Germany and its minorities, or surprises of the census of 2011
On 31 May 2013, Destatis, Germany’s Federal Statistical Office, announced the first findings of the census conducted two years earlier.
Surprise: Germany was found to have 1.5 million fewer inhabitants than expected. In France and Germany alike, the discovery elicited pessimistic remarks: “alarming finding,” “demographic decline,” “calculation error.” But in fact the discrepancy has nothing to do with the low birth rate in Germany. And this kind of disparity between inter-census annual updates and census results is found in all western countries, including England and France. There was another surprise, though, of a different nature: the Muslim population’s low response rate to census questions on religion. François Héran discusses these two results, resituating them in their legal and socio-political context.
The first data from Germany’s 2011 census were divulged on 31 May of last year. They were eagerly awaited, since the preceding censuses dated back to before reunification: 1987 for the Federal Republic and 1981 for the GDR. Several difficulties explain the lag (see Box 1). In the intervening years, demographers had had to settle for other sources: municipal registers, the accuracy of which was declining as there was no systematic recording of departures or effective transmission of address changes from one Land to another, and Mikrozensus, a wide-ranging annual survey of a sample amounting to an estimated one-hundredth of the population (830,000 persons) drawn from an obsolete residential database. It was high time to update the census.
Box 1: The new German census: why the delay?
The GFR was forced to suspend the 1983 census process due to organized protests against the transfer of sensitive (in this case religion-related) information to other databases. As there was no West German equivalent of France’s data protection law, the Constitutional Court in Karlsruhe stepped into the breach, handing down an unprecedented census act (Volkszälungsurteil) stating that anyone living in a society under the rule of law had a right to determine whether statistical data about him or her could be divulged and how they could be used. The law establishing this principle, called the right to “informational self-determination” (informationnelle Selbstbestimmung), required a total revision of data collection and processing procedures. A new census designed to comply with the new legal requirements was conducted in 1987 but once again elicited strong opposition.
It was not only a matter of legally securing census operations; the Federal state also had to obtain consent from all Länder to coordinate and fund them. These undertakings took time, but Germany did manage to meet the deadline fixed by a 2008 European Union regulation requiring member states to use 2011 as their reference year.
For financial reasons, census organisers in industrial countries—Germany included—now prefer to focus their efforts on sampling and estimation techniques. After cross-checking several administrative sources, Destatis questioned all homeowners. From this base it then drew a sample made up of approximately 10% of the population, who were then enjoined to answer a “census survey.” All retirement homes and other institutions were included. A post-census survey was also conducted to verify the soundness of the results. In all, one-third of the country’s population was contacted. The published findings are an extrapolation to the population as a whole at the census reference date: 9 May 2011.
Immigration: fewer “second-generation” residents than in France
Immigration was a major focus of the first published results. Immigrants were found to account for 13.2% of the population, half of whom have been settled in Germany for over twenty years. An additional 5.7% of the population has at least one immigrant parent (not including immigrants who arrived prior to 1955, mainly repatriated Germans from territories that are now part of neighbouring Slavic countries). These two generations taken together—18.9% of the total population—form Germany’s population “of immigrant background” (mit Migrationshintergrund). The figure is slightly below the 2011 Mikrozensus estimate of 19.5%.
These figures are comparable to those for France, but the generations are distributed differently. The proportion of “second generation” residents, i.e., individuals born in the host country but to at least one immigrant parent, is lower in Germany than in France: 6% as opposed to 10%. The period during which France took in the greatest number of immigrants (nearly twice current flows) was 1950-1974, and this labour migration was followed by family reunification and marital migration, though at lower levels. Germany had already received labour migration prior to 1974, but in the 1990s it became Europe’s leading destination country (to be overtaken by Spain in the 2000s decade). In addition to Turkish migrants, Germany had begun taking in refugees from the Balkans (many of whom have since left) as well as Russian and Kazakh migrants with claims to German origin.
Immigration flows diminished somewhat in the late 2000s, but since then Germany has again become a major receiving country, with nearly one million recorded entries in 2011 and more than a million in 2012, primarily from Central Europe but with a rapid increase in Spanish and Greek immigration due to the economic crisis. However, these flows are too recent to have engendered a “second generation” as numerous as the one in France.
Among migrants and migrants’ children settled in Germany and counted in the 2011 census, the most heavily represented countries of origin Turkey (17.9%), followed by Poland (13.1%), the Russian Federation (8.7%), Kazakhstan (8.2%), Italy (5.3%), Romania (3.7%), Greece (2.4%), Austria (2.2%) and Croatia (2.2%). Taken together, the two generations form a relatively youthful group: 42.1% are under 30 years, as opposed to 27.8% in the rest of the German population; only 9% are 65 years or over, as against 23.3% in the rest of the population.
The census brings to light marked variations in the geographical distribution of migrants and migrants’ children (Figure 1). While migrant settlement areas are largely determined by degree of urbanization and industrialization, a specific factor is operative in the new Länder: the fact that they used to constitute the GDR, which was closed to immigration. Though the Schröder government officially abandoned the idea that Germany “is not an immigration country” in the late 1990s, that idea seems to have persisted in the East, in fact as well as in representations.
Figure 1 . Share of the population represented by immigrants and persons with at least on immigrant parent, 2011 German census
More housing than expected—but fewer inhabitants
Immigration also plays a role in one of the major surprises of these census results. In mid-2011, the population estimates based on local registers and annual demographic updates predicted 81.7 million inhabitants, but the census found 1.5 million fewer, or 80.2 million. The number of housing units, on the other hand, turned out to be higher than expected by half a million (40.8 million as opposed to 40.3).
A journalist at Der Spiegel wondered if the million and a half inhabitants had just “evaporated overnight.” But this kind of estimation error is hardly surprising given the length of time elapsed since the last available censuses: 25 years for West Germany, 30 years for East Germany. In England-Wales, for example, one decennial census came out below annual update estimates (-1,140,000 in 2001) and another above (+464,000 in 2011).
The same oscillation has occurred in France: the 1999 census found half a million fewer inhabitants than predicted (after nine years of estimates); the 2004 edition found 420,000 more (due to a new, more detailed data-collecting procedure)—an uncertainty that INSEE is currently working to reduce by means of a rotating system of annual census surveys.
The new Länder’s phantom migrants
In the Germany of 2013, the “disappearance” of so many inhabitants has elicited a sense that the country is once again in demographic decline. Given that the fertility rate is far below replacement (at fewer than 1.4 children per woman), the country has been recording approximately 190,000 more deaths than births, whereas in France births exceed deaths by 250,000. However, in this case the need to revise estimates of total inhabitants downwards has nothing to do with the low birth rate. Study of the census makes it clear that the problem is due first and foremost to incomplete recording of departures by foreigners, specifically departures from the eastern Länder.
In fact, the deficit varies spectacularly by nationality and Land (Figure 2). It is tiny for the population holding German nationality (-0.6%) with modest spikes in the city-states of Berlin (2.5%) and Hamburg (3.3%)—a finding in itself remarkable. However, no less than 15% of the foreign population estimated on the basis of local registers is missing. The total number of enumerated foreigners thus falls from 7.3 million to 6.2 million, meaning that foreigners account not for 8.9% of the population but 7.7%.
Figure 2. Relative disparity between the population as enumerated in May 2011 and the results of annual population updates, by Land and nationality (%)
Should these deficits be attributed to insufficient census coverage of foreign inhabitants? That hypothesis must be discarded because the deficit variations are much too strong throughout Germany to point to uneven coverage. The method used to conduct the housing inventory prior to identifying residents, including foreign ones, was the same everywhere. Foreign residents could not have escaped census-takers in such variable proportions throughout the country. The problem, then, is not with the census but seems very likely due to how the files used for the annual updates of eastern Länder populations were managed. The numbers of central European migrants (from Poland Hungary, Slovakia) in these Länder were already quite low, and many left for the more prosperous west or returned home without ever being removed from local registers. When the moment of truth of the census came, those phantom migrants were no longer there to answer the questionnaire. It will be interesting to see how German statisticians analyse these findings.
Box 2. Unprecedented data on homosexual couples with children
Whereas many analyses of 2011 British census findings focused directly on “races,” religions and languages, publications on the German census tend to handle the question of minorities more discreetly. Among noteworthy innovations is direct mention of the sex of partners in civil unions. Destatis estimates that there are around 34,000 same-sex partnerships in Germany, 60% of them male and 40% female; also that 5,700 children are being raised by same-sex couples, 83% of whom by female couples.
Religion: spectacular—and disappointing—results
Religion as it relates to immigration is a major focus of the census findings published to date. Two questions were asked. Answering the first was compulsory for fiscal reasons, as members of religious communities in Germany pay a church tax by way of state administrative services; Islam is not covered by this procedure. The second question, the only optional one on the census questionnaire, was addressed to persons who did not declare membership of any of the official religious communities referred to in the first question: Did they nonetheless identify with a religion, belief system or philosophical view? Three options were available for Islam: Sunni, Shiite and Alevi (a more liberal version of Islam present in Turkey). Organizing the religion section of the census into two tiers this way directly reflects the German understanding of relations between state, individual and religion (Box 3).
Box 3. The state and religious faiths or denominations in Germany: the legal basis for census questions on religion
Article 140 of Germany’s 1949 Basic Law states that “no person is under obligation to declare their religious convictions.” That law reiterates the principle of the separation of church and state first formulated in the Constitution of Weimar (1919), but it in no way prohibits cooperation between church and state. The major Christian denominations have the status of “religious community under civil law” (öffentlich-rechtliche Religionsgesellschaft) but they are not integrated into the state apparatus. Believers who agree to pay a church tax also agree to let their employer deduct it from their pay and to have the tax office transmit their payment to the religious denomination of their choice. When a person moves, the administration accordingly updates the list of church taxpayers it transmits to the churches. The idea behind the law is that the census offers a good opportunity to check the accuracy of the tax lists.
This legal regime only covers longstanding religious institutions in Germany and therefore does not apply to Islam. However, Muslims and followers of other religions are at liberty to set up private religious associations protected by the fundamental German right to self-determination. The optional census question on religious convictions was addressed specifically to members of those religions.
The combined results for the two questions on religion are both spectacular and disappointing (Figure 3). Once again the weight of the communist heritage in the new, eastern Länder has made itself felt. Inhabitants of this part of Germany are largely de-Christianized and have no experience of church tax; three-quarters of them either refused to answer the question or stated no affiliation with the official churches, as against a mere one-fifth in the west. Furthermore, only 5.3% of responses cited an affiliation with non-Christian religions and a mere 1.9% reported affiliation with Islam. Destatis directors openly expressed their disappointment on this point: “unfortunately, this means that the 2011 census cannot provide reliable results for these religions in Germany.”
Figure 3. Reporting of religious affiliation in the 2011 German census
A test administered before the census showed that persons with low educational attainment found it difficult to grasp the legal notion of “religious community under civil law” and the filtering function of the first question in relation to the second. Nonetheless, legal considerations prevailed, and the questionnaire was only very slightly revised.
The extreme wariness of Muslims toward the census cannot be dissociated from the general context. The census was conducted in social waters recently roiled by the August 2010 publication of Thilo Sarrazin’s polemic essay, Deutschland schafft sich ab (Germany is doing itself in). Sarrazin, a member of the SPD, head of the Berlin Senate’s finance committee and later on the board of directors of the Bundesbank, sharply accused Muslim immigrants as a whole of not wishing to integrate, inordinately burdening social welfare system finances and “downgrading [Germany’s] human capital.”
These polemical declarations caused a great stir throughout the country. According to Die Zeit, sales of the book topped 600,000 in November 2010 and by February 2011 had broken the 1.5 million mark. Regardless of whether Sarrazin revealed a latent state of mind in Germany or exacerbated that state, it is fair to say that spring 2011 was hardly a propitious moment for calmly reporting an affiliation with Islam in the census.
 Destatis, Zensus 2011: Ausgewählte Ergebnisse. Tabellenband zur Pressekonferenz, Wiesbaden, May 2013.
 EC rule no. 763/2008 of the European Parliament and the Council of the European Union of 9 July 2008 on population and housing censuses.
 D. Coleman, “The Twilight of the Census”, Population and Development Review 38 (supplement s1, 2012): 334-351.
 P. Breuil et al., “Les immigrés, les descendants d’immigrés et leurs enfants”, France, Portrait social 2011 (Paris: INSEE, 2011).
 British Office for National Statistics, “Components of population change and the mid-year population estimates”, London, February 2013.
 F. Héran and L. Toulemon, “What happens when the census population figure does not match the estimates?” Population and Societies 411 (April 2005).
 F. Meisner, ed., Droit des religions (Paris: CNRS Editions, 2010), entries “Allemagne”, “Appartenance religieuse”.
 B. Gauckler, “Die Entwicklung des Fragebogens zur Haushalte-befragung des Zensus 2011. Ausgewählte Ergebnisse des quantitativen Feldpretests”, Statistisches Bundesamt, Wirtschaft und Statistik (August 2011): 718-734.
 B. Lestrade, “À propos du débat sur le livre de Thilo Sarrazin: quelle intégration scolaire et professionnelle réelle des immigrés?”, Allemagne d’aujourd’hui 195, (January-March 2011): 18-37.
 H. Sezgin, ed., Manifest der Vielen. Deutschland erfindet sich neu [Germany reinvents itself (Berlin: Blumenbar, 2011) (replies to Thilo Sarrazin by approximately 30 authors of Muslim immigrant background).