a sim ilar approach as the Nordic countries) and the UK in more detail. We will first, however, take a closer look at the notion of ‘ethnic m inorities’. Th e n o tio n o f ‘n atio n al m in o ritie s ’ an d ‘e th n ic m in o ritie s ’ Most countries lack a clear definition (and operationalisat ion) of ethnic m inorities. If public authorities do use the notion of m inorities it is often focused upon so- called national m inorities, in m ost cases historical linguistic m inorities, in line with the Fram ework convention on the protection of national m in orities of the Council of Europe. In the EU only France, Belgium , Luxem bourg and Greece have not signed or ratified this convention. The Baltic states have a particular statistical interest in keeping track of the ethnic composition of their populations . Estonia talks about “ethnic nationality” in this context and distinguishes the following groups: Estonian, Russian, Byelorussian, 12 For France the data are estimates. No Census data exists on country of citizenship for the UK (Cantisani & Poulain, 20 0 6). Estonia has Census data on the country of birth of the parents. Estonia equally has data on the self- indicated ethnic com position of its population, with an indicator of what they call ‘ethnic nationality’. 13 http:/ / www.oecd.org/ docum en t/ 51/ 0 ,2340 ,en _ 2649_ 33931_ 340 630 91_ 1_ 1_1_ 1,0 0 .h tm l Finnish, Tatar, Latvian, Polish, J ewish, Lithuanian, Germ an and other ethnic nationalities. Th e data on “ethnic nationalities” comes from the Census in which people are asked to self classify them selves. Parents determ ine the ethnic nationality of their children. If the child's m other and father were of different “ethnic nationalities” and the parents cannot agree on the ethnic nationality of their child, the ethnic nationality of the mother is preferred. Lithuania distinguishes as ‘ethnicities’ Lithuanian, Russian, Polish, Belarussian, Ukrainian, J ewish, Latvian, Tatar, Germ an, Rom any and Other. The Baltic states are som ewhat particular in this respect and their stance should be interpreted in the light of difficulties in dealing politically with (in particular) its Russophone m inorities after independency from the Soviet Union. It is clear ethnic m inorities in the context of the Baltic states does not refer to ‘classic’ im m igrant origin populations. Although the term “ethnic m inority” is widely used both in official policy and in public discourse in the UK, it is not a recognised legal term . Instead, the UK Governm ent defines a national ethnic m inority as a “racial group”, under the Race Relations Act 1976. A “racial group” is a group of people defined by race, colour, nationality and ethnic or national origins, and the Race Relations Act 1976 protects all racial groups from unlawful discrim ination. The courts have explicitly recognised Rom any Gypsies, Irish Travellers, J ews and Sikhs as constituting racial groups for the purposes of the RRA 1976 legislation. A m ore precise definition of “racial group”, based on ethnic origins, was given by the House of Lords 1 4 . In 198 3 the House of Lords em phasised that the word “ethnic” should be interpreted “relatively widely, in a broad, cultural/ historic sense”, but also observed that “the word ‘ethnic’ still retains a racial flavour”. Tribunals and courts have ruled on the basis of this that the English, Scots and Welsh, among others, are not racial groups by virtue of distinct “ethnic origins”. As we shall discuss in the specific section on the UK, “white” is considered to be a separate ethnic category. One country, the Netherlands, does have a strong tradition of using the term of ‘ethnic minority’ in public policy discourse to refer to immigrant origin populations without using an explicit racial discourse. In 1983 the Dutch government lauched a so- called “minorities policy” explicitly aimed at emancipation of officially defined categories of ethnic m inorities, with the objective of elevating the ‘ethnicized’ groups to equal social status with the indigenous groups in Dutch society, while at the same tim e propagating the ideal of a m ulticultural society. In the original Dutch governm ental discourse it was stipulated that the ethnic m inorities policy lim ited itself to those immigrants “for whom the presence is seen by the authorities as being t heir special responsibility (due to the colonial past or because they have been recruited by the authorities) and who find them selves in a m inority situation” (Minderhedennota, 1983: 12). The m inorities policy thus concerns on the one hand the Surinam ese, Antillians, Arubans and Moluccans (and their offspring) and, on the other hand, the Moroccans, Turks and other guestworkers from the Mediterranean area (Italians, Spanish, Greeks and (ex- )Yougoslaves) and their offspring. A foreign origin group is only considered to be an ethnic m inority group if one judges that the group is structurally trapped in a disadvantageous socio- economic position. As a result, gypsies and asylum seekers have equally become to be considered as part of the ethnic m inorities, but other groups as foreigners com ing from neighbouring countries (Belgium and Germ any) have not been defined as such. It is interesting to note that the Chinese have for a long time equally not been recognized as being an 14 Mandla v Dowell- Lee, House of Lords, 1983 IRLR 20 9 H.L. 1983 2 A.C. 548, 1983 1 All E. R. 10 62 ethnic m inority (as a policy category). In sum , the category of ethnic m inority was defined in a way cumulating both the social situation as criteria of foreignness. Th e n o tio n o f ‘allo ch th o n e s ’ in th e N e th e rlan d s Although the central term inology is still ‘ethnic m inorities’ in policies targeted at foreign origin groups the category of “allochthones” has in the meantime gained im portance through extensive use. The notion was introduced in the policy dom ain by the report Allochtonenbeleid (WRR, 198 9) of the academ ic advisory body for the governm en t (W etenschappelijke Raad voor Regeringsbeleid , in short: WRR). In this docum ent “allochthones” were defined as: “Allochthones are, generally speaking, all persons who com e from elsewhere and have durably settled in the Netherlands, including their descendants until the third generation, in as far as the latter want to consider them selves as allochthones. Minorities are allochthonous groups which find them selves in a disfavoured position: it has to be assessed periodically which groups have to be considered to be m inorities” (WRR, 1989: 10 ). It was also in this report that a plea was held to install a system of ethnic registration which goes further than the distinction between nationals and non- nationals. The report preferred a system of self- registration. In its reaction to the report, the Dutch governm ent, however, stated it preferred to stick to the notion of ethnic m inorities and it did not go into the m atter of ethnic registration. Although the notion of “allochthone” was starting to be routinely used in policy docum ents, it only got an operational basis in 1995, following the introduction of a new population adm inistration system at the m unicipal level (Gem eentelijke Basisadm inistratie, GBA). It was the national statistical office, the Centraal Bureau v oor de Statistiek (CBS), which de facto defined and constructed the new category of “allochthone” in a sem i autonom ous m anner, using inform ation com ing from the GBA. It is the definition of the national statistical office which would becom e hegemonic and is still the referen ce today. Since 1999 the CBS defines allochthones as: “every person living in the Netherlands of which at least one of the parents was born abroad” Note that the definition does not in itself suggest any racial or cultural connotatio n 1 5 , the criterion is place of birth of the parents. Place of birth of the parents is used as a proxy for foreign origin. It is ‘im precise’ as an ethnic category in the sense that it for instance equally includes children of Dutch expatriates. Before 1999, the CBS already used the category of “allochthone”: the allochthonous population was system atically counted on the basis of m unicipal data since 19951 6 . However, during the period 1995-1999, there were two definitions in use: an enlarged one and a restricted one. According to the enlarged definition, the allochthones were all persons who lived in the Netherlands and were either not born 15 It can equally be noted that, according to this definition, most members of the Dutch royal fam ily can be considered to be “allochthones”. 1 6 Figures are available on the basis of estim ations since 1972 and on the basis of data from the municipal administrations since 1995 (following the introduction of the Gem eentelijke Basisadm inistratie in 1994). in the Netherlands or were born in the Netherlands but had at least one parent which was not born in the Netherlands. In a more limited definition, the CBS only took account of people born abroad of whom at least one parent was equally born abroad and of people who were born in the Netherlands but who had two parents born abroad. In the year 1998, the CBS had the habit of priv ileging the restricted definition in its publications. Due to insistence by the governm ent, the CBS in 1999 however once again preferred to use the enlarged definition before finally opting for the new definition which is still in use in 20 0 6. The m ost recent definition of “allochthones” thus entails all people of the restricted definition, while adding all persons born in the Netherlands of whom at least one parent was not born in the Netherlands. The difference with the older enlarged definition is that it no longer includes people who were born abroad out of two parents born in the Netherlands. In the 1999 annual report regarding the m inorities policy, the governm ent stipulated it preferred to keep the children from “m ixed” couples in the new definition (as opposed to the old restricted definition). The (odd) argum entation was as follows: “The m ixed group is interesting because they seem to succeed better than the group of whom the two parents are born abroad » 1 7 . Whatever is the precise definition and operationalisation, in all cases the category of “allochthone” is broader than the one of “foreigner”, since it also includes people who hold the Dutch nationality. With the choice for place of birth as a criterion, the CBS could still trace people with Dutch nationality who originate fr om the form er Dutch colonies, without having to m ake any explicit racial distinction. As has been pinpointed before, these groups are official targets of the m inorities policy. Let us stress that the objective criterion of place of birth is com bined with a generational criterion in the notion of “allochthone”. At least in the CBS definition the third generation of im m igrants is in principle autom atically considered to be “autochthonous” and not “allochthonous”. The statistical administrative use of the category of “allochthone” by the CBS thus differs on this point from the proposition by the WRR, who (re)launched the category in 1989. In the definition of the national statistical office, “allochthone” is restricted to refer to the first generation of im m igrants (those born outside of the Netherlands) and to the second generation of people of foreign origin (born in the Netherlands but with at least one foreign parent)1 8 . Following its operationalisation by the national statistics office, the category of « allochthone » was increasingly used in policy docum ents, academ ia, public debate and the m edia. As a result, it was eventually even adopted in ordinary language. Not surprisingly, in the process the notion of “allochthone” underwent a change of m eaning and becam e increasingly used in ways differing substantially from its original adm inistrative definition. It began to be widely used to pinpoint people of Turkish, Moroccan, Surinamese and Antillian origin – the largest officia l “ethnic m inorities” - and for refugees from Africa, Asia and Latin Am erica. It was gradually bestowed with a connotation of the “non- white non- European Other”. Originally constructed as a m ere descriptive statistical category by CBS, the diffusion of the term 17 Our translation. Second Cham ber, 1999- 20 0 0 , docum ent 26815, p.5, note 2. In Dutch academ ia som etim es the term « one and a half generation » is equally used to pinpoint to children born abroad from immigrants of the first generation, who later cam e to the Netherlands in the fram ework of fam ily reunification schem es. 18 in ordinary speech acts led to a transform ation into a racial- culturalist category. It was now targeted toward everyone who was supposed not to have a “western” origin. European im m igrants and their offspring tended not to be included in the sem antic field of the notion, in contrast to its official definition. The pressure towards a racialising content was reflected in the statistical distinction which the CBS itself introduced in 1999 when distinguishing western allochthones and non- western allochthones 1 9 . This distinction is mainly used for statistical purposes in the field of education (given the established fact that nonwestern allochthones on average do worse than western allochthones), although it has not remained limited to that policy domain. Are part of the category of western allochthones according to the CBS: “the allochthones of European origin (with the exception of Turkey), of North- American origin, of Oceanic origin, of Indonesian origin and of J apanese origin”. Are part of the category of non- western allochthones according to the CBS: “people originating from Turkish, African, Latin- Am erican and Asian im m igration, except for people of J apanese and Indonesian origin”. The subdivision within the generic category of allochthones has thus more than an ethnic dimension. In the words of the national statistics office CBS the J apanese and Indonesians have to be excluded from the category of non- western allochthones because of “their socio- econom ic and cultural position” (http:/ / statline.cbs.nl) . The classification thus links up with two ideal typical contents of the immigrant: ethnic origin and inferior social origin. We can note that people of Indonesian origin are excluded from the category since a lot of (descendants of) Dutch colonizers ‘returned’ to Europe after the independence of Indonesia. In the definition of non- western allochthones the ‘im preciseness’ of the proxy of country of birth of parents was thus ‘corrected’ for a particular group of colonial expatriates (while at the sam e tim e introducing a new bias with regard to people of Indonesian origin without a genealogical link with white Dutch colonizers). As we have already stressed, the third generation of foreign origin is autom atically considered to be ‘autochthonous’ by the definition of the CBS. The category of “allochthone” hence does not function as an eternal racial category. Nevertheless, in ordinary life this lim itation of the definition of “allochthone” is not as strictly respected. Interestingly, although the CBS scrupulously avoids to use the term “allochthone” to designate the third generation, the national statistical office has tried to keep track of this third generation. Indeed, since 20 0 0 the CBS offers figures related to the “non- western third generation”, in which it classifies everyone who has at least one grandparents who was born abroad in a ‘non- western’ country (following the earlier distinction between ‘western’ and ‘non- western’). The data is produced in quite som e detail, allowing to distinguish those who have respectively 1, 2, 3 or all 4 grandparents of non- western origin. Specific data is provided for groups of Moroccan, Turkish, Surinamese and Antillian origin. It should be noted that com parable figures are not m ade available for the “western third generat ion”. A number of problems are evident when one uses the country of birth of parents or grandparents as a proxy for foreign origin or ethnicity. How m any generations (two, three?) can and should one continue to m ake distinctions between people on the basis of their ascendance? What rule should be applied to classify people with m ixed origins in one or the other category? What is the precise justification for making culturalist distinctions between western and non- western groups? Not to m ention the fact that quite som e people classified as being 19 Since 1999, statistical data on this distinction were produced, starting with data for the year 1996. ‘allochthones’ (in the Netherlands) , ‘im m igrants’ (in Norway) or even ‘foreigners’ (in Denmark), regar dless of their citizenship status, do not like this at all. One way out would be to allow people to classify them selves as is done in the UK. S e lf- id e n tificatio n o f e th n icity in th e U K In UK, the m ain criterion used when producing statistics on ethnicity is ethnic group, although the Office of National Statistics (ONS) also collects data on place of birth 2 0 (and religion 2 1 ) through the 10-yearly Census (and in the major household surveys). For England and Wales, statistics on ethnicity are collected and reported on by the Office of National Statistics, using Census data (from 1991 onwards, when a question on ethnicity was first included in the Census), and also on the following specific areas by key governm ent departm ents: housing (Office of the Deputy Prim e Minister), the labour m arket (Departm ent for Work and Pensions), health (Departm ent of Health), education (Depar tm ent for Education and Skills) and crim inal justice and citizenship (Hom e Office). Statistics on ethnicity are collected separately for Scotland, where this is the responsibility of the Scottish Executive, and for Northern Ireland, where they are published by the governm ental Northern Ireland Statistics and Research Agency. Unlike in m ost EU countries, citizenship and nationality data are not recorded in UK Censuses. The UK does furtherm ore not have a population register. The approach of using nationalit y as the principal criterion has not been taken in UK, since firstly, from a UK- policy perspective m igrants do not cease to be m inorities once they have taken British citizenship, and secondly, the nationality laws associated with Britain’s form er colonies are deem ed too com plex for nationality to be a useful variable on its own. The Census question on ethnic group records each person's perceived ethnic group and cultural background. The wording of the 20 0 1 England and Wales Census question 2 2 on ethnic group was the following: Eth n ic Gro u p . ‘What is your ethnic group? Chose ONE section from A to E, then tick the appropriate box to indicate your cultural background. A W h ite. Tick box options of: British; Irish or Any other White background (please write in). B Mixe d . Tick box options of: White and Black Caribbean; White and Black African; White and Asian or any other Mixed background (please write in). C As ian o r As ian Britis h . Tick box options of: Indian; Pakistani; Bangladeshi; Any other Asian background (please write in). D Black o r Black Britis h . Tick box options of: Caribbean; African; Any other Black background (please write in). E Ch in e s e o r o th e r e th n ic gro u p . Tick box options of: Chinese; Any other (please write in). 20 There is also some information available on the place of birth of the parents. Though not collected by the census, data has been collected via the British Register (for England and Wales since 1970 ) and by the General Household Survey. 2 1 A question on religion was present in the 20 0 1 Census. 2 2 This question was sim ilar to the one asked in 1991, but with changes in som e categories; for exam ple, people could tick ‘m ixed’ for the first tim e. Further inform ation on the 20 0 1 England and Wales Census questions on ethnicity and religion is available online at: http:/ / www.statistics.gov.uk/ census20 0 1/ profiles/ com m entaries/ ethnicity.asp. The ethnic classification question is in other words a self- classification system but one in which one has to choose among pre- established categories. Of course, these pre- established categories can provoke criticism . Tellin gly, the ethnic classification question for Scotland was under review by the Scottish Executive in 20 0 5, since there was concern about the categories that were used in the 1991 and 20 0 1 Censuses to classify the population along ethnic lines, and m inisters sought to establish categories that were acceptable both to data providers and data users. Questions were piloted in early 20 0 6, but the outcom e of the review is not yet publicly available at the end of 20 0 6. National Statistics of the UK states that collecting data on ethnicity is difficult because of the subject ive, m ulti- faceted and changing nature of ethnic identification and the lack of consensus on what is an ‘ethnic group’: “Mem bership of any ethnic group is som ething that is subjectively m eaningful to the person concerned and the term inology used to describe ethnic groups has changed over tim e. As a result, ethnic groups, however defined and m easured, will tend to evolve depending upon social and political attitudes or developm ents. Therefore we do not believe that basing ethnic identification upon an objective and rigid classification of ethnic groups is practicable” (National Statistics, 20 0 3: 7). This m ight be true, but the self identification in the UK system does im ply a choice between pre-established categories. These pre- established categories are piloted and pre- tested, but it does in the end boil down to a forced choice and the offered categories do still largely reflect the dom inant discourse (and legislative fram ework) of the state. Furtherm ore, there is the particular difficulty of self identification for people with m ixed ascendance (although the ‘m ixed’ option, which was introduced in the 20 0 1 Census, does provide som e kind of a solution to this problem ). Co n clu s io n a n d d e ba te Data on im m igrants and ethnic m inorities of different European countries are today hardly com parable. A num ber of countries can produce very detailed distinctions with regard to the foreign origin and com position of its population, while other countries feel the production of such data is inappropriate and dangerous. As a result, we have data on apples and pears and proper com parative social scientific work is being frustrated. I f we want to do serious (quantitative) com parative work with regard to foreign origin groups across Europe, we need com parable operationalisation system s which go beyond the sim ple distinction between nationals and non- nationals. The latter system is biased given the important variation in nationality legislations across Europe. The Regulation on harm onised statistics proposed by the European Commission is a step forward but does not resolve the issue of identifying and quantifying second generation im m igrants and longer established ethnic m inority groups. In the Netherlands and Nordic countries a formalized criterion (b irth place of parents) has been introduced in the 1990 s to pinpoint ethnic m inority and foreign origin groups. It has proven to be a useful instrum ent in docum enting discrim inatory practices and social exclusion of ethnic groups. At the sam e tim e, however, the differentiation between western allochthones/ im m igrants and non- western allochthones/ im m igrants, has added to the process of racialisation of society. Even worse, in public discourse these statistical notions som etim es function as (dis)qualifying social categories. As a result, the categories have incited quite som e resistance among those being classified against their will. One way out would be to allow people to classify themselves as is done in the UK. From a m ethodological point of view there are equally problem s. Country of birth of parents (or grandparents) can only function as a proxy for im m igrant background and ethnicity for a lim ited tim e span, especially because people with mixed origins are difficult to classify in a coherent and sensible manner. Self identification sh ifts the burden of this problem to the people we want to classify. That does not entirely resolve a num ber of fundam ental challenges. People m ight legitim ately want to classify them selves as part of the dom inant ethnic group or as part of no ethnic group at all but still be faced with discrim ination (or ethnic disadvantage) if they are judged to be part of a visible m inority or negatively racialised group. In the end every system of ethnic categor isation holds the risk of essentialism : it reifies ethnic groups. Furtherm ore, ethnic categorisations reflect (dom inant ) opinions about who is ‘in’ and who is ‘out’, which are em bedded in a specific tim e and place. Scientific classifications, and their statistical form alisation, - even if inform ed by self classification - are not im m une to this. They are equally subordinate to the societal context and power relations as other social products. We agree with Bourdieu when he writes: “every science which pretends to propose criteria which are in the best way anchored in reality should not forget that it does not do anything else than registering a particular state of the struggle of classification, that is to say, a particular state of m aterial and sym bolic relations of power between those who have an interest in this or that particular way of classifying and who, just as itself, call upon scientific authority to establish in reality and in reason an arbitrary division which it hopes to impose” (Bourdieu, 1980 : 66). The double hermeneutics which are inherent to social scientific activity does not allow us to im agine the constitution of scientific categories which are truly autonomous. Products of a social and political context, they are not immutable. They can be redefined when the context changes or they can loose the ir relevance when they have been instrum entally used – for instance when being used m ore as m eans of declassification than as m eans of classification. Categories which want to distinguish social groups and individuals should thus be treated with prudence and large reservations. Nevertheless, one should equally be able to name problems in order to resolve them and to identify particular groups in order to be able and study them . Sim on has nicely form ulated this dilem m a with which researchers and policy m akers are confronted: “(…) is it preferable to defend the invisibilisation of ethnic differences in the observational apparatus, while at the sam e tim e risking to allow hidden discrim inatory practices to prosper, or should one construct categories which, by their sim ple existence, can potentially reinforce a stigm atising designation of particular populations?” (Sim on, 1997: 9). Social scientists (and policy m akers) need analytical categories that allow to coun t and classify people according to their foreig n origin or ethnic background in order to be able to examine their integration into m ainstream society. We need reliable data to be able to m easure racial discrim ination or processes of social exclusion of which visible m inorities are victim . We should be conscious (and rem ain vigilant) with regard to the perform ative effects of ethnic categorisations, especially in their statistical form. As Keith puts it: “empirical academ ic studies potentially reify m inority presence through ascribed ethnicities that are m onitored, counted, and m easured in term s of dem ographic penetration of political system s, em ploym ent profiles, and attem pts to prom ote equal opportunities. Such m easurem ent m ay be pragm atically progressive and politically defensible but inevitably it highlights the ‘border problem s’ of definitions of dem ographic fixity that reveal the absurdity of racial languages enshrined in politics of affirm ative action and census m onitoring” (Keith, 20 0 5: 258- 259). The classification of ethnic groups in our view, however, constitutes a necessary evil in the construction of an efficient policy aiming at equal opportunities and in the struggle against racism . Furtherm ore, if we want to prom ote the quality of international com parative work on the issue, it is essential that classification system s of foreign origin and ethnic background are as sim ilar as possible. For the tim e being this is not (often) the case. Researchers should bare this in m ind. We cannot propose a ‘perfect’ system for classification here. We do , h owever , think that com parative research m ight profit from the availability of reliable data on country of birth of parents of the population across Europe to be able to investigate recent im m igrant groups of first and second generation. In the long term , an d for those countries which have already long established ethnic m inority groups, such a form alised classification system should in our opinion be com bined with a self identification procedure. It is not a m atter of one or the other system being better. For (com parative) research on immigrant origin groups and ethnic minority groups both strategies have their advantages and disadvantages. External classification (with proxies as country of birth of parents) is better suited for statistical com parative wor k on people linked to recent im m igration waves. Self identification allows to (somewhat ) rem ediate im position effects and is better equipped to deal with ‘historic’ ethnic m inority groups, but is m ore difficult to organize and m ore difficult to compare acr oss countries. Biblio grap h ie AMARO H. et ZAMBRANA R. (20 0 0 ), “Criollo, Mestizo, Mulato, LatiNegro, Indigena, White or Black? The US Hispanic/ Latino Population and Multiple Responses in the 20 0 0 Census”, Am erican Journal of Public Health, 90 (11): 17241727. ASPINALL P. (20 0 2), “Collective Term inology to Describe the Minority Ethnic Population: The Persistence of Confusion and Am biguity in Usage”, Sociology , 36 (4): 80 3- 816. BALIBAR E. (1992) Les Frontières de la Dém ocratie. Paris : La Découverte. BLUM, A. (20 0 2) ‘Resistance to Identity Categorization in France’ in KERTZER, D. & AREL, D. (eds). Census and Ethnicity : the Politics of Race, Ethnicity and Language in National Censuses, Cam bridge: Cam bridge University Press, pp. 121-147. BULMER, M. (1996) “The ethnic group question in the 1991 Census of Population” in COLEMAN, D. & SALT, J . (eds) Ethnicity in the 1991 Census, Volum e One: Dem ographic Characteristics of the Ethnic, Minority Populations . London: HMSO, pp. 33- 62. CANTISANI, G. & GRECO, V. (20 0 6a) “Registration of Acquisition of Citizenship”, pp. 167-178 in POULAIN, M. , PERRIN, N. & SINGLETON, A. (eds.) THESIM. Tow ards Harm onised European Statistics on International Migration. Louvain : Presses Universitaires de Louvain. CANTISANI, G. & GRECO, V. (20 0 6b) “Statistics on Acquisition of Citizenship”, pp. 261- 270 in POULAIN, M. , PERRIN, N. & SINGLETON, A. (eds.) THESIM. Tow ards Harm onised European Statistics on International Migration. Louvain: Presses Universitaires de Louvain. CANTISANI, G. & POULAIN, M. (20 0 6) “Statistics on Population with Usual Residence”, pp.181- 20 1 in POULAIN, M. , PERRIN, N. & SINGLETON, A. (eds.) THESIM. Tow ards Harm onised European Statistics on International Migration. Louvain: Presses Universitaires de Louvain. DUMONT, J .- C. & LEMAITRE, G. (20 0 5) Counting im m igrants and expatriates in OECD countries: a new perspective. OECD Social, Em ploym ent and Migration working papers No. 25. INSTITUT NATIONAL DE STATISTIQUE (1986) Loi du 4 Juillet 1962 relativ e à la statistique publique, m odifiée par la loi du 1er Août 198 5, Bruxelles, Ministère des affaires économ iques, 25 p. J ACOBS, D. & REA A. (20 0 5) "Construction et im portation des classem ents ethniques. Allochtones et im m igrés aux Pays- Bas et en Belgique", Rev ue Européenne des Migrations Int ernationales, 21 (2): 35- 59. J ENKINS R. (1994) ‘Rethinking ethnicity: identity, categorization and power’, Ethnic and Racial Studies, 17 (2), pp. 197- 223. KEITH, M. (20 0 5) ‘Racialization and the Public Spaces of the Multicultural City’, pp. 249- 270 in Karim Murji and J ohn Solom os (eds) Racialization. Studies in theory and practice. Oxford: Oxford University Press, 20 0 5. KERTZER, D. & AREL, D. (20 0 2) ‘Censuses, identity form ation and the struggle for political power’, in KERTZER, D. & AREL, D. (eds). Census and Ethnicity : the Politics of Race, Ethnicity and Language in National Censuses, Cam bridge: Cam bridge University Press, pp. 1- 42. KRIEGER N. (20 0 0 ) ‘Counting accountably: Im plications of the New Approaches to Classifying Race/ Ethnicity in the 20 0 0 Census’, Am erican Journal of Public Health, 90 (11), pp. 1687-1689. LEE S. (1993) ‘Racial classifications in the US census: 1890-1990 ’, Ethnic and Racial Studies, 16 (1), pp. 75- 94. LIE E. (20 0 2) ‘Num bering the nationalities: ethnic m inorities in Norwegian popula tion censuses 1845-1930 ’, Ethnic and Racial Studies, 25 (5), pp. 80 2- 822. MINDERHEDENNOTA (1983) Tw eede Kam er, zitting 1982 -198 3, nr. 1610 2 (21). Den Haag. NATIONAL STATISTICS (20 0 3) Ethnic group statistics. A guide for the collection and classification of ethnic data. London: National Statistics. NOBLES M. (20 0 0 ) ‘History Counts: A Comparative Analysis of Racial/ Color Categorization in US and Brazilian Censuses’, Am erican Journal of Public Health, 9 0 (11), pp. 1738-1745. PERRIN N. (20 0 6) “A Cohort Approach to Acquisition of Citizenship Statistics”, pp.321- 366 in POULAIN, M. , PERRIN, N. & SINGLETON, A. (eds.) THESIM. Tow ards Harm onised European Statistics on International Migration. Louvain : Presses Universitaires de Louvain. RICH E M. (1999) ‘Cultural an d Political Dim ensions of the US Census. Past an d present’, Am erican Behavioural Scientist, 42 (6), pp. 933- 945. SIMON P. (1997) ‘La représentation statistique de l’im m igration. Peut - on com ptabiliser l’ethnicité ?’ in RALLU J ., COURBAGE, Y. & PICHE V. (eds.) Old and new m inorities/ Anciennes et nouvelles m inorités, pp. 1-30 , Paris, INED. SIMON P. (1998 ) ‘Nationalité et origine dans la statistique française. Les catégories ambiguës’, Population, 3, pp. 541-568. SONDIK E., LUCAS J., MADANS J. & SMITH S. (20 0 0 ) ‘Race/ Ethnicity and the 20 0 0 Census : Implications for Public Health’, Am erican Journal of Public Health, 90 (11), pp. 170 9-1713. SPIRE, A. & MERLLIE D. (1998) ‘La question des origines dans les statistiques en France : Les enjeux d’une controverse’, Le m ouvem ent social, n°18 8 , pp. 119-129. VASENDEN, K. (20 0 5) Statistical definitions of persons w ith im m igrant background - new developm ents and international com parison. Paper written 26 April 20 0 5 for the Nordic Dem ographic Sym posium , Aalborg, Denm ark. WATERS M. (20 0 0 ) ‘Im m igration, Interm arriage, and the Challenges of Measuring Racial/ Ethnic Identities’, Am erican Journal of Public Health, 90 (11), pp. 1735-1737. WILLIAMS D. & J ACKSON J . (20 0 0 ) ‘Race/ Ethnicity and the 20 0 0 Census: Recom m endations for African Am ericans and Other Black Populations in the United States’, Am erican Journal of Public Health, 90 (11), pp. 1728-1730 .