Immigration to France has continued almost without interruption, albeit at varying levels, for at least the last 150 years. As in the past, the traces of this movement are visible in the composition of the country’s population. In 1999, France’s resident population of foreign or partially foreign origin (immigrants or persons born in France with at least one immigrant parent or grandparent) represented around 13.5 million people, equivalent to between a quarter and a fifth of the total population. This result, obtained by Michèle Tribalatusing the most recent census and, in particular, the Study of Family History (EHF) survey attached to the census, is very close to the same author’s 1986 estimate based on civil registration sources. But use of the EHF here allows the author to go further and to estimate the totals by country of origin, while excluding the descendants of repatriates from the former colonies, most notably Algeria. This refinement, which requires making some assumptions, proves valuable when the aim is to measure the discriminatory effects suffered by these populations in key areas, such as employment, which is the example the author analyses here.
France has been a country of immigration for the last 150 years. The two strongest phases of this immigration were the inter-war years and after the Second World War. In the late-nineteenth and early-twentieth century, immigrants were primarily Belgians and Italians, followed by Poles and Spaniards. France numbered nearly 500,000 Belgians in the census of 1886, while the number of Italians reached a peak (808,000) in 1931, ahead of Poles (508,000) and Spaniards (352,000). After the Second World War, migration flows from Spain and Italy experienced a fresh upsurge, at the same time as numbers of “French Muslims” (“Français musulmans”) from Algeria came to work in metropolitan France following the granting of free movement in 1946. The latter were 331,000 strong on the eve of Algerian independence. The Italian population reached a second peak in 1962 (629,000 persons), while the Portuguese population reached its maximum in 1975 (759,000) and the Algerian population in 1982 (805,000). The migration streams underwent diversification in the 1960s and 1970s, when they spread to include the other Maghreb (North African) countries, Turkey and later many others (notably of sub-Saharan Africa). Some of these migration streams originated in newly independent former colonies (Maghreb and sub-Saharan Africa). Independence was achieved in 1956 for Tunisia and Morocco, in 1960 for most of the countries of sub-Saharan Africa, and in 1962 for Algeria. The Algerian migration stream, however, which had started before independence, was unique in that maximum inflow occurred during the early years of the new Algerian state, in 1962-63, a few months after the end of the war.
The demographic effects on the population of France reflect the intensity and long standing nature of these migration streams, factors which influence the formation of families of immigrant descent in France. Marital status at the time of entry and the speed at which family reunification (regroupement familial) occurred also account for the differences in the timing of this process. For men who were already married when they arrived or who returned to their country of origin to marry, no child would be born in France as long as the wives had not come to join their husbands. Family reunification migration occurred much later among migrants from North Africa and Turkey than among those from southern Europe. Some streams were active during the two great migratory waves of the last century and thus contributed heavily to the first and second generations born in France, which are present today. This is the case for migration from Italy and Spain. But the totals given for foreigners or immigrants in successive censuses do not reflect the true impact of migration currents on the demography of France. The children and grandchildren of immigrants quickly acquire French nationality, when they do not have it from birth, and thus disappear into the undifferentiated mass of French nationals born in France. Ironical as it may seem for a country with a long tradition of immigration, few demographers have attempted to estimate the demographic contribution of foreign immigration. Only two such attempts were made before 1945. In 1927, Alfred Sauvy estimated the number of “Frenchifications” (“francisations”) that occurred between 1872 and 1927
That is, people who were French nationals in 1927 but... at 1.2 million, representing one-half of France’s population growth over this period. Fifteen years later, in 1942, Pierre Depoid attempted an exercise along similar lines, and used the numbers acquiring French nationality to estimate the number “Frenchified” (“Francisés”) between 1871 and 1940 at 2.9 million. He also gave an estimate of the children with one or both foreign parents (2.7 million). In 1945, he published another presentation of his results based on the assumption that the population of foreign origin is equal to all the people born to two foreign parents, to which he added half the persons with only one foreign parent (1.9 million). By adding to this the 2.2 million foreigners present in 1940, he estimated the foreign and foreign-origin populations at 4.1 million persons, equivalent to 11% of the population of France (Landry, 1945). After the Second World War, interest in the demographic effects of immigration did not reappear until the 1980s.
A new study conducted in the mid 1980s set out to reconstitute the population of foreign origin resident in metropolitan France produced by one hundred years of immigration (Tribalat et al., 1991). Immigration flows recorded since the end of the nineteenth century were used to estimate the number of people born in France and still resident on 1 January 1986, with at least one parent or grandparent who had immigrated during the previous one hundred years. This number depends of course on the volume of immigration, but also on the fertility of immigrant couples and on the extent of intermarriage between populations. The estimate we obtained was between 9.4 and 10.3 million people born in France with at least one parent or grandparent who had immigrated in the previous one hundred years. Added to 3.9 million immigrants, this gave a total of between 13.3 and 14.2 million. The available information was so incomplete that many assumptions had to be made, notably about fertility by the nationality of the fathers and mothers, to evaluate the first generation born in France. The validity of the results obtained could not be checked by comparing them with another source. An alternative way of making this estimation would have been to use a survey that supplied information on ancestry, but no such data were available. Information on ancestry had been collected in earlier surveys, in 1951 and 1971, on French attitudes towards immigration (Girard, 1971). But besides being restricted to nationals aged 20 and over, the sample size was small and the question content was not entirely adequate
The 1971 survey was on 2,693 persons and the questions.... More recently, a few surveys have indeed collected information on birthplace but on limited numbers. In 1992, the MGIS
This survey was designed to explore the assimilation... (Mobilité géographique et insertion sociale – Geographical Mobility and Social Integration) survey went considerably further, since by combining information on the country of birth of the parents and their present or former nationality, a sample of young adults who were children of immigrants could be extracted from the Demographic Longitudinal Sample (Échantillon démographique permanent – EDP). In 1999, the Family Survey conducted in conjunction with the population census, which took the name Etude de l’histoire familiale (EHF – Study of Family History) that year and was designed and conducted jointly by INSEE and INED (Cassan, Héran, Toulemon)
See also the survey’s bilingual French and English..., for the first time collected information on parental country of birth. When this is combined with the country of birth of respondents and their children, information is available on three successive generations (see Box). The exceptionally large sample size — 380,481 men and women aged 18 or over — makes it possible to update and validate the 1986 estimate. In addition, the survey includes a module on parental linguistic practices during the early childhood of the respondents and of their own children
The processing of this module received funding from..., information that is decisive in the new estimation.
In the 1986 estimation, the shortcomings of the available data and the complexity of the assumptions used meant that the total estimate could not be broken down by country of origin. With the EHF survey this is now possible. This precision is crucial to the extent that it replaces in a historical dimension — and reduces to their true proportions — the most recent migration streams, on which the attention of public opinion is focused.
Box: Information available in the Study of Family History survey and useful for the estimation
The Study of Family History (EHF) survey conducted by INSEE has been enriched by census data, making it possible, for example, to know the birthplace and nationality of the respondents (aged 18 or over). With this extra information, the immigrant population is reconstituted by adding the adult respondents born abroad, of foreign nationality or naturalized French (1), to the children born abroad of immigrant respondents (2). The first generation born in France is composed of children under 18 born in France to immigrant respondents (4) and the respondents (all of whom are 18 or over) born in France with at least one foreign-born parent (3). The second generation born in France is estimated by working at the “children” level and selecting those born to parents (respondents) born in France with at least one foreign-born parent (5).
But this survey suffers from a defective representation of immigrant populations relative to the 1999 Census, which was itself probably less satisfactory than the previous ones. Unlike previous Family Surveys, this one was not compulsory. Attempts by INED and INSEE researchers to devise a weighting that corrects for this defect have not been completely satisfactory. The least bad weighting produces an over-representation of some segments of the immigrant population (women from Turkey and male and female immigrants from Italy). The underenumeration rate in the EHF survey relative to the census remains high for men from Portugal (16%), Turkey (17%), Algeria (19%), sub-Saharan Africa (25%) and above all Tunisia (35%). As an extra precaution, therefore, the immigrant numbers were taken from the 1999 Census. For the children of immigrants (obtained from the reports of respondents in the EHF survey), we have re-weighted the results to correct for this underenumeration, based on the assumption that those who had started a family (and hence their children) were less affected by it than immigrants in general. An underenumeration rate of half as much was thus applied for children born in France to immigrant parents.
I - Method of estimation
As in 1986, the objective is to estimate the population composed of immigrants and of people born in France with at least one immigrant among their parents or grandparents, which we refer to as the “population of foreign origin”, even though some are not merely of foreign origin but are actually foreign nationals, while others by contrast have only one immigrant parent or grandparent. For consistency with our previous research, immigrants are taken to be individuals who were of foreign nationality when they came to France, whatever their age when they entered the country (Tribalat, 1989).
As was the case for the 1986 estimation, some assumptions are in part arbitrary, because of the exercise itself. In 1945, when Pierre Depoid added half the people with one foreign parent to those with two foreign parents he too was making an arbitrary assumption. This must be accepted or the exercise abandoned.
When aggregating populations of different origins and of different generations, we need to guard against two possible forms of double-counting. The same person can be both the child and grandchild of immigrants and can, through both parents or through several grandparents, have several foreign origins.
Regarding the former point, the same solution is adopted as in 1986: a person born to an immigrant parent and to a parent born in France who is also the child of an immigrant, will be classified in the first generation born in France and removed from the second generation born in France. The immigrant status of one of the parents is determinant.
On the second point, an a priori hierarchy of countries of origin had to be established, which in decreasing order is: Algeria, Morocco, Tunisia, country of sub-Saharan Africa, Turkey, Italy, Spain, Portugal, other EU country, other country of origin. For example, a child whose father was an immigrant from Algeria and whose mother was an immigrant from Spain, is considered as being of Algerian origin. This hierarchy was established in a way that does not under-represent the origins currently monopolizing attention, so as to avoid the charge of choosing an order which causes them to be under-estimated. But a different order would have made little difference to the results. With the exact opposite order, the composition of the first generation of adults born in France is little changed: 4% fewer adults of Algerian origin, and 6% and 2% more adults of Moroccan and Tunisian origin, respectively; 2% fewer adults of Italian origin, but 3% and 4% more adults of Spanish and Portuguese origin, respectively; 3% more adults from a country of sub-Saharan Africa and 4% more from Turkey; 9% more adults from other countries of origin. For all those from another EU country, the result would have been very slightly lower (– 0.5%).
Some specific origins cannot be identified with the country of origin coding currently employed by INSEE. It is clear that this was thought in relation to post-World War II migration. The contribution of Belgian, Polish or Russian immigration therefore cannot be evaluated.
A problem also arises over using reports by respondents about their children. The necessary conditions to be included with the children are, for obvious reasons, that the parents are not dead and are present in France at the time of the survey. The first source of underestimation — mortality — can be adjusted for using the probabilities of being fatherless or motherless
See the results drawn from the EHF survey by Alain.... The effect of parental mortality is taken to be negligible for minors (i.e. under-18s) of the first generation born in France. As regards the second generation born in France, for those originating in the old immigration countries
Italy, Spain, other EU country, other country of o..., the adjustment was made only for people born between 1945 and 1982, the numbers in the earlier cohorts being too small and too variable. In doing this we rejuvenate the population pyramid for this fraction of the second generation born in France by removing a large proportion of its members aged 55 or over. For the descendants of immigrants from Algeria, the second generation born in France is younger and the adjustment concerned people born between 1960 and 1982. For Portugal, it concerned those born between 1970 and 1982. Overall, the second generation born in France is underestimated by a few hundred thousand individuals. The second source of underestimation cannot be adjusted for since we do not know what proportion of the children of immigrants (first generation born in France) have decided to leave and settle abroad — in the country of origin of their parent(s), for example — after having raised their children who, for their part, have remained in France. It is assumed to be negligible.
1 - Estimation of the first generation born in France excluding that of African origin: the children of immigrants
The adults of the first generation born in France are obtained directly from the reports of respondents about their parents’ country of birth.
The survey used a sample from which minors are absent. Consequently the respondents’ reports about their children have to be used to estimate the children aged under 18 of immigrants, separately from the over-18s. But the reports of men cannot simply be added to those of women, since most children of the first generation have two immigrant parents. We have chosen the children aged under 18 present in France reported by female immigrants, to which are added the children under 18 present in France of male immigrants who have formed only one union and whose partner (or ex-partner) was born in France
The information about the partner comes from the 1999....
The specific case of those of North African and sub-Saharan African origins will be examined later. Additional assumptions are required for these origins in order to exclude the migrations of repatriates or more generally of former overseas’ residents, and their demographic effects. But the overall logic is the same.
2 - Estimation of the second generation born in France excluding that of African origin: the grandchildren of immigrants
Estimation of the second generation born in France is based on the reports to the EHF survey of adults about their children. It concerns adults and children under 18 without distinction. Overall, this generation is composed of the children of mothers born in France one or both of whose parents was an immigrant, and the children of fathers born in France one or both of whose parents was an immigrant. Its estimation depends on the elimination of double counting.
This is because a problem arises when one parent is the child of an immigrant and the other parent is an immigrant. The child has already been counted in the first generation born in France. This is the case for children whose mother is the child of an immigrant and whose father is an immigrant, who are already counted on the father’s side
According to the 1992 MGIS survey, among women aged... in the first generation. Symmetrically, the children of immigrant mothers and of fathers with one or both parents immigrants, already counted in the first generation on their mother’s side, are not counted a second time.
A similar problem arises with the children of parents who are themselves the children of immigrants by one or both parents, who can be counted in the second generation on either the mother’s or the father’s side. In this case, double counting is avoided by removing them from the second generation born in France on the father’s side.
In all, on the father’s side we count only the children whose mother was born in France to parents born in France. Here too, we obtained indications from the MGIS survey
We include 50% of the children of Algerian, Portuguese...
3 - Estimation of the first and second generations born in France originating from the countries of the Maghreb and sub-Saharan Africa
Decolonization triggered the return to France of former colonists who are usually referred to as repatriates (rapatriés). This was the case for the countries of the Maghreb and sub-Saharan Africa. The children of these repatriates must be distinguished from those of immigrants originating from these countries.
Having information about the parental country of birth is a definite progress but is still inadequate for analysing the contributions of the migration streams from the Maghreb and sub-Saharan Africa
This would require introducing an additional question,.... Algeria was a département of France until 1962, and Tunisia and Morocco remained French protectorates until 1956. At the time of colonization, these countries already contained large and long established populations of Jews, whose Dhimmi status
An inferior status reserved for Jews and Christians... had been removed by the arrival of the French (Ye’Or, 1980). Algerian Jews were granted full French citizenship by the Cremieux decree of 24 October 1870, while those of Morocco and Tunisia retained their nationality. During the colonial period, these territories, and especially Algeria, were settled by large numbers of French people but also by foreigners (Spaniards, Italians, etc.), many of whom later became French. Most Moroccan Jews left when the state of Israel was established and especially after Morocco gained independence in 1956: 90% emigrated to Israel, the remainder to Canada and France (Tolédano Attias, 2003 ; Bin Nun, 2003). Hence France received very few Moroccan Jews. On the other hand, many of the Jews who left Tunisia, mainly after it became independent, sought refuge in France (slightly less than half of the 110,000 Jews resident in Tunisia when the Jewish state was formed). The two most intense phases of this emigration were the Bizerta crisis in 1961
A crisis between France and Tunisia who “waged war... and the Six Day War in 1967
Many left after the pogrom of the Six Day War. This.... Lastly, “More than 140,000 Jews lived in Algeria on the eve of decolonization in 1962 […] Most emigrated to France with the other repatriated settlers”. Very few went to Israel, and of these half are believed to have done so before 1962 (Liling, 2003).
The situation is extremely complex and the solution adopted to separate the descendants of Maghrebin immigrants from the descendants of repatriated settlers is bound to be imperfect. What is obvious, however, is the need to include only descendants of immigrants who satisfy the definition of immigrants. In addition, as we shall see and as Richard Alba and Roxane Silberman (2002) have shown, they differ sharply by social origins
Alba and Silberman have also developed a method for.... To abandon the attempt would result in underestimating the scale of the difficulties encountered by the populations of Maghrebin origin while at the same time overstating their number. For Algeria, therefore, the repatriates, Jews or non-Jews, will not be considered as immigrants, since most were French nationals by birth. Nearly a million moved to France when Algeria became independent (Baillet, 1976)
Also arriving in France at this time were the “French.... Algerian-born French nationals by birth still numbered 660,000 in the EHF survey. The same solution is adopted for Morocco: since most Moroccan Jews chose to go to Israel, there are few among the immigrants in France. For Tunisia the situation is more complicated. The immigrants, as defined here, include a certain number of Jews, some of whom came to France when they had already acquired French nationality. According to the 1999 survey, 318,000 French nationals by birth were born in Tunisia and are thus repatriates. Given the information available in the EHF survey for distinguishing the children of repatriates from the children of immigrants, as will be seen, our estimate for the descendants of immigrants will probably not include the children and grandchildren of the few tens of thousands of Jews who sought refuge in France.
As a result of colonization
These African countries were the Territoires d’Outre-Mer... and later of overseas service, sub-Saharan Africa was also settled by French people who had children there, some of whom subsequently came back to live in metropolitan France after independence. Certain persons with the nationality of a sub-Saharan African country could also, after the independence of their country, if they “came to reside habitually and continuously on the territory of the French Republic, be attributed French nationality by declaration”, a possibility that existed until 12 July 1973. “The beneficiary was held never to have ceased being French even though becoming a national of the new state” (Massicot, 1986). In the EHF survey, more than 130,000 people born in a sub-Saharan African country reported themselves as French nationals by birth and are not counted among immigrants (because they were former overseas residents or were entitled to French nationality because of their residence in France). Whereas for the Maghreb we use the term repatriates, for sub-Saharan Africa we use the label “French nationals by birth from sub-Saharan Africa”.
In all, the problem of distinguishing between, on the one hand, the descendants of immigrants, and on the other, the descendants of repatriates and French nationals by birth from sub-Saharan Africa, arises uniquely for the over-18s of the first generation born in France and for the totality of the second generation, since with the EHF survey we know only the country of birth of the parents of the respondents. By contrast, the under-18s of the first generation born in France can clearly be identified because they were reported by their parents (respondents), whose birthplace and present or former nationality are known.
This problem can be overcome using the module of the EHF survey on the languages used in childhood and information from the MGIS survey.
Adults of the first generation born in France
The 1999 EHF survey collected information on the language spoken with the respondents’ parents, by means of the following question: “In which language, dialect, or ‘patois’ did your parents usually speak to you when you were a child, about five years old?” This question had to be answered separately for the mother and the father, and in two ordered categories: “usually spoke to you in …” ; “and also in …”.
The parents originating from Maghreb countries spoke to their children (the respondents) in Arabic or Berber
There are numerous variants but all contain a reference..., but also in other, usually European languages, by and large French. We consider that those who spoke Arabic or Berber to the respondents are necessarily immigrants and, consequently, that their children (the respondents) are the children of immigrants, and this even when these languages were spoken by just one of the parents in conjunction with French (or some other European language). The adults (i.e. over-18s) of the first generation born in France who are in this position will be referred to as of Arabic or Berber mother tongue. But a proportion of the respondents whose parents had never spoken to them in either Arabic or Berber are also the children of immigrants (and not all children of repatriates). For the sake of simplicity, and because it represents the overwhelming majority of cases, these respondents are said to be of exclusively French mother tongue; among them, we still have to estimate how many are children of immigrants. A similar procedure is adopted for the respondents whose parents were born in sub-Saharan Africa, who may have been exposed to many languages in childhood.
For immigrants originating from Algeria, Morocco and sub-Saharan Africa, the 1992 MGIS survey supplies information about the extent to which the languages of the country of origin were used for communicating with their children, in conjunction or not with French
60% for Algeria, 70% for Morocco, and 50% for sub-Saharan.... This information is used to estimate, for each origin, the total number of children of immigrants, that is, including those of French mother tongue. To do this the number of children of immigrants of Arabic or Berber mother tongue for the Maghreb, and of a sub-Saharan African mother tongue for sub-Saharan Africa, drawn from the EHF survey, is divided by the proportion of immigrants in the MGIS survey who reported speaking to their children in Arabic or Berber, or a language of sub-Saharan Africa, in conjunction or not with French.
The only way children of immigrants could be distinguished from children of repatriates or of French nationals by birth from sub-Saharan Africa, was by making the following assumptions: 1) children of immigrants who are also children of repatriated settlers or of French nationals by birth from sub-Saharan Africa are few in number; 2) repatriates who spoke Arabic, Berber or a language of sub-Saharan Africa and who used it with their children born in France are rare. Although many North African Jews spoke Arabic or Berber, it is reasonable to think that they did not use these languages with their children born in France, where they had sought refuge. The children of Jews who had come from Tunisia are thus probably not counted in the first generation born in France, although their parents do get counted among the immigrants if they report as Tunisian nationals or Tunisian-born. Conversely, the children of French nationals by birth from sub-Saharan Africa can appear in the first generation born in France if they were spoken to in a sub-Saharan African language, although their parents are not counted as immigrants.
One approach to estimating the size of the first generation born in France would have involved simply counting the number of children reported by the immigrants and adjusting the results to allow for the mortality of respondents (the probability of the children being orphans). But this solution has two disadvantages. First, the effect of mortality, which is large at high ages, can be only partially adjusted for. Second, it cannot be used on its own for estimating the second generation born in France, since in this case it is necessary to use the answers relative to parental birthplace of those present in the survey belonging to the first generation born in France. Although the method based on the number of children reported by the immigrant respondents was not used, it was nonetheless applied in the case of children aged over 18, so as to provide a point of comparison and thus of validation for the estimate based on language use for the over-18s of the first generation born in France (Table 1).
Table 1 –
Estimation of adults of the first generation born in France: estimate based on language use compared with that based on reports of immigrant respondents about the number of their children born in France (thousands)
Estimate by language use(1) Estimate based on number of children born in France reported by immigrant Respondents Algeria: 60% 472 428 70% 404 428 Morocco: 60% 161 173 70% 145 173 Tunisia: 50% 82 103 Sub-Saharan Africa: 50% 47 64 (1) Estimate based on the reports of respondents concerning (a) the country of birth of their parents and (b) the language spoken to them when aged five by their parents, to which an assumption is applied about the proportion of immigrants speaking to their children in a language of their country of origin. Reading: line 1: When it is assumed that 60% of immigrants from Algeria spoke Arabic or Berber to their children, the number of their children aged over 18 belonging to the first generation born in France is estimated at 472,000. By comparison, using the reports of immigrants from Algeria about the number of their adult children gives a figure of 428,000. Source: INSEE, 1999 EHF survey.
The second column of Table 1 gives the estimates based on language use, while the third column comes from the reports of immigrants on their children born in France who are now aged over 18 and resident in France (adjusting for the underenumeration of immigrants in the survey — see Box — and for the mortality of respondents). Two assumptions have been tried for the persons originating from Algeria and Morocco: these are that 60% and 70% of immigrant parents spoke an Arabic or Berber language with their children during childhood. The first proportion derives from the reports of parents from Algeria in the MGIS survey, and the second from the reports of parents from Morocco
The same proportion was applied uniformly to all a.... The estimation based on the proportion of 60% of parents who spoke Arabic or Berber with their children probably overestimates the number of over-18s in the first generation born in France of Algerian origin. But the 70% assumption certainly underestimates it. Accordingly the first assumption has been chosen (corresponding to 472,000). For persons of Moroccan descent, both estimations are slightly low, but here too the 60% assumption appears preferable (161,000).
For immigrants from Tunisia, who were not included in the MGIS survey, the proportion who spoke Arabic or Berber with their children was set at 50%
This proportion is lower than the assumptions made.... The number of over-18s, descendants of immigrants from Tunisia, is also slightly low (82,000). This may be explained in part by the fact that the estimate based on language use eliminates the children of Tunisian Jews who themselves are counted as immigrants (see above). Lastly, on the basis of a 50% assumption (taken from the MGIS) the number obtained is a little low for the over-18s of the first generation originating from sub-Saharan Africa (47,000).
Overall, the estimate of the number of persons aged over 18 in the first generation born in France obtained by combining use of the reports in the EHF survey on parental birthplace and language use, results in lower totals, excep for those of Algerian origin, than those obtained on the basis of the number of children born in France reported by immigrant respondents. It is possible, therefore, that it overestimates the children of Algerian migrants while underestimating the others. We can add that the total underestimation slightly exceeds the overestimation. Spread over some five million persons, however, as we shall see, this uncertainty appears acceptable
The information about the mother tongue also proves....
Adults and children of the second generation born in France
The second generation of descendants of immigrants from sub-Saharan Africa and from Turkey has not been estimated, since immigration from these sources is on the whole very recent and the second generation would represent only a few thousand people, which would not have substantially modified the results given the margin of error we are working with. For the second generation born in France of Maghrebin descent, the approach used is the same as for the first generation: estimation by language use is used to distinguish the descendants of immigrants from those of repatriates, by applying to the next generation the assumptions made about the proportion of Arabic or Berber speakers in the first generation.
II - The population of foreign origin by origin and generation
This processing of data from the 1999 EHF survey gives a total for the population of foreign or partly foreign origin of nearly 13.5 million persons, if we use the information available over a sequence of three generations: immigrants resident in metropolitan France, their children and their grandchildren. This population breaks down into 4.3 million immigrants, 5.5 million children of immigrants, and 3.6 million grandchildren of immigrants. Without the underestimate of the second generation born in France above age 55, the number of persons with at least one immigrant grandparent would probably be close to 4 million. Even after allowing for the underestimation, this population of foreign origin remains on average slightly younger than the population of metropolitan France as a whole — the proportion of persons aged under 40 is larger, while that over 60 is smaller.
The generation of immigrants contains proportionally fewer women than the first generation born in France where, logically, they outnumber men (Table 2). We should observe the same pattern from age 30 in the second generation born in France. But this is not the case. This generation was estimated from the reports of respondents about their children, and it is possible that the few instances of child omission more often affected girls — as happened for children aged under 18 of the first generation reported by immigrants from certain countries (Algeria and Morocco, for example). In addition, not adjusting for the effects of mortality of respondents above age 55 also works against the female presence.
Table 2 –
Estimation of the foreign-origin population, by generation and age group (thousands)
Immigrants First generation born in France Second generation born in France Total Men 2,167 2,736 1,865 6,768 Under 18 132 927 618 1,677 18-39 684 916 785 2,385 40-59 841 561 449 1,851 60 and over 510 332 13 855 Women 2,140 2,795 1,779 6,714 Under 18 128 876 592 1,596 18-39 743 919 758 2,420 40-59 742 578 416 1,736 60 and over 527 422 13 962 Total 4,307 5,531 3,644 13,482 Under 18 260 1,803 1,210 3,273 18-39 1,427 1,835 1,543 4,805 40-59 1,583 1,139 865 3,587 60 and over 1,037 754 26 1,817 Source: INSEE, 1999 EHF survey.
In the total population of metropolitan France (59 million), it would appear that 23% are of foreign origin (24% after adjusting roughly for the underestimation above age 55). This proportion varies little with age before 60 years (Table 3). Up to age 60, roughly a quarter of men are of foreign origin. This is also the case for women, but up to age 40. Thereafter the proportion of women of foreign origin declines because the contribution of female immigrants lessens. The relative weights of the three generations in the French population vary depending on age. Above age 40, immigrants dominate, while below this age it is the first generation born in France, especially among the under-18s. This latter age group also makes the largest contribution to the second generation born in France.
Table 3 –
Estimation of the contribution of the foreign-origin population to the population of France, by age group and generation (%)
Immigrants First generation born in France Second generation born in France Total Men Under 18 2.0 14.2 9.4 25.7 18-39 7.5 10.1 8.7 26.3 40-59 11.1 7.4 6.0 24.5 60 and over 9.6 6.3 0.2 16.2 Total 7.6 9.6 6.5 23.8 Women Under 18 2.1 14.1 9.5 25.6 18-39 8.2 10.2 8.4 26.8 40-59 9.7 7.5 5.4 22.7 60 and over 7.3 5.9 0.2 13.4 Total 7.1 9.3 5.9 22.3 Source: INSEE, 1999 EHF survey.
Public debate in France in recent years has been prompt to equate the population of foreign origin with its Maghrebin component. Yet individuals of Maghrebin origin belonging to the three generations studied here represent only around three million persons, or just 22% of the total foreign-origin population (Table 4). This is far short of the more fantastical figures that have appeared in the press
In particular the figure of six million put forward.... Among those of Maghrebin origin, slightly more than half are immigrants or descendants of immigrants from Algeria.
Table 4 –
Estimation of the foreign-origin population, by country of origin and generation (thousands and %)
Country of origin Immigrants First generation born in France Second generation born in France Total Number % Number % Number % Number % Maghreb 1,299 30.2 1,431 25.9 269 7.4 2,999 22.2 Algeria 574 13.3 783 14.2 220 6.0 1,577 11.7 Morocco 523 12.1 451 8.2 31 0.9 1,005 7.5 Tunisia 202 4.7 197 3.6 18 0.5 417 3.1 Sub-Saharan Africa 393 9.1 286 5.2 – – 679 5.0 Turkey 174 4.0 148 2.7 – – 322 2.4 Southern Europe 1,267 29.4 2,079 37.6 1,819 49.9 5,165 38.3 Italy 379 8.8 1,020 18.4 1,158 31.8 2,557 19.0 Spain 316 7.3 577 10.4 571 15.7 1,464 10.9 Portugal 572 13.3 482 8.7 90 2.5 1,144 8.5 Other EU country 363 8.4 646 11.7 793 21.8 1,802 13.4 Other country of origin 812 18.9 940 17.0 762 20.9 2,514 18.6 Total 4,307 100.0 5,531 100.0 3,644 100.0 13,482 100.0 Source: INSEE, 1999 EHF survey.
The population of Algerian origin (1.6 million) is heavily outnumbered by that of Italian origin (2.6 million), for which immigration began at the end of the nineteenth century, and is equal in size to that of Spanish origin (1.5 million) which is also older-established. For the Algerian stream, the first generation born in France is far from complete and the second generation is just beginning, whereas for the other two migratory streams these generations have had time to form. In all, when 1.1 million individuals of Portuguese origin are included, southern Europe has made the largest contribution to the population of France (5.2 million, equivalent to nearly 40% of the population of foreign origin). The impact of the older Belgian immigration is submerged in the population originating from another EU country (1.8 million) and that of the Polish and Russian streams in the other country-of-origin category (2.5 million).
The proportion of mixed unions made by the parents of the respondents is a major determinant of the growth of the foreign-origin populations down the generations. Since the four grandparents are not captured in the EHF survey, this impact on the second generation born in France cannot be measured. The parental countries of birth are not given for the partner(s) of the respondents. On the other hand, the parental countries of birth for the adults of the first generation born in France can be combined in order to distinguish those born to unions between two immigrants and those born to mixed couples in which one partner is an immigrant. For the Maghreb and sub-Saharan Africa, attention is limited, respectively, to respondents whose mother tongue is Arabic or Berber, or a language of sub-Saharan Africa, since among those of French mother tongue we cannot distinguish between the children of immigrants and the children of repatriates. The data are presented in Table 5, and together with those in Table 4, give a good idea of the impact of mixed unions.
Table 5 –
Distribution of adults of the first generation born in France, by type of parental couple (%)
Country of origin Two immigrant parents Immigrant father, mother born in France Immigrant mother, father born in France Total Algeria(a) 93 6 1 100 Morocco(a) 96 4 0 100 Tunisia(a) 89 8 3 100 Sub-Saharan Africa(b) 77 13 10 100 (a) Only those of Arabic or Berber mother tongue. (b) Only those of a sub-Saharan African mother tongue. Turkey 70 18 12 100 Italy 46 36 19 100 Spain 46 33 21 100 Portugal 73 20 7 100 Other EU country 20 38 42 100 Other country of origin 38 35 27 100 Source: INSEE, 1999 EHF survey.
Among adults of the first generation of Maghrebin descent of Arabic or Berber mother tongue, the proportion of mixed unions among the parents is low and contributes to reducing the size of this generation. But it is possible that those of exclusively French mother tongue more often had one parent born in France. The result for the children of Turkish migrants is not surprising when we remember that their number includes children of Armenians. For the latter, as for all the older streams, the immigrant parents usually came during childhood and were at least in part brought up in France
We had already observed this phenomenon in the MGIS.... Children of Italian and Spanish descent are more likely to have a mother or father born in France than to have two immigrant parents. This is also the case for persons of other European origins. The Spanish and Italian immigration streams were important before the Second World War (from the late nineteenth century for the latter) and reached a new peak in the immediate post-war period. For the long-established streams, an apparent heterogamy of origins for the parental couple can nonetheless hide a form of homogamy, when the partner born in France is in fact the descendant of an immigrant. But the marked predominance of the population of Italian origin also results from the frequency of mixed unions. The more recent Portuguese migration stream has produced fewer mixed unions, at least when these are apparent to statistical view, though probably far more than those from the Maghreb countries.
Depending on how long ago the migration streams began, young people have a greater or lesser relative importance in the foreign-origin populations. For the sub-Saharan African and Turkish streams, the first generation is very recent; more than three-quarters of the populations of these origins are below age 40, and under-18s account for 40% of the total. Conversely, the longstanding nature of Italian migration has allowed time for the first and even the second generation to be constituted, and all ages are present in more balanced proportions: 55% of persons under 40 and 20% of minors, which is a situation close to the national average.
The contribution of each stream to each of the generations, by age groups, highlights the temporal differentials in the formation of the foreign-origin populations (Table 6).
Table 6 –
Contribution of different migration streams to the formation of three generations of the foreign-origin population (%)
Country of origin Immigrants First generation born in France Second generation born in France Men Women Men Women Men Women Maghreb Under 18 28.6 27.1 40.7 41.6 18.5 19.2 18-39 33.0 30.4 34.4 33.5 2.2 3.1 40-59 36.2 30.9 5.9 4.4 - - 60 and over 30.4 16.4 1.6 0.6 - - Total 33.4 26.9 26.7 25.0 7.1 7.7 Maghreb + Sub-Saharan Africa + Turkey Under 18 49.6 49.2 59.5 61.0 18.5 19.2 18-39 54.1 51.1 37.6 36.9 2.2 3.1 40-59 49.1 40.7 7.4 5.7 - - 60 and over 33.9 19.0 3.6 2.1 - - Total 47.1 39.4 34.7 32.7 7.1 7.7 Southern Europe Under 18 9.2 9.5 23.4 22.6 50.4 50.2 18-39 20.1 18.4 40.6 41.2 48.9 48.2 40-59 30.2 32.6 47.9 47.6 51.4 51.6 60 and over 43.0 48.1 47.4 50.6 58.5 62.5 Total 28.7 30.1 37.1 38.1 50.1 49.8 Other EU country Under 18 10.3 10.1 5.6 5.4 14.2 14.4 18-39 6.3 8.4 8.4 9.2 24.1 24.7 40-59 5.9 9.2 19.2 19.7 27.1 27.6 60 and over 8.9 12.9 20.6 22.5 30.8 27.4 Total 7.0 9.9 11.1 12.2 21.6 22.0 Other country of origin Under 18 30.8 31.2 11.5 11.0 16.8 16.2 18-39 19.6 22.1 13.5 12.7 24.8 24.0 40-59 14.8 17.6 25.5 26.9 21.5 20.9 60 and over 14.2 20.0 28.3 24.8 10.7 10.1 Total 17.2 20.6 17.1 16.9 21.3 20.5 Source: INSEE, 1999 EHF survey.
While Maghrebins represent only 22% of the total population of foreign origin, they are the most numerous in the generation of immigrants (30%) and among the under-18s of the first generation born in France (41%). When the more recent streams from sub-Saharan Africa and Turkey are included, the total represents half of the immigrant generation under age 40 and 60% of the first generation born in France under age 18
For an estimate of the number of people living in France.... In other words, a very large proportion — six out of ten — of the children born in France to one or two immigrant parents, who are currently minors and going through the French school system, have origins in these countries. By contrast, above age 60, migrants from southern Europe clearly dominate in the generation of immigrants (43% of men and 48% of women). Proportionally less represented among the minors born in France to one or both immigrant parents, southern Europe makes the largest contribution to this generation among the adults, with its share coming close to 50% above age 40. In the second generation born in France, it forms the majority at practically all ages and still more so above age 60. Individuals originating in other EU countries, who count for so little (8%) in the immigrant generation, represent 12% of the first generation born in France, especially at the higher ages (21% above age 60), and 22% of the second generation born in France. Here too, the higher the age, the higher the proportion of persons originating from an EU country outside southern Europe. This attests in particular to the effects of the much older Belgian immigration, of the late-nineteenth and early-twentieth century. The “other origins” category includes both older (Russia, Poland) and more recent (China, former-Yugoslavia) streams, and its contribution to the three generations of the population of foreign origin reflects this strange combination: 30% of immigrant minors, over 25% of immigrants’ children aged 40 and over, and 24% of grandchildren aged 18-39.
III - The importance of the distinction between children of immigrants and children of repatriates
After the 1999 Census, INSEE did not publish a table showing the French nationals by birth born in the Maghreb countries and sub-Saharan Africa. This information has thus been taken from the EHF survey. Under the assumptions set out earlier for distinguishing between children of immigrants and children of repatriates or French nationals by birth from sub-Saharan Africa, this survey can be used to calculate the distribution by age group of the first generation born in France to parents who were French nationals by birth born in these former colonized countries
A person can be the child both of a repatriate and.... Over two generations, the population formed by the repatriates from Algeria and their children (1.37 million persons) is equal to that produced by Algerian immigration (1.36 million). The repatriates outnumber adult immigrants (660,000 against 545,000), and the first generation born in France is slightly smaller for the children of repatriates, though still slightly older.
If this estimation is correct, the error produced when all the children with at least one parent born in Algeria are equated with the children of immigrants from this country is potentially large and variable with age, unless it is assumed that they all have identical characteristics and behaviour. Overall, around half of the children born in France to one or both parents born in Algeria are descended from repatriates (Table 7). Less than 50% below age 40, this proportion rises at the higher ages. The different assumptions regarding the language spoken with the parents (see above) do not substantially modify these results.
Table 7 –
Estimation of the children of repatriates or of French nationals by birth from sub-Saharan Africa, as a proportion of persons born in France to parents born in Algeria, Morocco, Tunisia, or sub-Saharan Africa, by age group (%)
Algeria Morocco Tunisia Maghreb Sub-Saharan Africa Men Women Men Women Men Women Men Women Men Women Under 18 42 42 21 19 30 25 33 31 24 22 18-39 45 43 39 42 51 59 45 45 55 56 40-59 63 69 58 76 82 91 65 73 65 74 60 and over 81 89 100 100 57 100 80 91 100 87 Total 47 47 27 27 44 46 41 41 32 32 Note: These proportions are estimated on the basis of the assumptions made about the proportion of immigrants who spoke a language of their country of origin with their children (i.e. 60% for Algeria and Morocco, 50% for Tunisia and sub-Saharan Africa). Source: INSEE, 1999 EHF survey.
Around 370,000 persons are either repatriates or the children of repatriates from Morocco. The proportion of the children of repatriates among the children born in France to one or both parents born in Morocco is close to 30% overall. Situated around 20% for the minors, it rises sharply with age. For Tunisia, the number of repatriates and children of repatriates stands at 305,000 ; the proportion descended from repatriates in the first generation born in France, which is 45% on average, rises with age. Lastly, French nationals by birth born in sub-Saharan Africa and their children born in France represent close to 300,000 persons. The latter are on average older than the children of immigrants from sub-Saharan Africa, who are only just beginning to form families in France. Of the minors born in France to parents born in sub-Saharan Africa, only just over 20% are the children of French nationals by birth, as against more than half of the young adults (18-39 years). Moreover, if the calculation had been based on the number of children born in France reported by the EHF survey respondents born in Tunisia and sub-Saharan Africa (see Table 1), the proportion of repatriates and of children of French nationals by birth would still have exceeded 40% for the ages 18-39.
For the migration streams from the Maghreb and sub-Saharan Africa, it is thus misleading to think that the first generation born in France can be estimated using solely information about the country of birth of the parents. This is all the more troubling as it is precisely the populations produced by these migration streams that are the focus of current concerns. There is no reason why the situation and behaviour of the children of repatriates (or of French nationals by birth from sub-Saharan Africa) should be the same as that of the children of immigrants. Under our assumptions, the children of repatriates would represent 72% of all persons born in France to one or both parents born in Algeria of exclusively French mother tongue. This proportion would stand at 65% for persons from Morocco and 72% for those from Tunisia. Consequently, an aggregate-level analysis of the populations born in France to parents born in a Maghreb country and of exclusively French mother tongue will in fact contain more children of repatriates than children of immigrants.
Social origins differ sharply as a function of mother tongue. According to the EHF survey, 78% of the children of parents born in Algeria of Arabic or Berber mother tongue — i.e. 60% of all children of immigrants, under our assumption — have (had) manual worker fathers
A proportion close to that of young people aged 20-29..., compared with 32% for persons of exclusively French mother tongue. A large disparity is also observed among the children of parents born in Morocco (75% and 23%) or in Tunisia (67% and 30%). In the most favourable case, we can make the assumption that the characteristics of the children of immigrants of exclusively French mother tongue resemble those of children of repatriates: this supposes that the persons who in childhood were spoken to exclusively in French together form a homogeneous group. In the least favourable case, the profile of the children of immigrants with Arabic or Berber as mother tongue describes that of all immigrant children. The assumption that the children of immigrants of exclusively French mother tongue would be better placed than the children of repatriates will not be considered plausible. Depending on the country of origin, the two assumptions are not of equal likelihood.
This can be illustrated using the example of unemployment among the children of parents born in the Maghreb countries. The unemployment rate for persons with at least one parent born in Algeria (column 3 of Table 8) is the average of the very high rate for children of immigrants of Arabic or Berber mother tongue (column 1) and of the lower rate among children of parent(s) born in Algeria of exclusively French mother tongue (column 2). The latter is higher than the unemployment rate for persons of French descent
Born in France to two parents born in France. (column 4). The children of parent(s) born in Algeria and to whom only French was spoken include children of immigrants but above all children of repatriates from Algeria. The difficulties experienced by French nationals of Algerian descent
A shorthand term to denote adults born in France to... are thus seriously underestimated if we use the sole criterion of parental place of birth. The excess unemployment among adults of Arabic or Berber mother tongue increases with age, at least until 35-39 years, attesting to extraordinarily persistent difficulties. At ages 35-39, the overall unemployment rate for French nationals with one or two parents born in Algeria (3) is “only” two and half times higher than that for persons of French descent, compared with almost five times higher for those of Arabic or Berber mother tongue. Similarly, for men with at least one parent born in Algeria, for example, the unemployment rate at ages 40-44 would be half its level at ages 18-24, compared with a third lower among those of Arabic or Berber mother tongue.
Table 8 –
Effects on the unemployment rate of not differentiating between children of repatriates and children of immigrants from algeria
Unemployment rate on census definition (%) Excess unemployment for children with at least one parent born in Algeria (5) = (3)/(4) Assumption A* Assumption B** Born in France to at least one parent born in Algeria Born in France to two parents born in France (4) Notional unemployment rate (%) for children of immigrants (6) Excess unemployment for children of immigrants (7) = (1)/(4) Notional unemployment rate (%) for children of repatriates (8) Mother tongue Arabic or Berber (1) French exclusively (2) Total (3) Men 18-24 47.8 25.3 33.8 22.4 1.5 38.8 2.1 12.9 25-29 46.0 18.4 27.8 12.3 2.3 34.9 3.7 4.7 30-34 40.2 15.3 22.5 8.3 2.7 30.3 4.8 2.9 35-39 33.9 11.6 18.2 6.9 2.6 25.0 4.9 2.1 40-44 31.0 12.2 17.7 6.9 2.6 23.5 4.5 5.8 Women 18-24 52.7 37.2 42.5 30.5 1.4 46.5 1.7 28.9 25-29 34.6 21.8 26.1 18.7 1.4 29.5 1.9 15.2 30-34 37.9 19.4 25.4 14.5 1.8 30.5 2.6 12.2 35-39 33.6 22.2 25.8 12.2 2.1 29.1 2.8 17.3 40-44 30.1 11.5 16.5 9.9 1.7 22.7 3.0 4.1 * Assumption A: The children of immigrants of exclusively French mother tongue have the same unemployment rates as the children of repatriates. ** Assumption B: The children of immigrants of exclusively French mother tongue have the same unemployment rate as those of Arabic or Berber mother tongue. (6) Unemployment rate that the children of immigrants would have if it is assumed that those of French mother tongue have the same unemployment rate as the children of repatriates: (6) = (1) x 0.6 + (2) x 0.4. (8) Unemployment rate that the children of repatriates would have if the children of immigrants of French mother tongue have the same unemployment rate as those of Arabic or Berber mother tongue, the overall unemployment rate of the children with at least one parent born in Algeria of exclusively French mother tongue being equal to that of column (2). Source: INSEE, 1999 EHF survey.
Under the assumption that unemployment among the children of immigrants of exclusively French mother tongue is as low as among the children of repatriates (Assumption A), the unemployment of persons of Algerian descent is still much higher than would be expected just on the basis of parental country of birth (3).
If we assume that the children of immigrants of exclusively French mother tongue (representing 40% of all children of immigrants under the assumption selected) experience unemployment as often as the other young people of Algerian descent (Assumption B), the unemployment rates by age for the children of repatriates would, for the men, be too low to be credible. For the women, on the other hand, they are close to those of French female nationals born in France to parents born in France. Among men at least, therefore, adults from exclusively French mother tongue backgrounds have lower levels of unemployment than adults from Arabic or Berber mother tongue backgrounds
If the former represented not 40% but 30% of all children.... Taking into account only parental country of birth thus introduces errors of level and variations by age in the unemployment rate for the children of immigrants.
The Moroccan example (Table 9) supplies powerful reasons for not considering all children of immigrants as comparable with those from Arabic or Berber mother tongue backgrounds. Assuming the same unemployment rate for the children of immigrants of exclusively French mother tongue as for the latter has the effect of attributing a negative unemployment rate to the sons aged 18-24 of repatriates and a very low rate to those aged 25-49, neither of which are plausible. For French nationals of Moroccan descent, the conclusion has to be that the men who in childhood were spoken to exclusively in French have a level of unemployment closer to that of children of repatriates. An evaluation based on all the children of persons born in Morocco is less incorrect than if attention is limited to those of Arabic or Berber mother tongue. This is slightly less true at ages 25-49. And it is definitely less so for the women who probably occupy an intermediate situation, notably at ages 25-49. The excess unemployment of French nationals of Moroccan descent is thus high among young people, and primarily young men, but drops back to a more “normal” level above age 25, in contrast to the persistent difficulties faced by French nationals of Algerian descent.
Table 9 –
Effects on the unemployment rate of not differentiating between children of repatriates and children of immigrants from morocco
Unemployment rate on census definition (%) Excess unemployment for children with at least one parent born in Morocco (5) = (3)/(4) Assumption A* Assumption B** Born in France to at least one parent born in Morocco Born in France to two parents born in France (4) Notional unemployment rate (%) for children of immigrants (6) Excess unemployment for children of immigrants (7) = (1)/(4) Notional unemployment rate (%) for children of repatriates (8) Mother tongue Arabicor Berber (1) French exclusively (2) Total (3) Men 18-24 50.9 28.1 41.7 22.4 1.9 41.8 2.3 – 5.7 25-49 26.7 9.6 13.0 8.2 1.6 19.8 3.3 5.9 Women 18-24 46.8 37.6 42.5 30.5 1.4 43.1 1.5 23.2 25-49 38.7 19.1 22.8 13.0 1.8 30.8 3.0 15.1 * Assumption A: The children of immigrants of exclusively French mother tongue have the same unemployment rate as the children of repatriates. ** Assumption B: The children of immigrants of exclusively French mother tongue have the same unemployment rate as those of Arabic or Berber mother tongue. (6) Unemployment rate that the children of immigrants would have if it is assumed that those of French mother tongue have the same unemployment rate as the children of repatriates: (6) = (1) x0.6 + (2) x 0.4. (8) Unemployment rate that the children of repatriates would have if the children of immigrants of French mother tongue have the same unemployment rate as those of Arabic or Berber mother tongue, the overall unemployment rate for the children with at least one parent born in Morocco of exclusively French mother tongue being equal to that of column (2). Source: INSEE, 1999 EHF survey.
The Tunisian example produces the same result, i.e. supports the hypothesis of a low excess unemployment among the children of immigrants of exclusively French mother tongue, at any rate among the men. Among male French nationals of Tunisian descent, the difficulties for those aged 18-29 are certainly less than those faced by their counterparts of Algerian descent, but are still much greater than those for the children of French descent (Table 10).
Table 10 –
Effects on the unemployment rate of not differentiating between children of repatriates and children of immigrants from Tunisia
Unemployment rate on census definition (%) Excess unemployment for children with at least one parent born in Tunisia (5) = (3)/(4) Assumption A* Assumption B** Born in France to at least one parent born in Tunisia Born in France to two parents born in France (4) Notional unemployment rate (%) for children of immigrants (6) Excess unemployment for children of immigrants (7) = (1)/(4) Notional unemployment rate (%) for children of repatriates (8) Mother tongue Arabic or Berber (1) French exclusively (2) Total (3) Men 18-29 38.1 17.0 23.7 16.1 1.5 27.6 2.4 – 3.0 Women 18-29 42.5 28.4 31.9 22.8 1.4 35.5 1.9 17.1 * Assumption A: The children of immigrants of exclusively French mother tongue have the same unemployment rate as children of repatriates. ** Assumption B: The children of immigrants of exclusively French mother tongue have the same unemployment rate as those of Arabic or Berber mother tongue. (6) Unemployment rate that the children of immigrants would have if it is assumed that those of French mother tongue have the same unemployment rate as children of repatriates: (6) = (1) x 0.5 + (2) x 0.5. (8) Unemployment rate that the children of repatriates would have if the children of immigrants of French mother tongue had the same unemployment rate as those of Arabic or Berber mother tongue, the overall unemployment rate for children with at least one parent born in Tunisia of exclusively French mother tongue being equal to that in column (2). Source: INSEE, 1999 EHF survey.
These three examples are strong grounds for not favouring one hypothesis a priori. They also show the shortcomings of the “Maghrebin origin” category which, though it has the advantage of allowing analysis on larger total numbers, lumps together situations that are extremely variable by origin, sex, age, language use, etc. French nationals of Algerian descent face great difficulties and particularly at the ages where these have usually eased. The difficulties are less acute though still disturbing for the other French nationals of Maghrebin origin. These measurement problems will arise for all phenomena studied, with the same uncertainty and the same complexity.
The estimate of the population of foreign origin over three generations based on the 1999 EHF survey is ultimately very close to that obtained thirteen years earlier using another method. Allowing for the underestimation of the second generation born in France above age 55 and the few tens of thousands of people originating from a Maghreb or sub-Saharan African country missing from the first generation born in France, our figure for 1999 is closer to a total of 14 million immigrants or persons with at least one immigrant parent or grandparent, and thus stands at the high end of the range obtained in 1986: 14.2 million. Even at an interval of thirteen years, this closeness of the results lends credibility to both estimates, that of 1986 slightly overstating reality while that of 1999 may understate it slightly. The new estimate presents the great advantage of including a distribution by country of origin, therefore making it possible to situate the contributions of the various migration streams to the foreign-origin population in their historical dimension. The constitution of the first generation born in France of Maghrebin descent is far from complete and that of the second generation is barely started. We are situated at the heart of the process by which new Frenchmen are made. For the older streams, the impression today is that the matter is settled and that we have nothing left to learn. Although the assimilation of the Italians and their descendants generated anxieties at the time, with hindsight it appears to have gone relatively well. The question arises, therefore, of whether the present worries reflect an objectively greater difficulty, based on the specificity of the cultural and religious zones from which the new streams originate and on the hostility they arouse, or instead result from the questioning that any process of assimilation is bound to inspire among its contemporaries. An absence of anxiety may itself be indicative of the successful functioning of the French “creuset” or melting pot (Noiriel, 1988) into which millions of descendants of Italians, Spaniards, Russians, Poles, etc. have gone, but that is not a reason for not attempting to understand how integration worked for these migration streams, while there is still time. The calculations presented here show that a “window” still exists for studying the assimilation of these streams, or at least those from southern Europe, which remain dominant in the first generation of adults born in France. Conversely, this study is still premature, at least above age 40, for the streams that are the focus of media attention.
The example of unemployment gives an idea of the scale of the problems encountered by the French population of Maghrebin origin. Difficulties in entering and remaining in the labour market are a major obstacle for persons of Algerian origin, something already highlighted by the MGIS survey (Gaymu and Parant, 1996). These difficulties, remarkably persistent with age, are slightly less acute for men originating from another Maghreb country (Tunisia, Morocco). If the results presented here were confirmed it would mean two things. First, that it is a mistake to treat the populations of Maghrebin origin as forming a homogeneous whole. Second, while unemployment is singularly high for all men and women of Maghrebin origin, there is clearly a problem specific to the population of Algerian origin, which requires investigation: possible factors include lasting mutual hostility, unrealistic expectations, negative effects of disparaging representations, discriminatory practices, etc. (Silberman and Fournier, 1999).
But we should not let ourselves be blinded by the present difficulties, thereby running the risk of an unrelieved pessimism. The dynamics of integration must be appraised over the very long term, something which assumes that we develop the tools necessary for understanding the processes at work.
Institut National d’Études Démographiques, Paris. Translated by Godfrey I. Rogers.
That is, people who were French nationals in 1927 but who would have been foreigners had the law contained no provision for changing nationality.
The 1971 survey was on 2,693 persons and the questions were formulated thus: “Have you one or several foreigners among your great-grandparents? Your grandparents? Your parents?” Nationality, it has been established, is far less well known than country of origin, and one can still be a foreigner even though born in France. In 1971, 14.1% of French people aged 20 and over re ported having at least one foreign parent or grandparent, compared with 7.1% twenty years ear lier.
This survey was designed to explore the assimilation process for immigrants and the children of immigrants in France. It resulted in a report and numerous publications (Tribalat, 1995, 1996).
See also the survey’s bilingual French and English website: http:// www-ehf. ined. fr
The processing of this module received funding from the Délégation générale à la langue française et aux langues de France (DGLFLF).
See the results drawn from the EHF survey by Alain Monnier and Sophie Pennec: “Orphelins et orphelinage”, in the collective book Histoire familiales, histoires de famille, to be published in the Cahiers de l’INED (2004).
Italy, Spain, other EU country, other country of origin.
The information about the partner comes from the 1999 census. Male immigrants who have formed more than one union are excluded so as not to take into account any children living in the home of a male immigrant who were from a previous union of his second partner born in France. In all, missing from the total are the children of mothers born in France and immigrant fathers who have formed a second union, whether these children are of their father’s first or second union. But the approximation is not too wide of the mark.
According to the 1992 MGIS survey, among women aged 20-29, born in France to immigrant parents, the proportion living with an immigrant is 36% for women of Algerian origin, 14% for those of Spanish origin, and 29% for those of Portuguese origin. We have assumed the frequency of these unions to be identical to that of women of Spanish origin for women of Italian origin or from another EU country, and that for women of Portuguese origin to apply to women of other origins.
We include 50% of the children of Algerian, Portuguese and other origins, and 60% of those of Spanish, Italian and other EU country origins. We probably overestimate slightly the second generation born in France of Algerian origin. A proportion of one-third would reduce its numbers by 20,000.
This would require introducing an additional question, though this would not be straightforward. For Algeria, the question could be on the lines of: “Was your father Algerian in the present-day sense, even if at the time like all Algerians he had French nationality, or was he a descendant of those who settled in Algeria in the period of French colonization, or indeed even earlier for the Jews?”
An inferior status reserved for Jews and Christians under Sharia law.
A crisis between France and Tunisia who “waged war on each other for a few days”, the latter wanting to recover the air and sea base that France occupied “by virtue of the agree ments between the two countries” (Liling, 2003). Jews suspected of supporting the French feared for their safety and some fled to France.
Many left after the pogrom of the Six Day War. This chronology of the exile is quite closely mirrored in the acquisition of French nationality by Tunisians: 2,756 between 1956 and 1960, most of which were granted in Tunisia, 7,355 between 1961 and 1966 and 9,860 between 1968 and 1972, the majority of which were granted in France. A number of them had thus come to France when already French nationals.
Alba and Silberman have also developed a method for distinguishing “the children of pieds noirs [European settlers from Algeria] broadly defined” from the children of Maghrebin immigrants based on a number of socio-demographic characteristics. Their method is far from per fect since it risks giving extra weight to the very unfavourable social characteristics, thus casting the situation of the children of Maghrebin immigrants in a still more negative light. But they believe it preferable to making no distinction between the two populations, which are too different in many respects.
Also arriving in France at this time were the “French Muslims” (“Français musulmans”). In 1968, the last year in which INSEE used the “Français musulmans” category, they numbered 85,000, half of whom reported themselves as French by birth and half as naturalized French. The latter are thus counted with the immigrants.
These African countries were the Territoires d’Outre-Mer (TOM–French overseas ter ritories).
There are numerous variants but all contain a reference to Arabic or to Berber (Berber, Kabyle or Chaoui), or to a dialect (or “patois”) specified as Arabic, Berber, Algerian, etc.
60% for Algeria, 70% for Morocco, and 50% for sub-Saharan Africa.
The same proportion was applied uniformly to all ages.
This proportion is lower than the assumptions made for immigrants from Algeria and Morocco so as to allow for the presence of Tunisian Jews who would have spoken French with their children (see above).
The information about the mother tongue also proves useful in the case of the adult children of immigrants from Turkey. The 1920s saw an influx to France of large numbers of Armenians, traumatized by the genocide of 1915 and disappointed by the Treaty of Lausanne of 1923 which ended all hope of an independent state. Some immigrants born in Turkey, among the oldest, are in fact Armenians, probably numbering several thousand. The same applies, a fortiori, to the first generation born in France. Among the children of fathers and/or mothers born in Turkey, those spoken to in Armenian, exclusively or alternating with French, during childhood, represent slightly more than 20% of the total (just over 9,000 of 43,000). These persons belonging to the first generation born in France are elderly. It seems unwise to study the behaviour of Turkish-origin adults in the higher age bands on the assumption that they are of the same “ethnic” origin as the younger ones.
In particular the figure of six million put forward by Le Monde in an article about the new TV station set up by Beur FM (Beur TV): “It will seek to answer the expectations of the six million Maghrebins who live in France …” (Le Monde, 5 April 2003). [Tr. The word “Beur” originates in the Parisian back slang term for Arab].
We had already observed this phenomenon in the MGIS survey: two-thirds of the Spanish immigrants aged 20-59 had arrived before the age of 16.
For an estimate of the number of people living in France who could be Muslims, based on these figures, see Michèle Tribalat, “Le nombre de musulmans en France, qu’en sait-on?” in Yves-Charles Zarka (dir.), L’Islam en France, Special issue of the journal Cités, 2004.
A person can be the child both of a repatriate and of an Algerian immigrant. The number of such cases is assumed to be negligible. This assumption is less acceptable for the next generation.
A proportion close to that of young people aged 20-29 of Algerian origin in the MGIS survey.
Born in France to two parents born in France.
A shorthand term to denote adults born in France to immigrant parents from Algeria. Almost all are French nationals.
If the former represented not 40% but 30% of all children of immigrants, the unemployment rates, both male and female, for the children of repatriates would be close to those of French nationals of French descent. The children of immigrants could then well be represented by those of Arabic or Berber mother tongue.