The German statistical category “migration background”: Historical roots, revisions and shortcomings

The term “migration background” is commonly used in Germany today, but this neologism is only 20 years old. As an official category, it is even much younger. There has been only little research concerning the new population category, which emerged around the turn of the millennium. Thus, the question how the “migration background” could become the central category describing migration related diversity in Germany is not answered yet. This article fills this gap by exploring the context of the emergence of the “migration background” including the history of ethnic categories in German official statistics. It describes the actual definition of a “migration background” which became an official category in 2007 when the German Federal Statistical Office started publishing data regarding “the population with a migration background” based on the microcensus, a 1% household survey with mandatory participation. The central questions are: how national membership is imagined, how is it inscribed in definitions, and what adaptions had to be made over time? To answer these questions, different sources as questionnaires, publications of results of the microcensus and national reports on children and youth are analysed. Using interpretative methods, it is shown how a new taxonomy of the population in Germany was created, how it was influenced by international and national educational research, and to which extent it reshaped the perspectives on newcomers and natives. It is shown that the new category is tightly bound to citizenship and summarizes a number of older ethnic categories, but excludes also immigrated Germans who immigrated shortly after Second World War and from the former German Democratic Republic. Therefore, the label “migration background” is misleading because inherited citizenship and ancestry is in the centre of the definition rather than migration experience.

to citizenship and summarizes a number of older ethnic categories, but excludes also immigrated Germans who immigrated shortly after Second World War and from the former German Democratic Republic. Therefore, the label "migration background" is misleading because inherited citizenship and ancestry is in the centre of the definition rather than migration experience.

Keywords
Migration background, ethnic categories, national statistic, German microcensus, statistical category, citizenship, national membership In 2007, the German Federal Statistical Office started a new publication series entitled "Population with a migration background -results from the microcensus 2005". To this purpose, the office developed a definition for people with a migration background. Around the millennium social, educational and health sciences had indeed employed this neologism, but very seldom and without any consistent definition. In the following pages, I will first recount the historical context of the official statistical category "migration background" and its antecedents; secondly, analyse how this category was conceptualized by the Federal Statistical Office and end with discussing some shortcomings and problematic implications of this concept for recent membership in the German nation.
Why did the concept "migration background" emerge around the millennium?
Since the end of Second World War, the common official population categories in Germany 1 in "natio-ethno-cultural" terms (Mecheril, 2003: 23ff.) have been citizenship and religion. 2 Henceforth, the German state has not collected explicit data on ethnicity and race, but rather related data on foreigners, migrants and their descendants, most consistently as part of the microcensus. This is Germany's largest representative survey, conducted annually since 1957 by the Federal Statistical Office, working in cooperation with the statistical offices of Germany's federal states (Bundesl€ ander). Data are gathered on population characteristics and living conditions, including questions about citizenship.
The categories in census questionnaires, such as this microcensus, can be seen as artefacts, which reveal information about membership within a society at a given time (Yanow, 2003: 22). A first observation concerning the German microcensus is that categories specifying different foreign citizenships were not considered in the first six years of the survey, but from 1964 onwards in increasing detail (see Figure 1). 3 While the number of categories (and questions) regarding foreigners increased over time, it is important to note that until 1982 statisticians were much more concerned about tracking different groups of Germans, specifically people of Figure 1. Categories regarding residents' citizenship, 1957-2004. Source: Statistisches Bundesamt (1988; Gesis (2017); own representation.
German descent who fled the GDR or left socialist countries such as Poland, Romania, Czechoslovakia, Hungary or Yugoslavia (see Figure 2). Ethnic German immigrants from the Eastern bloc were differentiated according to the region they came from and the kind of federal identity card they had received as "expelled persons" (Bundesvertriebenenausweis). If they had lived before 31 December 1937 within the borders of the then internationally recognized German eastern territories, they received a Bundesvertriebenenausweis A. If they settled during Second World War in these or in German-occupied territories, they received a Bundesvertriebenenausweis B. Refugees from the GDR received a Bundesvertriebenenausweis C.
In short, the growing lists of detailed information on citizenships illustrate that foreigners with different citizenships were seen as an increasingly important category in the course of the 1980s, while questions regarding the ID-card of expelled Germans were asked in 1982 for the last time. After the reunification in 1989, the category of refugees from the Soviet-occupied zone or Soviet sector of Berlin became obsolete, and in 1993 the law on expelled persons changed. From then on, only persons from the former USSR with German ancestry, being discriminated for their Germanness, could enter Germany as "late resettlers" (Sp€ ataussiedler). The categories of "expelled persons" and GDR-refugees lost their importance shortly after reunification.  Indeed, Sp€ ataussiedler became an important group during the 1990s, as their numbers increased. However, how many Sp€ ataussiedler immigrated to Germany can only be estimated. Since they were considered Germans, their entry was not documented separately in immigration statistics and they were not asked any questions about their immigration or place of birth in the microcensus until 2005. Therefore, Sp€ ataussiedler were invisible as a population category. Nonetheless existed concerns about this group, and problems of integration, especially of youth, were reported (for example, Bayer, 1998;Claus, 1994;Steffen and Elsner, 2000).
At the same time, two other resident groups in Germany were growing: naturalized persons and the so-called "second generation" migrants, i.e. children of immigrants, who remained foreigners due to the ius sanguinis policy of the German citizenship law, first instated 1913. According to the ius sanguinis principle, German citizenship was granted to those whose parents were German citizens; children of foreigners inherited the foreign citizenship(s) of their parents. In 2000, German citizenship law reforms in part rescinded this ius sanguinis requirement. Since then, children who have at least one foreign parent residing legally for a minimum of eight years in Germany may receive German citizenship.
As a consequence, the category of "Germans" became more diverse around the turn of the millennium -due to the immigration of Sp€ ataussiedler, naturalizations and German-born children with foreign parents. By this point it had become obvious that citizenship did not really differentiate between natives born in Germany and newcomers. The category of "Germans" actually included immigrants with the status of "expelled persons" and "refugees" in former rounds of the microcensus as well as naturalized and Sp€ ataussiedler. At the same time, the category "foreigners" actually included many natives, that is, people who had been born in Germany as foreigners, even if they had never left the country. With the annually growing groups of "German immigrants" due to immigration and "native foreigners" due to children born to foreigners in Germany, political advisors and academic experts increasingly demanded that new categories be developed to include them adequately in population statistics (Salentin and Wilkening, 2003;Unabh€ angige Kommission "Zuwanderung", 2001).
The term "migration background" was used as early as 1998, as part of an expert discussion on youth welfare documented in a regularly issued governmental report, but still unsystematically and without further definition (Deutscher Bundestag, 1998). In the reports over the following two decades, the term was applied with increasing frequency (see Figure 3). The usage of "migration background" did not wane until 2017, when it was overtaken by the term "asylum", due to the arrival of about a million people seeking refuge in 2015, directing public interest to this specific group of newcomers and away from those who had already been in Germany a longer time.
Starting in 2005, when "questions on migration" were introduced in the microcensus, and until 2013, the term "migration background" was used much more frequently in these governmental reports. The reports drew heavily from the microcensus data and thus refer frequently to the new category "migration background", which had been operationalized for the first time in 2005 using these "questions on migration" (see Figure 4). In the following years, other surveys also included more

Frequencies of Terms
The * indicates that also compounds are included which are frequent in German, for example asylum seeker (Asylbewerber/-in). Figure 3. Frequencies of the terms foreign/er, migration background, resettler, refugee and asylum in the governmental reports on children and youth from 1965 to 2017. Source: DJI (2017); Bundesministerium für Familie, Senioren, Frauen und Jugend (2017); own analysis and representation. In which year did you immigrate as a Spätaussiedler(in) or were you naturalised?

Changed 2009:
In which year were you naturalised?
7. Which citizenship did you possess before naturalisation? (according to coding list)

Changed 2007:
Which citizenship did you possess before immigrating as a Spätaussiedler(in) or before you were naturalised? (according to coding list) The following questions were asked every 4 years (

12.
In which year was your mother naturalised?
13. Which citizenship did your mother possess before naturalisation? (according to coding list) Questions 8 to 13 are also asked of the father. or less similar operationalizations of the concept "migration background". These observations thus mark the "surface of emergence" of the concept "migration background", as described by Foucault (2002: 45). The "questions on migration" made people who were German resettlers, naturalized and the second generation migrants visible in German population statistics. Strikingly, the justification of the microcensus law introducing these questions only explains their introduction with an interest in naturalized citizens and second generation migrants: The immigration and integration of migrants is an increasingly important political issue, for which there is still a lack of basic information. Information on naturalization is significant because through naturalization a formal integration takes place which allows one to make conclusions about these migrants' willingness to integrate. (Deutscher Bundestag, 2004: 11) Questions about respondents' parents were included every four years: "The question about the citizenship of one's parents and possibly their naturalization allows a differentiated analysis of the whole spectrum of features of the microcensus for the group of persons belonging to the second migrant generation" (Deutscher Bundestag, 2004: 13). Resettlers and third generation immigrants were not explicitly mentioned at all. However, these two groups were also statistically included in the group defined as "people with a migration background". Despite its use as a statistical category, the term "migration background" was then not mentioned in the microcensus law nor in the justification of the parliamentary proposal of it.
But why was "migration background" chosen and not another category? To understand this, one has to look at the context: in 2000 the OECD had started the Programme of International Student Assessment (PISA), which analysed the achievement of different categories of students. In 2003, the English version of the PISA report mentioned the term "migrant background" for the first time. Without an actual definition, it was used to describe the opposite of "native": "When interpreting performance gaps between native students and those with a migrant background [. . .]" (OECD, 2004a: 169). Generally, the term "migrant background" was not used as often as the terms "immigrant background" and "immigrant students", which would become the standard denomination in PISA for the overall category of students who have two foreign parents, separable into first and second generation by their own birth country. The German translation of the 2003 PISA report uses the German term "Migrationshintergrund" 22 times (OECD, 2004b) and over the following years this became an established part of the PISA publications in German (Will, 2018a).
Two years after this first German translation, an important point in the discursive development of this term is marked by the publication of two national reports: In 2006, the first German national educational report was released. It used data generated from the microcensus' questions on migration to categorize particular population groups in Germany as "persons with a migration background", as "foreigners" of the first, second or third generation, "Germans" of the first or second generation and "persons without a migration background" (Konsortium Bildungsberichterstattung, 2006: 140). In turn, this taxonomy, originally developed for the monitoring of the population's educational achievements, constituted the blueprint for the Federal Statistical Office's new publication series focussing on migration, this being a part of the subseries titled "foreign population". In the strict sense, it is inconsistent to statistically document "persons with a migration background" as part of the subseries titled "foreigners", because more than half of the "persons with a migration background" are German citizens. This anomaly is increased by the fact that the contrasting reference group in the Federal Statistical Offices' subseries is "Germans without a migration background", changed from "persons without a migration background" in the educational report. Set in opposition to the subseries' focus "persons with a migration background", this change implies the equation "Germans ¼ those without a migration background" while those with a "migration background" are not presented as German. Even if unintended, this is at the very least an implicit denial that many Germans have ancestors who had immigrated to Germany at one point in time, which is a constitutive part of German history. At the same time, "persons with a migration background" are framed in the statistical report as a problematic part of the population, which has to be monitored until the third generation because " [a] ccording to scientific studies from all classic immigration countries, members of the 3rd generation are extraordinary 'difficult' in terms of integration politics" (Statistisches Bundesamt, 2009Bundesamt, (2007.
The result is a taxonomy with 16 immigrant population subgroups, presented in contrast to a majority referred to as "Germans without a migration background". Strikingly, the migration experience of "ethnic Germans", which had been recorded until 1982 with the same accuracy as that of (former) foreigners, is partly excluded from this new taxonomy, because "only immigration into the actual territory of the Federal Republic from 1950 onwards is taken into account" (Statistisches Bundesamt, 2009Bundesamt, (2007. These changes and omissions raise the question: How exactly is the category "migration background" defined? The category "migration background" in Germany -The official definition(s) In 2005, the Federal Statistical Office defined this category for the first time: Persons with a migration background are defined as 'all migrants who entered the current territory of the Federal Republic of Germany after 1949, and all foreigners born in Germany and all those born in Germany as Germans with at least one parent who immigrated to Germany or who was born as a foreigner in Germany'. (Statistisches Bundesamt, 2009 (2007): 6) The wording was changed and simplified in 2016. The current definition is: "A person has a migration background if s/he or at least one of his/her parents did not acquire the German citizenship at birth" (Statistisches Bundesamt, 2017: 4).
Following Yanow (2003), I try to trace the assumptions and inscriptions in this powerful category. The official definition of "migration background" was (and is) very complex. Possibly that is the reason why the classification is not commonly debated. In science and technology studies, one describes similar processes as a "black-box"-effect, for instance, complicated algorithms are accepted and not questioned once they have become established (Porter, 1996: 45). The same is true for the category "migration background": for a long time there has been a lack of scientific and public critique of it, 4 which is surprising given the controversial nature of the topic and of decisions made in this policy field.
The categories' creators had expected some discussion about their newly designed taxonomy of immigrant groups (Sch€ afer andBru¨ckner, 2009: 1047), but perhaps did not foresee that the category "migration background" would have such a significant impact on wider discourse. However, there is no such thing as naı¨ve science, especially not in state offices. Statistical offices are usually not only aware that statistics will have an effect on discourse and policy, but even gather them to this purpose. Official categories and figures inform and structure public and scientific discourse and lead to the passing and implementation of regulations, laws and measures. In this way, the official representations and definitions of certain phenomena become factual and impact our reality.
Figures function additionally as evidence for a variety of argumentations and can be described as a "generalized communication medium", to borrow and adapt Luhmanns concept (Heintz, 2007: 65). Everybody understands figures and can communicate about them, even if in detail the categories, upon which the numbers are based, are actually diffuse. By providing such easy-to-use figures, also the Federal Statistical Offices' official categories shape the discourse and define what can be said and how (Foucault, 2002). Scientific categories are powerful actors that structure our world once they enter a discourse (Berger andLuckmann, 1990 (1969) ;Porter, 1996;Bowker and Star, 1999;Latour, 2001). Even though these categories are human constructs and do not exist as part of nature, they are nonetheless bound to social interactions and become real through daily enactments (Yanow, 2003: 22, 44).
Similar to the Dutch categories of authochton and allochthon (and its subcategories), the category "migration background" has been created by statisticians. Respondents were not asked if they have a "migration background" or not. In the Netherlands, the data stem from the population registry, but the categories are created and people are classified within them, without asking the classified if they belong to these categories. In Germany participants of the microcensus initially had to answer 3 to 17 questions, depending on the skip rules (see Figure 4 for the questions). In 2009 and with a new microcensus law in 2017, the number of questions increased to 19 questions and then 24, respectively (see Figure 1 in the online Supplemental material).
The answers to these questions are processed to come up with a classification of the respondents, using additional data of parents living in the same household. The Federal Statistical Office ascribes a "migration background" to the surveyed persons employing a complex program for their classification. The complexity resulted in some minor "mistakes", and revisions have had to be made of the statistical publications based on this program. For example, German children of foreign parents (so-called ius soli children who were granted German citizenship according to the reissued law of 2000) were "miscategorized" as foreigners despite their German citizenship. Foreign-born Germans were re-categorized -recently, in 2016/17. All former volumes of the subseries have since been revised (see Will, 2018b).
It is important to note that there was limited data on residents' country of origin until 2017. 5 Until then, the microcensus only asked if somebody was born on the current territory of Germany with the answer options "yes" and "no". Thus, no data on concrete countries of origin were available, only "Germany within its current borders" and "abroad". Participants from "abroad" have also been asked when they entered Germany and which citizenship(s) they currently possess and possibly possessed earlier. All participants born in the current territory of Germany are asked about their (former) citizenship(s) and the basis for having German citizenship. Therefore, "origin" is operationalized in microcensus-data in terms of "citizenship". A second important detail: there are no tick boxes in the microcensus asking for any self-categorization. Only (apparently) objective facts are asked about territories, citizenships, if applicable naturalization year, resettler status and year of entry, and all this is also asked about both parents if they do not live in the same household (between 2005 and 2016 every four years, since 2017 annually).
The Federal Statistical Office provides a four-quadrant classification matrix, organizing people according to their migration status using the criteria "birth inside/outside current German borders" and "German citizenship/non-German citizenship" (see Figure 5). The authors describe the categorizations in the first, second and third quadrant as rather unproblematic as far as discriminatory issues are concerned (Statistisches Bundesamt, 2009Bundesamt, (2007: 325f.). In the fourth quadrant, persons with and without a "migration background" are grouped together and should be separated. In order to do this, additional data about the respondents are needed concerning their country of birth and citizenship at birth; this kind of boundary-drawing is portrayed as controversial by the authors (Statistisches Bundesamt, 2009Bundesamt, (2007).
However, the third quadrant is problematic too, since Germans with and without a "migration background" are also combined in the same quadrant and not all immigrated Germans are counted as having a "migration background"; the cut-off year is 1949. Immigration of Germans until that date does not count as a criterion for ascribing a "migration background" to someone. Taking into account how much energy was invested in endless tables of German expelled persons and refugees until a little more than 30 years ago, this decision is striking. A footnote at the cited four-quadrant matrix explains: All interviewed experts deemed it consensus that it is inappropriate to confuse the definition of migration background by including those who came to the current Federal Republic of Germany as part of the migration flows during the Third Reich or in the years immediately following WWII as a consequence of resettlements, flight or expulsion. Therefore only those immigrants who immigrated after 1949 will be categorized as having a migration background. (Statistisches Bundesamt, 2009Bundesamt, (2007 By the time you read this article, the cited text will have been revised in the mentioned recent revision of the whole series, including the results from 2005 until 2016. 6 In the revised version the "Annex 1 -persons with a migration background" is no longer included, and the cut-off year 1949 is not mentioned (Statistisches Bundesamt, 2017: 4ff., also in the revised series back to 2005), even if the latter is still being used (Statistisches Bundesamt,6.3.17). While mentioning the cut-off   Figure 6. *"Both-sided migration background" means that both parents have a migration background. "One-sided migration background" means that only one parent has a migration background. Source: Statistisches Bundesamt (2009Bundesamt ( (2007.

Country of birth
year, at least let the reader know that the migration experience of many Germans has been "silenced" (Bowker and Star, 1999: 5), not even mentioning it deepens the silence about this part of German migration history.
The As mentioned, this is part of a superordinate subseries analysing the foreign population in Germany, titled "Subseries 2 -Foreign Population" ("Reihe 2 -Ausl€ andische Bev€ olkerung"), which contains annually published results from the central registry of foreigners (Ausl€ anderzentralregister). This illustrates the logic behind the new subseries: even though it contains information on the whole population in Germany and not only foreigners, the focus is clearly directed toward "the others".
The Federal Statistical Office used the taxonomy for the first national educational report as blueprint for the new categories in the subseries. Not only does a comparison of the two taxonomies below ( Figure 6) include similar subgroups, "Appendix 1" of the statistical subseries 2.2 explicitly states, "The herewith displayed subcategorization of the population was first used in the classification of the population by migration status in the federal educational report" (Statistisches Bundesamt 2009: 332).
Despite the similarities in these two taxonomies, there is an important difference in the order of the represented groups. In the educational report, the first group is referred to as "persons with a migration background", leading to the conclusion that "persons with a migration background" are the centre of the study's interest. In contrast, the statistical subseries 2.2 names "persons with a migration background" in its title, but the first group referred to are "Germans without a migration background". This makes the majority population the reference group -which is consistent with other publications of the Federal Statistical Office -additionally assigning it the number 1 and placing all other subgroups under number 2.
With this repositioning of the main and reference group in their taxonomical representation, the Federal Statistical Office automatically widened the visual distance between the reference majority population and the group "Germans with a one-sided migration background". The taxonomy in the educational report illustrates a growing convergence, from "persons with a migration background" with foreign citizenship, followed by those with German citizenship, to "persons without a migration background". In contrast, the statistical subseries illustrates the greatest distance between "Germans without a migration background" and children of mixed German and foreign descent (group 2.2.2.2.2.2). This observation draws attention to just one of the symbolic shortcomings of this taxonomy using "migration background". What are the shortcomings of the category "migration background"?
The category "migration background" was an innovative concept in the discourse of the early 2000s. It widened the scientific and public understanding of migrationrelated, intergenerational transmissions of inequalities, regardless of citizenship, and contributed to the recognition of German society as one structured and shaped by migration. However, it also had exclusionary effects, especially due to the way the Federal Statistical Office operationalized the official concept of "migration background" and represented the microcensus results. Strictly speaking, already the label "migration background" is misleading, since only certain migration experiences are considered in the definition and handedover to descendants, and others are not. A "migration background" is ascribed to grandchildren of people born as foreigners, but not to children of immigrated German-born persons. The concept is grounded on citizenship, not migration experience. Given that the German citizenship law has historically been based on blood-ties (until 2000 exclusively 7 ), it is still an "ethnic" rather than a migration category.
In addition, the term "migration background" is used as part of both population categories, "people with a migration background" and "people without a migration background", and thus could be used for discussions on both groups. In fact, since "people without a migration background" are characterized as the majority, it would make sense that this category would come up more often. However, the opposite is the case. That is because "persons without a migration background" normally go unmarked and are equated with the category "Germans", which is problematic since more than half of the "persons with a migration background" possess German citizenship. Thus, the latter are excluded from the category "German" within statistical representation, just as they are excluded in the everyday use of this categorization. The Federal Statistical Office reinforced this incorrect representation on its website, stating that the above-mentioned subseries displays "the situation of the population with a migration background [. . .] in relation to the German population" (Figure 2 in the online Supplemental material). It was corrected only in September 2016. It is no surprise that linguistic analysis found exclusivist patterns for the use of the term "migration background" within the media (Scarvaglieri and Zech, 2013) and parliamentary discourse (Elrick and Schwartzman, 2015).
Furthermore is the category too complex. The Federal Statistical Office issued three categorical revisions in 10 years, not counting the changed wordings of the questions. In 2009, the changes were: (1) the so-called ius soli children were recategorized into "Germans with a migration background", prior to this they were categorized as foreigners; (2) the information on other people in the household was given priority over the statements of the respondent (Statistisches Bundesamt, 2009Bundesamt, (2007: without page number) and (3) the definition of "migration background" was changed completely, focussing clearly on citizenship in 2016. The complex nature of the categorization resulted in a lack of public and scientific discourse. However, the described decisions are far-reaching and also questionable in the light of social research on integrational processes. Different categories of immigrants and their descendants are collapsed for the category "persons with a migration background", increasing their number artificially. The category "persons with a migration background" includes first, second and parts of third generation "immigrants". To join these three groups into one contradicts, the logic of migration theories which foresee an on-going adaption of immigrants from one generation to the next (Esser and Friedrichs, 1990;Portes and Rumbaut, 2001). The first generation is separable through its own migration experience in the tables, but the second and third generations are not. The Federal Statistical Office explains, It is not readily possible to separate persons born in Germany into 2nd generation (parents immigrated) and 3rd generation (grandparents immigrated), because parents can belong to different immigration generations. [. . .] This would mean that an up-todate, pragmatic differentiation can only be made between persons with their own migration experience (1st generation) and persons without their own migration experience (2nd and 3rd generation). (Statistisches Bundesamt, 2017: 6) As a result, the category of "persons with a migration background" has grouped together people who recently immigrated with others who have possibly only one immigrated grandparent and three grandparents, who are "Germans with German ancestors since birth". This echoes -probably unconsciously -the Nazi definition of Aryans: all four grandparents had to be provable of Aryan descent to count as an Aryan. This legacy, and the fact that the data are available only if children and parents live together in one household, suggests to take the third generation (from the 2.5 generation onwards) out of the definition, so that first and second generation may be represented more clearly Another important and disputable decision made by the Federal Statistical Office is that children with one German-since-birth-parent and a (former) foreign parent count as having a "migration background". Yanow shows for the U.S. that the "one-drop" rule is still active in classificatory practice, sorting people into the categories "black" or "non-White" if only one parent is not "White" (2003: 121). Why does the German-born parent not matter more than the foreign one, when a person is born and living in Germany? Austria, for example, also uses the category "migration background", but counts children of mixed descent with an Austrianborn parent as "without a migration background", justifying it with the "Recommendation of the 2020 censuses on population and houses" of the United Nations Economic Commission for Europe (Statistik Austria, 2016).
Also PISA did not include students with one native parent in its first reports, explaining that "students born in the country [of assessment, addition by author] who have one foreign-born parent (children of 'combined' families) were included in the native category, as previous research indicates that these students perform similarly to native students" (OECD, 2006: 26). These findings hold for Germany as well and were confirmed by the Federal Statistical Office's microcensus data (Statistisches Bundesamt, 2008: 15). Nonetheless, this empirical evidence was ignored in the Federal Statistical Office's interpretation of its categories, classifying children with only one German-since-birth-parent as "persons with migration background". This procedure breaks the rule of similarity, an important assumption underlying the construction of categories. The invented categorization suggests that children with one German-since-birth-parent are more similar to "persons with a migration background" than to "Germans without a migration background", contradicting the statistical data. As such, the aim of the Federal Statistical Office does not appear to be the representation of statistical data, but rather that of the different national, ethnic and cultural origins within Germany's resident population. The core concept in the decision to count children with "only" one German-since-birth-parent as part of the group including foreigners, their descendants and German immigrants after 1949, is one of (lacking) ethnic (German) purity.
The Netherlands have a similar policy as Germany, excluding persons from the category autochthon if one parent is born abroad. But in the Netherlands and Austria, citizenship is not taken into account; both countries categorize solely on the basis of the country of birth of the respondent and his or her parents. Assumed migration is in the centre of the concept. In Germany, the category "migration background" is grounded in descent, although it is framed as citizenship.
Even if the microcensus collects this information as a kind of ethnic data, these data are not meaningful and usable for equal opportunities measures. The classification takes only bureaucratic facts such as (former) citizenship(s) and resettler status into account, not lived experience or self-identification of the respondents. For the collection of data for equal opportunities policies, seven principles are crucial: 1. Self-identification 2. Voluntary participation/information 3. Information about the aim of the survey 4. Protection of personal data 5. Minority groups have to be involved in the development of the questions, categories, data analysis and interpretation 6. Possibility to indicate multiple identities, different reasons for discrimination and perceptions of others 7. Principle of doing no harm (data must not be misused) (Ahyoud et al., 2018: 33).
The microcensus meets the first criteria of data protection, but none of the other requirements. There is no place for self-identification, "migration background" is a state-administered ethnic category, whereas ethnic German origin inherited from two German-born parents is the reference category. The broader group combinations in the four-quadrant matrix in Figure 5 do not account for important differences within each group, even if other tables complement this with a list of predominantly geographic subgroups (Europe, EU28, other Europe, Africa, America, North America, Middle and South America, Asia, Australia and Oceania) and selected national groups by (former) citizenship. A similarly "lumpy" categorization is described by Yanow for the U.S. census practice (Yanow, 2003: 146f.), which she criticizes as an insufficient approach to gathering equal opportunities statistics, since underprivileged groups will be "overseen" within the broader group categories. The differences are greater than assumed, with the result that better off-groups in such a "lumpy" category can profit from equal opportunities measures more than those who are in greater need of it.
This means the category "migration background" is confusing and does not adequately portray diversity or diversity-related disparities in German society. It is time for a revision, maybe even a replacement of the category with one based on self-identification, accompanied by a question regarding one's discrimination experiences. To do this properly, organizations of "new Germans" and minority groups should be consulted in order to develop options for voluntary self-identifications.

Conclusion
Analysing German natio-ethno-cultural categories in the microcensus reveals a shift in attention, steering the focus away from German immigrants and toward foreigners. It also traces a shift to an overarching category which includes German immigrants after 1949 and onwards, foreigners and their descendants. The latter shift took place around the turn of the millennium and became official with the 2005 microcensus results concerning the "population with a migration background". The microcensus definition silences the war-related immigration of Germans, which had been of concern until the 1980s. Furthermore, the concept of "migration background" is exclusionary with regard to 15% of all Germans and half of all "persons with a migration background". The Federal Statistical Office represents them as not being part of the general population in Germany -even if they are Germans. Thus excluded, they are othered through their separation from the German natio-ethno-cultural mainstream. Therefore, the formerly innovative category "migration background" evokes questions about national membership, especially for persons who may feel German, but whose belonging to Germany is questioned by the official category. Finally, the concept of "migration background" is bound to "inherited citizenship" which values German descent higher than other descent. All of these points raise questions about the ability of the official category "migration background" to adequately represent the officially acknowledged diversity in the German population, perhaps indicating that Germany needs another concept innovation.