Trajectories and Origins: Survey on the Diversity of the French Population

INED Population Studies 8

Cris Beauchemin · Christelle Hamel  Patrick Simon Editors

Trajectories and Origins: Survey on the Diversity of the French Population

INED Population Studies Volume 8

Series Editors Éric Brian Département de sciences sociales Ecole normale supérieure Centre Maurice-Halbwachs (CNRS_ENS_EHESS), Paris, France Jean-Marc Rohrbasser Institut national d’etudes démographiques (INED), Paris, France Editorial Advisory Board Isabelle Attané (INED), Didier Breton (University of Strasbourg), Olivia EkertJaffé (INED), Lionel Kesztenbaum (INED), Anne Lambert (INED), Cécile Lefèvre (University Paris-V), Godelieve Masuy-Stroobant (University of Louvain-laNeuve), Nadine Ouellette (INED), Arnaud Régnier-Loilier (INED), Claudine Sauvain-Dugerdil (University of Geneva).

This book series is devoted to publications of international relevance in population studies and demography as promoted by the French Institute for Demographic Studies (INED, Paris). As one social science among many, demography is interlinked with related disciplines including sociology, anthropology, history and linguistics, and continuously explores its boundaries with neighbouring disciplines, ranging from epidemiology and biology to economics. The studies published in the series are based on solid empirical research and firm methodological foundations. Particular attention is paid to long-term and collaborative surveys. Guided by its distinguished Editorial Advisory Board, the INED series aims to provide international visibility to works of high academic standard recognized in the French-­ speaking scientific community. The series supports an internationally acknowledged style of demographic research, championed by INED for more than half a century and rekindled in such fields as the study of demographic situations around the world; the relationship between demographic conditions and development; international comparisons; migration, identities and territories; family studies; gender studies and sexuality; ageing, health and mortality; trajectories, mobility and social networks. More information about this series at http://www.springer.com/series/11579

Cris Beauchemin  •  Christelle Hamel Patrick Simon Editors

Trajectories and Origins: Survey on the Diversity of the French Population

Editors Cris Beauchemin Institut national d’études démographiques (INED) Paris, France

Christelle Hamel Institut national d’études démographiques (INED) Paris, France

Patrick Simon Institut national d’études démographiques (INED) Paris, France

ISSN 2214-2452     ISSN 2214-2460 (electronic) INED Population Studies ISBN 978-3-319-76637-9    ISBN 978-3-319-76638-6 (eBook) https://doi.org/10.1007/978-3-319-76638-6 Library of Congress Control Number: 2018942529 © Springer International Publishing AG, part of Springer Nature 2018 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer International Publishing AG part of Springer Nature. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

The TeO survey was a major collective endeavour. We would like to thank all those who brought this venture to fruition, first and foremost François Héran, Director of INED when the survey was first conceived. He defended the project throughout the 4 years of survey preparation and implementation, offering scientific and political support at difficult times when the survey's legitimacy was brought into question. Chantal Cases, his successor, also provided her unstinting encouragement. She ensured optimal conditions for data analysis and followed the progress of our research with close attention. A large-scale data collection operation of this kind would have been impossible without the close collaboration of our colleagues at INSEE, both for project governance and for survey design and implementation. Stéfan Lollivier, Head of Demographic and Social Statistics at the time, took the necessary steps to include the survey in the INSEE work programme. Guy Desplanques, Head of the Demography Department, followed by Pascale Breuil, Head of the Demographic and Social Studies unit, supervised its design and implementation. Our colleagues at the Statistics and Immigration Studies unit, Catherine Borrel, unit head, Elisabeth Algava, then Bertrand Lhommeau, and the survey leaders, Jacqueline Perrin-Hayes and Pascal Germé, played a central role in all the survey phases, from its initial design to the release of data just 1 year

after collection. Cécile Ménard, followed by Pascale Pietri-Bessy, coordinated the operations at INSEE and the data collection process. Data collection was organized by the INSEE teams, and we are very grateful to the regional managers and the interviewers who deployed the questionnaire in the field and who recorded the slices of life that respondents kindly made available to us. The survey design, notably the content of the questionnaire, owes much to a multidisciplinary team of colleagues. Most of them were also members of the broader data analysis group. We extend our thanks to Maryline Bèque, Yaël Brinbaum, Martin Clément, Stéphanie Condon, Hugues Lagrange, Maud Lesné, Dominique Meurs, Laure Moguerou, Muriel Moisy, Mahrez Okba, Ariane Pailhé, Jean-Louis Pan-Ké-Shon, Jean-Luc Primon, Corinne Régnard, Mirna Safi, Emmanuelle Santelli and Vincent Tiberj for their respective contributions to our collective endeavour. The survey coordinators at INED received fantastic assistance in their task from Karine Wigdorowicz, Stéphane Bernard and Amélie Charruault, who also helped to prepare the data files and their documentation. This book is an abridged translation of the book Trajectoires et origines: enquête sur la diversité des populations en France published by Éditions de l’Ined in 2015 (Grandes Enquêtes book series). For the original publication, we wish to thank Jean-Marc Rohrbasser, head of the editorial committee, the various anonymous reviewers and the INED editorial team: Martine Rousso-Rossmann, Nicole Berthoux, Dominique Paris and Agnès Belbezet. The book was translated into English by Harriet Coleman, Paul Reeve and Brian Stacy. We wish to thank them for their hard work and for the excellent quality of their translations. Our thanks also to Catriona Dutreuilh who did an amazing job of coordinating the translation work and editing the final version. A survey of this kind would not have been possible without the financial support of numerous institutions interested in questions of integration and discrimination. Alongside INSEE and INED, they include ACSÉ,

AFPA, ANPE, ANR, DARES, DREES, HALDE, IAURIF-ÎdF and ONZUS.  Working together with scientific experts in the steering committee, the funding institutions were present throughout the project. This book is above all a tribute to the respondents who kindly agreed to share their experiences and trajectories, thereby illustrating the diversity of contemporary French society.

Contents

1 Introduction.............................................................................................. 1 Patrick Simon, Cris Beauchemin, and Christelle Hamel 2 Migration Histories and Socioeconomic Profiles.................................. 11 Cris Beauchemin, Bertrand Lhommeau, and Patrick Simon 3 Educational Trajectories and Transition to Employment of the Second Generation........................................................................ 39 Jean-Luc Primon, Yaël Brinbaum, and Laure Moguérou 4 Employment and Wages of Immigrants and Descendants of Immigrants: Measures of Inequality and Perceived Discrimination................................................................. 79 Dominique Meurs 5 Union Formation in a Multicultural Context........................................ 107 Christelle Hamel, Bertrand Lhommeau, Ariane Pailhé, and Emmanuelle Santelli 6 The Living Environment of Immigrants and Their Descendants: Perceived Discrimination and Segregation.................... 143 Jean-Louis Pan Ké Shon and Claire Scodellaro 7 Migration and Living Conditions: Their Impact on Health................ 171 Christelle Hamel and Muriel Moisy 8 Discrimination in France: Between Perception and Experience......... 195 Yaël Brinbaum, Mirna Safi, and Patrick Simon 9 The Place of Racism in the Study of Discrimination............................ 223 Christelle Hamel, Maud Lesné, and Jean-Luc Primon 10 Language Use and Family Transmission in Migration Context.......... 249 Stephanie Condon and Corinne Régnard

ix

x

Contents

11 Registers of Identity. The Relationships of Immigrants and their Descendants to French National Identity.............................. 277 Patrick Simon and Vincent Tiberj 12 Secularization or a Return to Religion? The Religiosity of Immigrants and Their Descendants................................................... 307 Patrick Simon and Vincent Tiberj 13 Transnational Links and Integration: Between Here and There........ 331 Cris Beauchemin, Hugues Lagrange, and Mirna Safi 14 Conclusion: Diversity of Origins and the Emergence of Minorities............................................................................................. 359 Cris Beauchemin, Christelle Hamel, and Patrick Simon  ppendix: Trajectories and Origins: Survey Methodology......................... 369 A Elisabeth Algava, Bertrand Lhommeau, and Cris Beauchemin

Preface

 cience by Exception, or How the TeO Survey S Fulfilled Its Mission Anyone in France who follows the news on political and social life with a minimum of attention will have heard of the Trajectories and Origins survey, subtitled “Survey on population diversity in France”. Carried out jointly by INED and INSEE in 2008 and 2009, the survey caused something of a stir on its launch (I will return to this below). Since then many articles, working papers, and dissertations have contributed to sustaining public interest in the survey and popularizing its acronym, TeO. But until 2015, no reference document had been published to bring together these publications, expand upon them and systematize their content. You are holding just such a document in your hands (or viewing it on your screen), thanks to the extensive team of collaborators led by Cris Beauchemin, Christelle Hamel, and Patrick Simon. Beyond INED and INSEE, the project leaders agreed from an early stage to provide open access to the survey datasets. They were thus able to include contributions from universities and research institutions, for whose participation I would like to express particular thanks. It must be emphasized from the outset that the TeO survey is a lone star in the universe of research and public statistics in France. How many surveys have explored in depth the lives and experiences of immigrants and their descendants in France, their integration in French society, and the successes and obstacles that have marked their trajectory over two generations, including experience of discrimination? The only comparable precedent is the geographical mobility and social integration survey (Mobilité géographique et insertion sociale, MGIS) carried out by Michèle Tribalat, an INED researcher, in 1992. This pioneering survey, performed in collaboration with INSEE, focused on a set of seven migration streams dating back to different periods of recent French history. It resulted in a set of important publications, and remains a reference today for anyone seeking to understand the process of migrant integration in France in the second half of the twentieth century. Its questionnaire not only retraced migrants’ personal trajectories, it also looked xi

xii

Preface

back at the previous generation. This was the very first time in France that the relevant supervisory bodies, the CNIS and the CNIL,1 gave authorization to collect information on the countries of birth of individuals and their parents, paving the way for the comparative study of what are now known as the “first generation” (migrants who came from elsewhere to reside in France) and the “second generation” (their children born or raised in France). As the MGIS survey covered several decades retrospectively, regular updating was not necessary. The addition of a few further years of observations would have made little difference to the overall pattern of results. But this argument does not justify the fact that 16 years were allowed to pass between MGIS and TeO. Why such a long delay? The first reason, and not the least among them, is the cost of such surveys, notably the cost of collecting data by means of a long and complex questionnaire. The objective was to record the trajectories of migrants and their descendants in their various dimensions: geographical, residential, familial, educational, occupational, religious, civic, not forgetting their social networks, cultural activities, and their own perceptions of their personal trajectory. Hence the need to ask hundreds of questions, through interpreters when needed, in face-to-face interviews that took over an hour to complete on average. The next factor was the cost of reaching the relevant minority populations. In France there is no population register providing a sampling frame that can be used to reach all descendants of migrants. The census includes questions on origins, including those of migrants who are naturalized French citizens, but they are limited to the respondents’ own generation, and do not cover their parents. To prepare for the TeO survey, the INSEE team thus had no choice but to trawl by hand through many thousands of birth certificates held by French municipalities, with authorization from judges, in order to identify the children of immigrants who might be eligible to enter the sample. It was no easy matter to obtain funding for the TeO survey. In the lively debate that surrounded its launch, one commentator – not particularly well-informed on the funding of public surveys, but convinced that self-interest runs the world – became convinced that the primary motivation of the project’s leaders was to grab all the available grant money, supposedly showered so generously upon studies of “diversity” in France. But nothing could be further from the truth. Neither INSEE nor INED had the means to cover the full costs of data collection. Heaven and earth had to be moved before adequate funds could be obtained from various ministries, supplemented by the decisive contributions of a number of agencies and authorities – to whom we are most grateful.2 Unable to supervise the content of the questionnaire, the Direction des populations et des migrations (Department for populations and 1  Respectively, the Conseil national de l’information statistique and the Commission nationale de l’informatique et des libertés (national bodies charged with personal data protection in France). 2  Direction de l’animation de la recherche, des études et des statistiques (DARES), Direction de la recherche, des études, de l’évaluation et des statistiques (DREES), l’Agence nationale pour la cohésion sociale et l’égalité des chances (ACSÉ), Association pour la formation professionnelle des adultes (AFPA) la Haute Autorité de lutte contre les discriminations et pour l’égalité (HALDE), Agence nationale de la recherche (ANR), l’Institut d’aménagement et d’urbanisme (IAU-ÎdF), Observatoire national des zones urbaines sensibles (ONZUS).

Preface

xiii

migrations) which at the time was being absorbed by the Ministry of Immigration, abstained from contributing to survey funding, leaving this to the DREES, the shared research and statistics unit of the Ministry of Health and the Ministry of Social Affairs. The hunt for the funds needed for the TeO survey involved exhausting efforts on the part of the project leaders, illustrating the chronic deficit of large-­ scale infrastructure that afflicts social science research in France. I would like to pay particular tribute to these young researchers, who spared no effort in initiating this venture against all odds, despite the risk that their publications would be delayed by several years. This brings us to one further factor behind the difficulty in launching a new survey on the integration of immigrants and their descendants. Fifteen years after the MGIS survey, a new generation of researchers was needed at INED, along with a new generation of statisticians at INSEE. And yet the initial number of personnel fell far short of what was required. As the director of INED since the late 1990s, I had the backing of the board of administration and the scientific council on this point. In 2002 they supported the creation of a research unit on international migration and minorities, and agreed to staff it through recruitment at the highest level, while in parallel INSEE increased the staff resources devoted to the question of migration. These renewed and rejuvenated teams valiantly led the TeO survey project. But the emergence of a new generation of researchers was evidently a lengthy process. Finally, the concern, and even hostility, aroused by the very idea of the TeO survey must also be mentioned. Its designers had constructed a questionnaire that explored migrants’ integration over time in the various facets of social life, but they also wanted to obtain as much detailed information as possible on experience of discrimination. How can different forms of discrimination be studied without taking into account the categories used by its perpetrators? Given the impossibility of questioning those perpetrators, the survey had to rely on the victims’ capacity to perceive discrimination and describe the experience. In France, it was out of the question for the TeO survey to apply the kind of “ethno-racial” framework used in censuses in the United States or the United Kingdom – a list of “races”, potentially subdivided into cultural affinities, with the respondent invited to tick one or more boxes. A first solution consisted in posing questions on various experiences of discrimination or unfair treatment, followed by a long list of possible grounds, including country of origin, surname, neighbourhood of residence, skin colour, accent, etc. These were maintained in the final version of the questionnaire, and detailed studies based on these data can be found in this volume. It is thus possible, based on the TeO survey data, to determine what proportion of persons who experienced discrimination attributed it to one or other factor, including physical appearance and skin colour. The reverse, however, is not possible. What is the probability that individuals will experience discrimination if they are unlucky enough to be perceived as black, Asian, North African, but also white, etc.? The data do not say. This is comparable to a scenario where the proportion of tourists among visitors to the Louvre is known (it is very high), without knowing the probability for a randomly selected

xiv

Preface

tourist of visiting the Louvre (which is much lower), nor indeed even having an approximation of the number of tourists in a year. Louis Schweitzer, who at the time was the head of the national authority in charge of combating discrimination (Haute autorité de lutte contre les discriminations et pour l’égalité, Halde), took a lively interest in the TeO survey, to the point of co-financing it. He spoke out in favour of a second solution, so-called hetero-­ perception. This begins with a question such as “Do you feel that you are perceived as black?”, followed by a second question such as “If yes, have you been discriminated against for this reason?” This general question would doubtless then have to be broken down by context: family, work, public space, dealings with the administration, etc. Several options of this kind were studied, but they met resolute opposition from the association SOS-Racisme: they saw the “Schweitzer solution”, as it came to be known, as nothing other than a “racial census”. In their view, under the cover of studying racial discrimination, the TeO researchers were in fact planning to practise discrimination themselves. The first chapter of the present volume recounts the special treatment to which the TeO survey was subjected in its preparatory phase of examination by the competent authorities: a double examination by the control bodies of the CNIS; an unauthorized leak of the questionnaire to the general public; an online petition launched by SOS-Racisme; and scathing opinion pieces in several daily newspapers either defending or condemning the famous “ethnic statistics” that TeO had come to symbolize. Some were even signed by INED researchers from outside the research unit on migrations. Not to mention the considerable reluctance of the Haut conseil à l’intégration, a body that has since ceased to exist, whose erstwhile president affirmed in an official letter that the TeO project was decidedly too focused on studying discrimination and not enough on integration. No previous INED or INSEE survey had ever kindled such fierce debate. A coincidence of timing fanned the flames. After the questionnaire had been approved by the CNIS and was about to enter its final test phase, an initiative taken in 2007 by parliamentarians who were also members of the CNIL began to make headlines. It consisted in introducing an amendment to article 63 of a proposed immigration law (known as the “loi Hortefeux”, after the Minister of Immigration at the time) aiming to facilitate CNIL scrutiny of “operations needed to conduct studies that measure the diversity of the origins, discrimination, and integration”, and specifying that such operations were to remain strictly anonymous and be subjected to a full examination procedure. This amendment was immediately interpreted by the media as a green light for “ethnic statistics”, of which the TeO survey was taken as the illustration par excellence. On the legal level, the amendment took the form of an exception that was to be added to the ten exceptions already set out in article 8 of the Data Protection Act (loi Informatique et libertés) of 6 January 1978, modified in August 2004. This point merits further explanation. Article 8 of the Data Protection Act, which sets out a general principle inspired by Article 1 of the French Constitution, is of key importance for statisticians and social science and health researchers: “The collection and processing of personal data that reveals, directly or indirectly, the racial and

Preface

xv

ethnic origins, the political, philosophical, religious opinions or trade union affiliation of persons, or which concern their health or sexual life, is prohibited.” Such data are labelled as “sensitive” by law (note in passing that income and assets are not included in this list...). This list is a singular one for those interested in research. Are there not countless surveys that draw on individual data in these areas? Does the press not regularly report on opinion polls and surveys that measure political, trade union, and religious affiliations? And what of surveys on health, disabilities, and sexual behaviour? If the 1978 law applies in the same way to private polling institutes and public institutions (INSEE, INED, INSERM, etc.), what, then, is the legal basis for the studies that regularly appear under the “Science” and “Society” headings of print and audiovisual media? The answer can be given in a single word: exceptions (dérogations in French legal language). The same article of law that forbids the collection and use of sensitive personal data then immediately lists no less than ten exceptions, themselves subject to variable conditions. The collection and use of such data is authorized if the protection of persons is guaranteed (written consent is obtained, data are anonymized, statistics compiled by INSEE under the control of the CNIS), if it serves certain ends (favouring the assertion of rights or progress in health research, an association studying its own members...), if it is authorized by a decree from the Conseil d’État, or if it is judged to be in the “public interest”. But the CNIL must always examine which of these exceptions applies, if any, on a case-by-case basis. Anonymity at source, for example, eliminates personally identifiable information from data and takes them out of the scope of the Data Protection Act, but it is again up to the CNIL to determine the technical conditions under which the data collected can be “depersonalized” (as for example in the case of INSERM and INED telephone surveys on sexual behaviours). The law thus responds in the affirmative to the crucial question of whether collecting and using “sensitive data” is permitted. But instead of saying: “It is allowed, as long as...”, it states that “It is prohibited, unless....” The general principle of prohibition remains, although it is made more flexible through a range of duly controlled exceptions. The CNIL amendment to article 63 of the Hortefeux immigration law thus proposed an additional exception in favour of studies on the diversity of origins and the extent of discrimination. But the Conseil constitutionnel, in a decision issued on 15 November 2007, declared the amendment unconstitutional, on the grounds that it was a “legislative rider” that was out of place in a law on immigration controls. Would it have been found acceptable in another, more relevant law? Looking back in hindsight some years later (and this type of analysis cannot be rushed), I strongly doubt it. The amendment’s sin was not the desire to introduce an eleventh exception to the general prohibition, but the fact that this exception was a thematic one: it was supposed to cover all studies on the diversity of origins, instead of being restricted to an exception of a technical or procedural nature (bearing on privacy protection guarantees). It is in this respect that, rider or not, the amendment proved to be contrary to Article 1 of the Constitution, as the Conseil constitutionnel mentioned in a commentary on the same decision (obiter dictum, in legal terms). An exception cannot suspend a general principle; it must remain an exception. The amendment

xvi

Preface

sought by the CNIL gave the impression that any and all surveys on “the diversity of origins” could thus automatically be classified as in the public interest, whereas the spirit of the law implied that the fulfilment of this condition should be examined on a case-by-case basis. Needless to say, neither INED nor INSEE had called for such an exception, nor had the designers of the TeO survey. The application of the standard procedures of the CNIL was largely sufficient. Here I must draw the reader’s attention to a fundamental point that, to my knowledge, has never before been raised. Most of our socio-demographic knowledge on changes in French society and questions of public health has been obtained by way of derogation. This is true of studies on the evolution of social mores and family structures, and it is equally true of studies on migrants’ origins and the impact of these origins on their interactions with the host society. As INED researchers gain knowledge of social mores, in the old sense of “moral and political sciences”, they refine their questions, explore individual biographies in greater detail, and trace back histories over several generations. In doing so they are constantly navigating the boundaries between the public, the private, and the intimate.3 The list of sensitive – sometimes extremely sensitive – topics that INED has explored in its surveys over the last 15 years is long: non-marital cohabitation, children born outside marriage, medically assisted procreation, abortion, sexual behaviour, genital mutilation, sexual dysfunction, family violence, disability, adoption, homelessness, end-of-life medical decisions, etc. How have such surveys been possible? Only by exception. In this respect, there is nothing unusual about the study of immigrant populations. Of course, the sensitive or intrusive character of an issue can change over time. Questions on non-marital cohabitation were still sensitive in 1980, but became commonplace over the following years. In the first years of the PACS civil partnership, compiling statistics on the sex of the partners in civil unions was prohibited by law, until homosexual organizations themselves requested that this secrecy be lifted. I could give many more examples. In the eyes of young researchers today, the questionnaires of the 1980s seem timid, as those of the 1950s were for my generation. Will the same not be true a decade from now, in 2025, to those rereading the passionate debates of the late 2000s on the TeO survey? Will the suspicion that certain questions on origins and appearances seek to “undermine the foundations of the Republic” still be understood, when their modest aim was to better capture the mechanisms underlying the various forms of discrimination that undermine the principle of equality? More than a decade has passed since the polemics stirred by the TeO survey, and already the lava has cooled on at least one theme: questions about religion. Who remembers the fiery accusations spurred by these questions, on the same grounds as (and perhaps to an even greater extent than) ethnic origins and physical appearance? In an official letter to the CNIS on the TeO survey, SOS-Racisme issued this absolute accusation: “Apparently, INED and INSEE wish once again to verify that 3  This line of thought was already broached in the preface to the INED volume on the Family history survey associated with the 1999 census (Histoires de familles, histoires familiales. Les résultats de l’enquête Famille de 1999, Paris, Éditions de l’Ined, coll. « Les Cahiers », 156).

Preface

xvii

a­ nti-­Semitic adage, ‘Jews are rich’.” And further down, this definitive statement: “Presuming to understand the contribution of religious influence to the behaviour of individuals is totally unacceptable.” This casts opprobrium on the very principle of a sociology of religion, already present in 1897 in Émile Durkheim’s universal classic Suicide, and practised in countless research centres around the world. Who can deny the influence of the religious factor on behaviours as varied as non-marital cohabitation, fertility, abortion, divorce, and end-of-life care? There is no major stage in the life cycle that is not deeply affected by individuals’ relationship to religion. What demographer or sociologist would dare to forbid research of this type, risking exclusion from the entire scientific community? Before the data had been examined, how could anyone dictate that religion has no effect on the integration of migrants and their descendants in different spheres of social life in France? While Article 1 of the constitution may sweepingly condemn legal distinctions by “origin, race, or religion”, not a single supervisory body, be it the CNIS, the CNIL, or even the Conseil constitutionnel in its decision of November 2007, casts doubt on the legitimacy of questions about religion in public statistical surveys. It is abundantly clear that scientific analysis of the religious factor in an anonymous survey cannot be construed as a form of unequal treatment granting rights to certain individuals or withdrawing them from others on the basis of their religion; and neither can it be seen as an attempt to label individuals on the basis of their religious affiliation. The antiracist group’s demand that questions on religion in the TeO survey be prohibited found no support. The present volume thus presents invaluable in-depth analyses of discrimination by religion, changes in religious references from one generation to the next, and levels of religious endogamy among practitioners of different religions, be they Catholic, Protestant, Jewish, Buddhist, or Muslim. The classical argument holding that the relationship to religion is a shifting reality that resists all measurement also does not hold water: all the realities of this earthly existence are shifting ones, beginning with social realities. The response to this ancient problem does not consist in prohibiting statistics, but in diversifying them to better capture behaviours in all their complexity. This is why the TeO survey distinguishes between questions that reflect a simple sense of affiliation, those that speak of levels of practice, and those that establish a strong attachment to religion. It bears repeating, although it contradicts a widespread belief: public statistics and public research have the right to include direct questions on the religion of respondents, their spouses, and their parents, on the condition that the scientific purpose of the survey justifies this inclusion and is formally recognized. From 1980 to 2008, INSEE and INED had to content themselves with a question on respondents’ religious affiliation that did not specify the religion in question (the distinction drawn was simply: “neither sense of affiliation nor practice”, “sense of affiliation only”, “both”, “do not know or prefer not to respond”). Since then, at least three public statistical surveys have included direct questions on the religion of respondents and their families. Besides the TeO survey, there have also been the successive waves of the ERFI survey (the French version of the International Generations and Gender Survey, carried out in 2005, 2008, and 2011 in 15 European countries), as well as the MFV survey (Migration, families

xviii

Preface

and ageing in the overseas départements, 2009–2010). These surveys were all carried out jointly by INSEE and INED, with the approval of supervisory bodies. They gave rise to publications that are accessible online. None of them has ever led to the slightest stigmatization of populations who believe in or practise the religions concerned. As the years go by and the dust begins to settle, there is no doubt that the TeO project stands out for its novelty and for its exceptional results. To critiques of principle, it responded with a demonstration in practice. Its leaders decided to show that the proof of the pudding was in the eating. It is now clearly established that social science research can expand our knowledge of discrimination and its mechanisms without the slightest threat to individual freedoms, and without any damage to minorities. More precisely, TeO demonstrates – if a demonstration were needed – that the technique of “proxies” or substitutes, whereby a problem is approached by talking about something else, has severe limitations. Country of origin cannot be equated with religion, any more than religion with physical appearance, or language with country of origin. These factors can act cumulatively or interact, but they only partly overlap. To take just one example, religious discrimination does indeed exist, but quite separately from discrimination by origin. It has now been demonstrated, through the systematic use of logistic regression as a tool for modelling and differentiation, that for a given duration of residence, generation, age and level of education, migrants and their descendants still encounter difficulties of highly unequal scale in their integration into French society, depending on their origin and on the relationship of the native-born population to that origin group. Taking the redundancy of these different variables as axiomatic and excluding one or another of them ex-ante is not a scientific approach. In good science, the only valid criterion for introducing a variable into a model, or excluding it, is its added explanatory value, empirically attested. Opponents of a model are free to refute it, but it is up to them to demonstrate that the observed differences can be better explained with other variables. In the same line, the designers of the TeO survey understood the importance of creating control samples for purposes of comparison, i.e. French citizens born in France to parents also born in France, but also persons born French in the overseas départements, or born in metropolitan France to parents born overseas. Complementary questions were used to carefully distinguish between repatriates and migrants (as well as their respective descendants). Here again, these methodological distinctions, far from dismembering our one and indivisible Republic, have treated it with total respect, working to measure the gap that separates the reality from the ideal. TeO is the first survey to compare the experiences of discrimination faced by members of different populations who, in the eyes of those practising discrimination, share a common visible trait: the colour of their skin; some because they descend from African populations reduced to slavery in the Antilles and long since legally included in the French nation, others because they came to France from sub-Saharan Africa. The results confirm that while a family and community history of several generations of French citizenship may guarantee access to jobs in the civil service, it does not protect individuals from racial discrimination.

Preface

xix

The authors do not evade any of the difficulties that the second generation faces on the labour market, which are often greater than those of their parents when they arrived in France in a different economic context. Reinforcing the findings of Sciences Po’s 2005 CEVIPOF survey by Sylvain Brouard and Vincent Tiberj, they confirm that the children of immigrants from the Maghreb and Turkey have tended to turn to Islam, and that this religious revival is not unrelated to frustrations arising from experience of discrimination. On a question as crucial as the unequal school outcomes of girls and boys, as measured notably by the probability of leaving school with no qualifications, they reveal that even after controlling for a series of socio-­ economic and sociolinguistic factors, large differences by country of origin remain. All other things being equal, school failure is less common among girls from families with origins in the Maghreb or sub-Saharan Africa than among boys from these families, whereas the opposite is the case among families of Turkish origin, who sometimes appear to deliberately curtail girls’ education. Here again, these differences highlight a need for further research, both qualitative and quantitative. Many more examples could be given. Generally speaking, the TeO survey reminds us that discrimination is not merely a postulate – it can be empirically demonstrated. Its mechanisms need not merely be guessed at – they can be measured. Nor can it be reduced to vague representations. Much to the contrary, it is crucial to study the relationships between subjectively perceived discrimination and the objective experience reflected in life trajectories, with their histories of success and failure, progress and marginalization. The TeO survey makes this possible. No doubt the first results presented here point to the need for further, more detailed studies, but we have already learnt a major lesson: the subjective and objective dimensions of discrimination are closely linked. At this stage, one last clarification is needed. Indeed, perhaps I should have begun here. In collecting sensitive data on origins, four levels of knowledge and practice can be distinguished. In the nominative databases of administrations and businesses, recording data on origins and religion is prohibited: no exception is possible under current jurisprudence. The second level is currently that of the population census, which since the nineteenth century has included a question on respondents’ country of birth and previous nationality, even if they have since acquired French citizenship. The third level goes back one generation, asking about country of birth and the previous nationality of the respondent’s parents. This is the case, for example, of the family history survey (EHF), associated to the 1999 census and, since the years 2003–2004, of INSEE’s major surveys (labour force, family and housing, living conditions, etc.). At last we are able to measure access to employment, housing, and social and career mobility among the descendants of immigrants. The fourth and final level is reserved for specialized research surveys on sensitive topics. This includes the study of discrimination by origin, which calls for the description of characteristics linked to discrimination, including physical appearance, provided that strict technical and legal guarantees are set in place. Surveys in this category, which includes TeO, are extremely rare, but in my view they should be carried out every decade, or even every 5 years. Such surveys must address a

xx

Preface

strong social demand: in the case of TeO, the need for international knowledge and comparisons on the dynamics of integration and the scale of discrimination. I do not consider, however, that it should fall upon the researchers themselves – be it with the best possible intentions – to demonstrate that their work will necessarily have a “positive impact” on the fight against discrimination or, for example, on the success of integration programmes. No one expects a survey on incomes to demonstrate that it will improve their distribution. Conversely, it is impossible to prevent statistical studies from being occasionally misused by the ignorant or the ill-intentioned. The work of refutation in these cases must take place in the public arena. The researcher’s role is not to reform society but to methodically inform social actors through new knowledge, as objectively as possible. The designers of the TeO survey and the researchers who have explored the resulting data have fulfilled this mission perfectly. Now is the time for other social actors to take hold of the results, for competing researchers to do better, and for the responsible authorities to ensure that the means are made available for future surveys on the same topic. Institut national d'études démographiques (INED) Paris, France

François Héran

Chapter 1

Introduction Patrick Simon, Cris Beauchemin, and Christelle Hamel

1.1  Genesis of a Survey Even before it was a project, the Trajectories and Origins (TeO) survey was a clear necessity. France, a country of immigration throughout the twentieth century, has become a multicultural society, and the diversity of origins represented in its population is unprecedented. And yet little was known about the situation of populations with an immigrant background. In public debate, they are portrayed through false representations and stereotypes. With industrial restructuring, immigrants, initially seen as useful when the French economy needed low-skilled labour, came to be treated as undesirables. Economic crisis and endemic mass unemployment undermine the legitimacy of their presence in the country. Although born and raised in France, their children too are sometimes seen as separate from the national community. Between the Marches for Equality of the years 1983 and 1984 and the riots of November 2005, the “second generation” question became a running theme in the mass media (Lapeyronnie 1987; Hajjat 2013; Beaud and Masclet 2006). The list of anxieties and fears is long; living in segregated neighbourhoods, failing at school, and lacking national identity, the children of immigrants, it was argued, would be tempted by communautarisme.1 Political leaders have diagnosed a crisis of the French “model of integration”, while the descendants of immigrants demand equal rights and denounce the discrimination that they face. And indeed, research on 1  The meaning of communautarisme does not translate easily in English since it is a mix of sociological and political terms. It refers to a propensity of the members of a group (be it ethnic, racial, sexual, social class, etc) to favor their membership of the group over other identities and participation, leading to a cluster of social interactions within the in-group and reducing any interactions with the out-group.

P. Simon (*) · C. Beauchemin · C. Hamel Institut national d’études démographiques (INED), Paris, France e-mail: [email protected]; [email protected]; [email protected] © Springer International Publishing AG, part of Springer Nature 2018 C. Beauchemin et al. (eds.), Trajectories and Origins: Survey on the Diversity of the French Population, INED Population Studies 8, https://doi.org/10.1007/978-3-319-76638-6_1

1

2

P. Simon et al.

discrimination has demonstrated that citizenship does not protect French nationals of immigrant background from unequal treatment based on their origins (Simon 2007). What should we think about this? Does the problem ultimately lie in the behaviours and strategies of individuals with an immigrant background, or in the way that French society responds to its growing population diversity? It is easy to imagine that the alternatives are not limited to this Manichean dichotomy, and that the realities of the trajectories, positions, and practices of immigrants and their descendants are much more complex (Guénif, 2006). Although the difficulties are real and should not be minimized, high achievers exist alongside school dropouts; immigrants’ levels of qualification have increased considerably in the last three decades; spatial concentration is not necessarily a sign of geographical exclusion; and representations of community isolation are at odds with the evidence that the social worlds of immigrants and their descendants are increasingly mixed (Santelli 2007; Safi 2006; Pan Ké Shon and Verdugo 2014). However, these findings are patchy, often published in monographs or community studies at neighbourhood or city level. Statistical data collected on a large scale and providing detailed knowledge of different origin groups are often lacking. While knowledge gaps on immigrant populations have been filled since the early 1990s, the situation of the descendants of immigrants in French society has not been captured in census data or in most public statistical surveys. There has been progress in the last decade, notably with the addition of questions on parents’ country of birth and nationality in the Labour Force survey (Enquête emploi), but that survey does not cover many areas of social life. What are the family, residential, and employment trajectories of migrants after arriving in France? To what degree are the milestones in their trajectories dictated by choice or by constraint? How strongly are their trajectories still influenced by the cultural and material assets they possessed – or did not possess – before arriving in France? Once length of stay and socioeconomic characteristics have been taken into account, do we still find significant differences in the processes of integration of migrants from different origins? These questions regarding “immigrants’ outcomes” take different forms for their descendants, who were born and socialized in France. What role do the origins of the second generation play in their trajectories, and how do these differ from the trajectories of the descendants of French families with a comparable social background? Are familial and marital behaviours and social practices reproduced across generations, or are substantial changes observed? Does intergenerational transmission vary widely across origin groups? In launching the TeO survey in late 2004, the aim of INED and INSEE was to answer these and many other questions. Funded by various ministries and public institutions (DARES, DREES, ACSÉ, AFPA, HALDE, ANR, IAU-ÎdF and ONZUS), the survey was led by a joint INED-INSEE team, along with a multidisciplinary group of 15 researchers and academics from diverse institutions and research centres. The list of contributors to this volume reflects that diversity, as most worked first on the survey design and then on analysis of the data. As we mark the close of the adventure that began in 2006 with the development of the

1 Introduction

3

q­ uestionnaire, followed by the field survey between autumn 2008 and February 2009, it is important stress the collective nature of this undertaking. The TeO survey was a necessity: society at large, the scientific community and the public authorities were voicing an urgent need for reference data on the situation of immigrants and their descendants. Our aim was to respond to these demands while developing an independent research programme. In short, it was up to us to define the ambitions and content of a survey that would simultaneously serve the public interest and contribute to the advancement of scientific research. There are many ways to tackle questions of integration and discrimination in surveys, and they are widely debated, not only in the political arena, but also among social scientists (Lorcerie 1994; Blum 1998; Fassin and Simon 2008; Rea and Tripier 2010). Let us recall the main research question to be addressed by the survey, as described at its inception: “The survey will investigate to what extent origin, as such, is a factor of inequality, or simply of specificity in access to the various resources of social life (housing, language and education, employment, leisure, public services and social benefits, contraception, healthcare, nationality, social networks, marriage market, etc.). It will investigate the link between origin and other status categories in French society (gender, class, phenotype, age, neighbourhood...) in order to analyse processes of integration, discrimination, and identity construction in French society as a whole”. Stemming from the political and social debates of the 1990s and 2000s, the issue of integration became a central concern in both research and politics. This was highlighted by the previous survey of immigrants and their descendants carried out by INED and INSEE in 1992 – the geographical mobility and social integration survey (Mobilité géographique et integration sociale, MGIS) – which examined immigrants’ assimilation and concluded that the French republican model of integration was working well (Tribalat 1995). However, integration has become a cliché, and the contribution of this concept, in terms of knowledge, has been devalued as its political charge has increased. The rising awareness of discrimination and its recognition on the political agenda has fostered a re-orientation of scientific and social interest towards society and its institutions (Fassin 2002; De Rudder et al. 2000). It was in this context that the TeO survey was launched, marking a new step in quantitative research on populations of immigrants and their descendants. The project explicitly called for a reformulation of research questions around issues of discrimination. The survey, and the analyses built upon it, examine integration over time by studying trajectories and comparing successive generations in terms of living conditions and access to resources, but steer away from a normative approach to integration, wherein the practices and behaviours of immigrants and their descendants are expected to converge toward a reference represented by the mainstream population, itself understood as uniform. The project included two strands of research: first, a resource-based approach; second, an examination of different forms of discrimination and their consequences for identity construction and practices. The objective was to describe and analyse levels and modes of access to the various resources of social life (language and education, employment, housing and place of residence, public services and social

4

P. Simon et al.

benefits, nationality, healthcare services, contraception, etc.). In studying levels of access to resources, we sought to identify obstacles and inequalities arising over the life course, and to determine the respective roles of individual or collective strategies and of structural constraints (family and social background, type of housing, spatial segregation, income, etc.). The trajectories and practices of immigrants and their descendants follow two types of processes across French society: differentiation and singularization, on the one hand, and convergence and cohesion, on the other. This approach served as a guiding thread, first for questionnaire design, and then for data analysis. Below, in brief, are some of the main features of our research strategy, as set out over the course of this volume. –– Although the principal goal was to obtain information on immigrants and their descendants, the survey covered the whole population. Individuals with no immigrant parentage in the last two generations responded to most questions, including those on migration trajectories, transnational practices, experience of discrimination, and forms of national belonging. We were thus able to compare differences by origin in the practices, experiences and trajectories of immigrants and their descendants, and of the mainstream population. This was truly a survey on the diversity of all populations in French society. –– Questions about transmission and reproduction from one generation to the next were central to the project. By comparing the positions of immigrants and the trajectories of their descendants, born and socialized in France, we could document the “unchaining of the generations” (Attias-Donfut and Wolff 2009) and its variations across ethnic backgrounds. Analysing the trajectories of descendants of immigrants provides a crucial means to assess French society’s ability to ensure equality of opportunity and reach cohesion with diversity of origins. We are thus able to observe the invisible boundaries within society, revealed by the filtering and selection of individuals based on their origin. –– The survey had two complementary ambitions, sometimes difficult to reconcile: to collect information on wide-ranging areas of social life on small targeted groups, while at the same time focusing on more specialized research themes. The compromise consisted in remaining superficial on certain topics, while going into detail on others. We gave particular priority to the study of migration trajectories, union formation, and the description of educational trajectories and current employment. Experience of discrimination was also a central component of the survey. The questionnaire included questions on discrimination in every area of social life where it can occur (in education, during job recruitment and in the workplace, in healthcare, housing, access to services), and a specific dedicated module that also explores the experience of racism. The comparative approach adopted for the analysis examines similarities and differences across generations and between the various population groups that form “French diversity”. Alongside immigrants and their descendants, we also identified individuals from the overseas départements (DOMs) and their children born in metropolitan France as a group of interest. In their relationship to migration and discrimination, the situation of these French citizens is distinct from that of others with

1 Introduction

5

several generations of French parentage, i.e. persons born as French citizens to French parents. This last group formed by persons with no immigrant parentage in the last two generations remains difficult to name, as it is constructed by contrast with minority groups. We were not satisfied with the terms “control group” or “reference group”. We ultimately chose the term “mainstream population” to refer to its position both from a demographic viewpoint and in relation to French social stratification.2 The knowledge obtained through the survey has been disseminated not only via the publications of the survey analysis team, but also by making the data available to the scientific community. We adopted a very proactive strategy on this point, making the survey datasets available via the Quetelet web plateform a year after the end of data collection, in February 2010, enabling the research community to work on the data in parallel with the survey analysis team. This strategy proved effective: many of the publications based on TeO data have been authored by researchers external to the project.

1.2  An Innovative Survey The first challenge for the survey was to find a methodology that would enable us to reach populations that are difficult to identify in the usual sources of household surveys. It was important to cover all groups making up the population residing in metropolitan France, to capture all trajectories and experiences, both their singularities and their similarities, in relation to processes of integration and discrimination. This included immigrants, descendants of immigrants who were born in France, persons from the overseas départements, their children born in metropolitan France, and French native-borns with native-born French parents – the “mainstream population”. In total, nearly 22,000 questionnaires were administered to cover these various sub-samples, with substantial over-representation of several statistically rare ethnic groups (for example, immigrants from Southeast Asia, Turkey or sub-­Saharan Africa and their descendants) to ensure a statistically reliable sample size. Moreover, in the absence of any ready-to-use sampling frame to select individuals born in France to immigrant parents, the constitution of a sample of descendants of immigrants was a considerable feat. The sampling method is described in a note at the end of this volume. The questionnaire was administered by INSEE and its network of interviewers (566 interviewers in 18 Regional Divisions) from September 2008 to February 2009. Despite the respondents’ high levels of mobility, which called for ingenious address recovery strategies, the response rate was 61%, which is high for this type of individual survey. Direct refusals were relatively rare, as the survey was well received by the persons contacted by the interviewers. In addition to residential mobility, linguistic issues of non-native French speaking immigrants and the  See the lexicon in Chap. 2 of this book.

2

6

P. Simon et al.

c­ oncentration of respondents in disadvantaged neighbourhoods, where data collection is generally more challenging, were also problems specific to this survey. At the end of data collection process, 21,800 interviews had been conducted, broken down as follows: –– –– –– –– ––

8300 immigrants, 86% of the initial target (9600); 8200 descendants of immigrants, 85% of the initial target (9600); 700 DOM native-borns, 88% of the initial target (800); 700 descendants of DOM native-borns, 88% of the initial target (800); 3900 persons from the “mainstream population”.

Two other operations connected to the main survey were also implemented in parallel, to flesh out the data available from the TeO project: –– A postal survey of young people, comprising an 8-page self-administered questionnaire on family transmission and parent-child relations to be completed by all offspring aged 18–24 in the target household. Out of the 6163 questionnaires sent out, 3353 were returned to INSEE (a return rate of 54%) and data capture was performed by an external service provider. Analyses based on this distinct database are not presented in this volume, and have been published elsewhere (Moguérou and Santelli 2013; Moguérou et al. 2013). –– Qualitative post-surveys conducted with a subset of 1000 respondents by 19 teams selected by competitive tender, on a varied set of themes, for more in-­ depth examination of the survey questionnaire results. Some projects have been presented in separate publications.3 These post-survey interviews are not covered in this volume. Finally, communication has been a central concern from the outset. A survey such as this one cannot be carried out without explaining its ins and outs, offering reassurance about its objectives, and involving civil society organizations, wherever possible, with its preparation. A website was created to post the survey documents online (http://teo.site.ined.fr/); a short English version is also available. The questionnaire was also posted online and translated into English. A forum of associations was created, and the survey was presented at a one-day event focused on gauging reactions to the project as a whole. These constant contacts with civil society and the research community were also useful in defusing the various controversies that arose during survey implementation.

1.3  A Survey Under Close Scrutiny Given the sensitive topics addressed in the questionnaire, and the sampling by ethnic origin of the populations surveyed, the survey was a delicate operation. Moreover, although TeO was a scientific survey, the fact that it was co-headed by  See the list of publications on the TeO survey website: teo.site.ined.fr

3

1 Introduction

7

INSEE and that the sample was derived from the census gave it a special status. Its design and implementation were subject to close – and, in many respects, exceptional – scrutiny, on the part of both the institutions supporting the project and the bodies charged with verifying and certifying scientific quality and compliance with the ethical principles governing statistical surveys. Its status as a public statistical survey conferred many advantages, but limited the development of methodological innovations on sensitive questions. For a survey on populations with an immigrant background, and with a focus on themes of integration and discrimination, the balance is sometimes difficult to achieve. As aptly noted by the Constitutional Council in the comment relative to its decision of 15 November 20074 on the conduct of studies measuring diversity of origins, discrimination, and integration: “In the realm of statistics, not everything is possible”. Given its objectives, the TeO survey found itself at the centre of a wider societal debate that accompanied its implementation and that had a non-negligible impact on its content. Here we will briefly review these episodes. The standard pathway of public statistical surveys on demographic questions begins at the CNIS (National council for statistical information), where the committee on demography and living conditions issues an opinion on its appropriateness (avis d’opportunité). This opinion, formed on the basis of a grounded presentation of the survey objectives, must ensure that it corresponds to the public interest, without duplicating the work of previous surveys. Once this opinion has been issued, surveys are evaluated by the Comité du label, another CNIS body, which examines the questionnaire methodology and content in detail, before ultimately issuing the “label of statistical quality”. In parallel, a file is submitted to the French data protection agency, CNIL, which rules on respect for privacy and the conditions for collecting sensitive information (ethnic or racial origins, religion, health status, political convictions, sexual orientation, etc.). The TeO project took a significantly different path. The CNIS issued an initial favourable opinion on 15 May 2006. But, breaking with the usual procedure, other sessions of the CNIS were scheduled to further discuss the survey project. An extraordinary supplementary session was held on 24 May 2007 to examine the sensitive questions contained in the questionnaire (foreigners’ residence permits, questions on religion, ethnic origin, and skin colour). At the end of this session, as the questions on respondents’ skin colour and religion had generated debate, another session was scheduled with the stated objective of “obtaining a consensus” on these questions. In the meantime, on 31 May 2007, the Comité du label validated the survey protocol and questionnaire, issuing an opinion confirming the general interest and statistical quality of the survey. However, this opinion was subject to the “condition that the CNIS subsequently issue a favourable opinion on the sensitive questions in the questionnaire”. The CNIS session of 12 October 2007 was thus entirely devoted to examining the TeO survey. At the end of a lively debate, the questions on religion and identity – including those on skin colour – were validated, accompanied by restrictive conditions on dissemination of the corresponding data.  Decision no. 2007–557 DC.

4

8

P. Simon et al.

Access to certain variables on religion and political orientation would thus be subject to a specific procedure managed by the CNIS’s committee on statistical confidentiality. The first pilot survey, scheduled for late November 2007, thus included two questions on skin colour. They began with the way the respondent was seen by others: “When someone meets you, what colour(s) do you think they see you as?” The response was open, with no predefined categories. It was followed by a self-­ identification: “And what colour(s) would you say that you are?” – again with an open response. But while the survey questionnaire had ultimately been approved after multiple examinations by the CNIS, the questions on the respondents’ skin colour was considered incompatible with Constitutional Council decision of 15 November 2007, recalling that the “the processing required to conduct studies measuring diversity of origins, discrimination, and integration [...] may bear on objective data but may not, in light of the principle set out in article of 1 of the Constitution, be based on ethnic origin or race.” In the wake of this decision, INSEE and INED removed these two questions from the final questionnaire.5 The removal of the questions on skin colour may retrospectively be seen as an episode with no substantial impact on the survey data analysis. Nevertheless, the fact that we were unable to confront this aspect of identity and discrimination in a scientific survey – albeit one carried out under the aegis of the French public statistical service – raises questions about freedom of research on these themes, and about the persistent confusion between an ethno-racial frame of reference that might be used by administrations, on the one hand, and, on the other, categories constructed by researchers for purely statistical purposes in order to advance knowledge on these topics. Beyond these specific questions, the procedure itself  – which was exceptional in various respects  – and the debates surrounding the conduct of the survey demonstrate the extreme sensitivity of the topics that it covers. On several levels, the TeO survey took place under close scrutiny.

1.4  The Contents of this Book This volume is a translation of selected chapters from the French book presenting the TeO findings.6 Compared to the French version with its 19 chapters, this English volume contains 13 chapters and a shorter version of the methodological note. We have merged two chapters on education into one, and have done the same with two chapters on employment. The chapters on migration trajectories, fertility histories, residential decohabitation, and the measurement of discrimination in the survey were not included. We wanted to cover the main themes of the questionnaire in this 5  For a discussion on the consequences of the decision of the Constitutional Council and the associated comments, see the COMEDD report (Héran et al. 2010). 6  Beauchemin Cris, Hamel Christelle, Simon Patrick (2015) Trajectoires et origines. Enquête sur la diversité des populations en France, Paris, INED, 607 p. (Grandes Enquêtes)

1 Introduction

9

volume, in order to offer a complete overview of the situation of immigrants and their descendants in contemporary French society. The authors of the various chapters worked together for several years and developed a shared analytical framework that was applied to the data. The definition of population groups, age groups, and family characteristics are standardized across chapters; the variables describing the phenomena under study, such as transnationalism, employment, education, migration, and discrimination are present in multiple chapters, and for the most part are constructed in the same way throughout. This standardization of conceptual tools ensures the overall consistency of this volume despite the large number of contributors. We presented descriptive analyses of the initial results in a previous publication (Beauchemin et al. 2010). The objective of the present volume is thus to deepen our analyses in the light of existing knowledge on the questions examined in the survey. Chapter 2 looks at the sociodemographic profile and social characteristics of the population groups studied in this volume. Chapters 3, 4, 5 and 6 focus on the resources of migrants and their descendants: education, work, health and housing. The accumulation of human capital may open pathways to mobility for migrants, while social capital that descendants inherit from their families is transformed through access to the resources available in French society. However, the benefits of these resources are modulated by differential access to education, the labour market, and housing. Chapter 7 addresses the dynamics of family formation of immigrants and their descendants, their reproduction or transformation. After identifying the obstacles linked to origins in different areas of social life, the last chapters look at experiences of discrimination and racism from the point of view of the respondents, and address topics that are central to the experiences of minorities, looking at identities, citizenship, civic and political participation, and relationships to religion in a society that is already highly secularized. These combined approaches offer many lessons, which we summarize in the concluding chapter.

References Attias-Donfut, C., & Wolff, F.-C. (2009). Le destin des enfants d’immigrés. Un désenchaînement des générations. Paris: Stock. Beauchemin C., Hamel C., & Simon P. (Eds.). (2010). Trajectories and origins. Survey on population diversity in France. Initial findings (Document de travail 168). Paris: INED. Beaud, S., & Masclet, O. (2006). Des “marcheurs” de 1983 aux “émeutiers” de 2005. Deux générations sociales d’enfants d’immigrés. Annales, Histoire, Sciences Sociales, 4, 809–843. Blum, A. (1998). Comment décrire les immigrés? À propos de quelques recherches sur l’immigration. Population, 53(3), 569–588. De Rudder, V., Poiret, C., & Vourc’h, F. (2000). L’inégalité raciste: l’universalisme républicain à l’épreuve. Paris: Puf. Fassin, D. (2002). L’invention française de la discrimination. Revue française de science politique, 52(4), 403–423. Fassin, D., & Simon, P. (2008). Un objet sans nom: l’introduction des discriminations raciales dans la statistique française. L’Homme, 187–188, 271–294.

10

P. Simon et al.

Guénif, N. (Ed.). (2006). La République mise à nue par son immigration. Paris: La Fabrique. Hajjat, A. (2013). La marche pour l’égalité et contre le racisme. Paris: Éditions Amsterdam. Héran F., Simon P., Debet A., & Boisson M. (2010). Inégalités et discrimination. Pour un usage critique et responsable de l’outil statistique: Rapport du Comedd. Submitted to the Commissaire à la diversité et à l’égalité des chances, Paris. Lapeyronnie, D. (1987). Assimilation, mobilisation et action collective chez les jeunes de la seconde génération de l’immigration maghrébine. Revue française de sociologie, xviii, 287–318. Lorcerie, F. (1994). Les sciences sociales au service de l’identité nationale. Le débat sur l’intégration en France au début des années 1990. In D.-C. Martin (Ed.), Cartes d’identité. Comment dit-on “nous” en politique ? (pp. 245–281). Paris: Presses de Sciences Po. Moguérou L., & Santelli E. (Eds.). (2013). Dossier: “Des jeunes comme les autres ?”. Migrations-­ Société, 25(147–148). Moguérou, L., Santelli, E., Primon, J.-L., & Hamel, C. (2013). Taille de la fratrie et statut social des enfants d’immigrés issus de familles nombreuses. Politiques sociales et familiales, 111, 17–30. Pan Ké Shon, J.-L., & Verdugo, G. (2014). Ségrégation et incorporation des immigrés en France. Mise en perspective temporelle, 1968 à 2007. Revue Française de Sociologie, 55(2), 245–284. Rea, E., & Tripier, M. (2010). Sociologie de l’immigration. Paris: La Découverte. Safi, M. (2006). Le processus d’intégration des immigrés en France: inégalités et segmentation. Revue Française de Sociologie, 47(1), 3–48. Santelli, E. (2007). Grandir en banlieue. Parcours et devenir de jeunes français d’origine maghrébine. Paris: CIEMI. Simon, P. (2007). La question de la seconde génération en France: mobilité sociale et discrimination. In M. Potvin, P. Eid, & N. Venel (Eds.), La deuxième génération issue de l’immigration. Une comparaison France-Québec (pp. 39–70). Outremont: Athéna. Tribalat, M. (1995). Faire France : une grande enquête sur les immigrés et leurs enfants. Paris: La Découverte.

Chapter 2

Migration Histories and Socioeconomic Profiles Cris Beauchemin, Bertrand Lhommeau, and Patrick Simon

2.1  Introduction The Trajectories and Origins (TeO) survey1 was designed to be representative of the total working-age population living in metropolitan France (mainland France and Corsica). Its main objective is to describe the population’s diversity, that is, both that of social trajectories and origins but also that of migration origins. The survey is representative of the entire population living in metropolitan France (living in ordinary households, i.e. excluding communal establishments) and not only minorities from migrant backgrounds, which makes it unique in the landscape of socio-­ demographic surveys aimed at studying the trajectories of first or second generations (i.e. migrants and their French-born descendants; descendants being defined in this text as all French-born children of migrants, excluding grandchildren). Compared with other statistical sources, the TeO survey can be used to describe the trajectories of migrants’ descendants, often unidentifiable in French socio-­ demographic surveys, or present in such small numbers that researchers cannot analyse situations by origin. The TeO survey is representative of minority groups not picked up by these statistical sources which, in general, can only be used to study immigrants belonging to the largest migration flows, in the form of aggregate groups (Maghreb, Southern Europe). It also offers an opportunity for an in-depth

 The TeO survey methodology is presented in an Appendix at the end of the book.

1

C. Beauchemin (*) · P. Simon Institut national d’études démographiques (INED), Paris, France e-mail: [email protected]; [email protected] B. Lhommeau Direction de l’animation de la recherche, des études et des statistiques (DARES), Paris, France e-mail: [email protected] © Springer International Publishing AG, part of Springer Nature 2018 C. Beauchemin et al. (eds.), Trajectories and Origins: Survey on the Diversity of the French Population, INED Population Studies 8, https://doi.org/10.1007/978-3-319-76638-6_2

11

12

C. Beauchemin et al.

study of the situation of French citizens born in the overseas départements2 (DOM in the ensuing text) and their descendants born in metropolitan France, who also constitute a population of interest for the themes developed in the survey. The terms used in this book to designate the survey populations reflect an initial division into two groups based on the migration history of persons across two generations, which distinguishes migrants (born outside metropolitan France, also known as the first generation) and their children born in metropolitan France (the second generation) from the “mainstream population” (defined by convention in this text as all persons who are neither migrants nor descendants of migrants, see Box 2.1). The purpose of this chapter is to present these populations by origin, and show the diversity of their socio-demographic structure (age-sex distribution, social origin, age on arrival for migrants, etc.). It is important to bear these disparities in mind when analysing trajectories in all areas of economic and social life (education, employment, family formation, housing, health, etc.). For example, it comes as no surprise that among immigrant groups where migration is often undertaken as a couple and/or at a relatively advanced age, the proportion of mixed unions is low. This chapter’s second section illustrates the diversity of migration histories among the population living in metropolitan France. It presents a categorization of populations based on linkage with migration, understood as a combination of nationality at birth and a change of residence resulting in settlement in metropolitan France. It also defines the origin-based classifications used throughout the book and shows the geographical spread of the migrant population across metropolitan France’s various regions. The third section covers migrants’ main socio-­demographic characteristics: age and time of arrival in France, age structure at the time of the survey, sex ratio and social background. The fourth section is devoted to the structural differences which distinguish the descendants of immigrants according to their parents’ social and geographical origins. Box 2.1: Glossary • Migrants are people born outside metropolitan France, irrespective of their nationality at birth or at the time of the survey. They are sometimes referred to as “first generation”. By extension and according to a now standard terminology, migrants who arrived in early life and who spent time in education in metropolitan France are said to belong to “Generation 1.5” (in this book, migrants who arrived before the age of 16). Among migrants, the following categories are distinguished: –– Immigrants: persons born outside the current borders of metropolitan France and without French nationality at birth. This category forms the largest group in the migrant population. (continued) 2  In France, a département is one of the three levels of government below the national level, between the administrative regions and the commune.

2  Migration Histories and Socioeconomic Profiles

13

Box 2.1 (continued) –– DOM native-borns: persons born in one of the French overseas départements (DOM), i.e. French Guiana, Guadeloupe, Martinique and Réunion. –– Repatriates: persons with French nationality at birth, born in any of the former French colonies before independence. These countries include Algeria, Benin, Burkina Faso, Cambodia, Cameroon, Central African Republic, Chad, Comoro Islands, Côte d’Ivoire, Djibouti, Gabon, Guinea, Laos, Madagascar, Mali, Mauritania, Niger, Republic of Congo, Senegal, Togo and Vietnam. Repatriates are included in the mainstream population. This terminology does not refer to a specific legal status of persons at the time of their arrival in metropolitan France. –– Other French nationals born outside metropolitan France: persons with French nationality at birth, born outside metropolitan France but not in a DOM or a former colony before independence. This group is included in the mainstream population (defined below). • Natives of metropolitan France are people born in metropolitan France, irrespective of their nationality at birth or at the time of the survey. They include the so-called second generation, i.e. those with at least one migrant parent (and who are thus the second generation in terms of presence in France). They can be broken down into the following categories: –– Descendants of immigrants: persons born in metropolitan France with at least one immigrant parent. Descendants of mixed parentage are persons with one immigrant parent. –– Descendants of DOM native-borns: persons born in metropolitan France, with at least one parent born in a DOM. –– Descendants of repatriates: persons born in metropolitan France, with at least one parent born French in one of the former French colonies before independence. This group is included in the mainstream population. –– Descendants of other French nationals born outside metropolitan France: persons born in metropolitan France, with at least one parent who is a French national by birth but was born outside metropolitan France, elsewhere than in a DOM or a former French colony before independence. This group is included in the mainstream population. –– Native-borns with no migrant parentage: persons born in metropolitan France of French parents who are themselves French-born in metropolitan France. This category forms the largest group in the mainstream population. (continued)

14

C. Beauchemin et al.

Box 2.1 (continued) • The term “mainstream population” means all persons residing in metropolitan France who are neither immigrants nor natives of a DOM, nor descended from DOM immigrants or native-borns. The mainstream population thus includes natives of metropolitan France without direct migrant parentage (mother and/or father), repatriates from the former colonial empire, and all other persons born abroad (excluding French DOMs and colonies) and their sons and daughters. In the non-mainstream population, “visible minorities” are distinguished. Taken from Canada, the term refers in this book to groups whose otherness is made perceptible and who are subject to potentially negative prejudices and stereotypes, i.e. immigrants and descendants of immigrants from the Maghreb, Turkey, Africa and Asia, along with DOM native-borns and their descendants born in metropolitan France.

2.2  The Diversity of Migration Histories 2.2.1  A Typology of Linkage with Migration The TeO survey does not attempt to reconstruct the history of migration in France. It does, however, give an overview of the diversity of migration experiences, on both the personal and family levels (see “typologies of linkage with migration”, Fig. 2.1 and Box 2.1). Within the population living in France, an initial distinction can be made between those who were born in metropolitan France (native-borns) and those – born elsewhere – referred to here as migrants (see Box 2.1). The latter represent 14% of 18–60  year-olds (Table  2.1). Among them, foreign persons born abroad (traditionally referred to as immigrants in French official statistics3) are the most numerous. They number 3.6 million and represent 10% of the working age population of metropolitan France. But the immigration experience in France is not limited to immigrants. It also encompasses French people born outside metropolitan France. This includes DOM native-borns (some 290,000 persons aged 18–60, Table 2.1), but also, in much greater numbers (close to a million people, Table 2.1), 3  In 1991, the High Council for Integration proposed the “immigrant” category, which has since been adopted by INSEE (Simon 1999a). Note that immigrants – foreign born by definition – may acquire French nationality after settling in France. At the time of the survey, an immigrant could therefore be French and/or foreign. On the other hand, French-born persons are never considered as immigrants in French public statistics. This definition of immigrants is original: In many other countries, immigrants are defined as foreign-born persons irrespective of their nationality, or  – alternatively – as individuals who are not nationals of the country concerned, irrespective of their migration history (Poulain et  al. 2006). In the context of the TeO survey, a distinction is made between immigrants and migrants (see Box 2.1 and Box 2.2)

2  Migration Histories and Socioeconomic Profiles

15

Fig. 2.1  A Typology of linkage with migration … without French nationality at birth

Immigrants

Migrants: born outside metropolitan France DOM native-borns

… with French nationality at birth

Repatriates

Other French nationals born outside metropolitan France

Inhabitants of metropolitain France

Mainstream population

without migrant parentage

Parent(s) born in metropolitan France with French nationality at birth

Persons born in metropolitan France with no direct migrant parentage

Descendants of repatriates Persons born in metropolitan France, with French or foreign nationality

Parent(s) born outside metropolitan France, with French nationality at birth

Descendants of other French nationals born outside metropolitan France

Descendants of DOM native-born(s) with migrant parentage

Parent(s) born outside metropolitan France, without French nationality at birth

Descendants of immigrant(s)

Note: For the sake of simplicity, this chart does not include specific cases which concern very small numbers of individuals (persons born in metropolitan France or in a DOM without French nationality, individuals with at least one unknown parent, etc.). These specific cases are presented on teo. site.ined.fr/annexes (Appendix 1)

French people born abroad, only part of whom are “repatriates” (270,000, Table 2.1), i.e. French persons born in the former colonial empire, and who came to live in metropolitan France after independence while they were still children (under 12 years). This figure excludes the former “subjects” of colonial France classified among immigrants (Box 2.3). This leaves 670,000 persons living in metropolitan France who were born French outside France and its colonial empire (Table 2.1).

16

C. Beauchemin et al.

Table 2.1  Population living in metropolitan France according to link with migration Unweighted numbers Migrants born outside metropolitan France Immigrants 8456 DOM native-borns 712 Repatriates 68 Other French persons born outside 225 metropolitan France Natives of metropolitan France Descendants of immigrants 8161 Descendants of DOM native-borns 651 Descendants of repatriates 279 Descendants of other French persons 189 born outside metropolitan France Native-borns with no migrant parentage 3020 Total 21,761 Mainstream population 3781

Weighted numbers, in thousands

Weighted percentages

3583 291 269 667

10 1 1 2

14

3636 220 885 543

10 1 2 2

86

24,606 34,699 26,969

71 100 78

100 78

Coverage: 18–60 year-olds living in an ordinary household in metropolitan France Interpretation: 10% of 18–60 year-olds living in ordinary households in metropolitan France are immigrants Note: See Box 2.1 for definitions of population categories Source: TeO survey, INED-INSEE, 2008

Box 2.2: Interpreting the Results: What “Weighted” Means Persons from the mainstream population make up only 17% of TeO respondents (calculated on the basis of unweighted numbers in Table 2.1). To obtain their true weight in the metropolitan population, weighting must be used (for the calculation method, see the methodological note at the end of the book). Throughout this book, only the weighted results reflect the actual relative weight of respondents in the population, i.e. 78% for the mainstream population (Fig. 2.1 and Table 2.1 for 18–60 year-olds).

This reminds us that historical colonization is not the only way in which French nationals come to live abroad and that France is not only a country of immigration but also, to a lesser extent of course, a country of emigration. Among natives of metropolitan France, those who do not have migrant parents (i.e. without direct migrant parentage) represent the largest category (71% of the metropolitan population aged 18–60, Table  2.1), with it being understood that at least one of their ancestors may have been born outside France. The TeO survey did not collect information on grandparents’ origins, but other sources suggest that, among children born between 2006 and 2008, 15% have an immigrant grandparent without having immigrant parents (Breuil-Genier et al. 2012).

2  Migration Histories and Socioeconomic Profiles

17

Box 2.3: Recording Nationalities and Classifying Them According to Linkage with Migration: The Contributions of the TeO survey The TeO survey questionnaire is particularly thorough when collecting information on the respondent’s nationality and that of their parents. Unlike most statistical sources, the questionnaire records current nationality by allowing the respondent to indicate up to three different nationalities. Declaring oneself as French does not preclude having one or several additional foreign nationalities. The survey therefore enables us to identify bi(tri) nationals. Respondents who report being French are asked to specify by what means they became French. Here too, TeO goes to a much finer level of detail than other sources. Five response categories are proposed: (1) since birth, (2) by reintegration, (3) by naturalization, (4) by marriage, (5) by declaration or option at adulthood or prior to adulthood. The “reintegration” response category relates to persons born in the former French colonies, who were born French and changed their nationality at their country’s independence before regaining French citizenship after migrating to France. Persons who are “French by reintegration” are classified among immigrants (see the rules for classifying the survey populations in on teo.site. ined.fr/annexes). Note that in other sources, such as the census, they are usually classified as French-born rather than immigrants. But the most delicate part of recording nationality concerns the parents’ nationality at birth. Based on this information, the descendants of immigrants from former colonies can be distinguished from those of “repatriates”, i.e. French (former settlers, mostly) who left the colonies to (re)settle in metropolitan France (see Box 2.1). If the parent(s) was (were) born abroad, the respondent must first indicate whether he/she/they had French citizenship at birth. A second question is then asked to respondents with parent(s) born in a former French colony to identify whether they adopted the citizenship of the newly independent country. In this case, the parent is regarded as an immigrant, and the respondent is then defined as the descendant of an immigrant. Only the TeO survey allows such a precise distinction to be made between descendants of immigrants and of former settlers.

The remaining category consists of French native-borns with direct migrant parentage (at the parent level), who represent around 15% of the working age ­population (18–60 years). These are mainly persons with at least one immigrant parent (3.6 million, or 10% of the metropolitan population),4 but also the sons and daughters of DOM native-borns, a relatively small group (220,000 persons) 4  Since the late 1990s, a number of sources have estimated the size of the population of immigrants’ descendants (see Breem 2010). While limited from the viewpoint of age groups, the TeO survey provides more reliable data, thanks to more detailed questions on the nationality of the respondents’ parents (Box 2.3).

18

C. Beauchemin et al.

c­ompared to the 1.4 million French native-borns descended from foreign-born French people. Among the latter, children of repatriates (see Box 2.1) – 885,000 individuals – far outnumber repatriates themselves (whose numbers are limited by the survey’s 60-year age limit). Their number reflects the importance of colonial history in the formation of the French population, although it does not include descendants of immigrants from former colonies who may have reported having French-born parents even if they were not, out of confusion about the rules concerning citizenship in colonial times (Box 2.1). In this book, the term “mainstream population” refers to all persons residing in metropolitan France who are neither immigrants, nor DOM native-borns, nor direct descendants of DOM immigrants or native-borns. This category covers the numerical majority of the population living in metropolitan France (78%). This description suggests that, by contrast, immigrants, DOM native-borns and their children are minority groups. This is true in statistical terms (they make up 22% of 18–60 year-­ olds, Table 2.1) but only partly true in reference to the sociology of racism, which usually reserves the term “minorities” for groups placed in a position of subordination (Simon 2006). We shall see in the results on discrimination and racism presented in this book (Chaps. 8 and 9) that immigrants, DOM native-borns and their descendants are not treated in the same way in France with respect to their origin. The experience of discrimination and racism more directly concerns those that can be grouped under the term “visible minorities” (see Box 2.1), i.e. the populations originating, over one or two generations, from the DOMs, sub-Saharan Africa, Maghreb, South-East Asia and Turkey.

2.2.2  Migration Origins Migrants and their descendants do not form a homogeneous population, but rather a diverse population descended from successive immigration waves (Blanc-Chaléard 2001). Because the TeO survey is based upon an effort to sample the most represented origins in France – even if the numbers of persons are low in absolute terms – the survey makes it possible to produce detailed results by origin. However, because of sampling constraints, countries had to be grouped (see Table  2.2). These aggregated-­origin categories correspond to relatively homogeneous regional sub-­ groups in terms of migration history. Southeast Asian countries (e.g., Cambodia, Laos and Vietnam) share  – beyond colonization and independence  – a common history of exile which spans the 1970s and 1980s. Their history is different from that of Chinese immigration, both more recent and more limited in numerical terms, and therefore not included in this group but rather classified in the residual category “other countries”. Sub-Saharan Africa, which forms a single category in most socio-­ demographic surveys, is split into three parts:

2  Migration Histories and Socioeconomic Profiles

19

Table 2.2  Origin variable’ categories Origin group Sahelian Africa Guinean and Central Africa

Southeast Asia Other EU-27 countries

Other countries African countries formerly under French administration

List of countries Senegal, Mauritania, Gambia, Guinea-Bissau, Guinea, Mali, Burkina Faso, Niger and Chad. Côte d’Ivoire, Ghana, Togo, Benin, Nigeria, Cameroon, Central African Republic, Gabon, Republic of Congo (Brazzaville), Democratic Republic of Congo (formerly Zaire) and Equatorial Guinea. Vietnam, Laos and Cambodia. All EU-27 countries, excluding France, Spain, Italy and Portugal. EU-27 immigrants mostly came from Belgium, the UK, Germany, Poland and Romaniaa. This category covers 95 countries; the principala countries are China, Lebanon, Sri Lanka, and Haiti. Benin, Burkina Faso, Cameroon, the Central African Republic, the Comoros, Congo (Republic), Djibouti Congo-Brazzaville, Gabon, Guinea, Côte d’Ivoire, Madagascar, Mali, Mauritania, Niger, Senegal, Chad and Togo.

The list of principal countries is drawn up based on immigrants’ places of birth. Survey numbers are available on teo.site.ined.fr/annexes (Annex 2)

a

1. Sahelian Africa comprises the Sahel countries of West Africa which produced the first significant flows of sub-Saharan migrants from the 1960s onwards. Imigrants at the time consisted mainly of workers with low levels of education, recruited often in the Senegal river valley, which lies at the border between Senegal, Mali and Mauritania. Workers supplied the industries of the Seine valley with labour. These migrants were preceded, in the early twentieth century, by pioneers, sailors who had left to settle in Marseilles and intellectuals who came to study in French universities (Barou 2011). 2. The “Guinean and Central Africa” category is flanked by the countries of the Gulf of Guinea and the French-speaking countries of Central Africa, marked by later emigration of more educated migrants often fleeing civil wars in their countries (Congo, Côte d’Ivoire, etc.). 3. The other sub-Saharan countries are included in the residual category, “other countries”. These three categories are sometimes switched around to regroup them into sub-­ Saharan African countries formerly under French administration which are subject to special arrangements in respect of rights of residence and which have remained French-speaking (see Table 2.2). Table 2.3 shows the widely known composition of the immigrant population in France (Borrel 2006). It also presents the distribution by origin of descendants of immigrant(s). While immigrants’ origins are relatively easy to define through nationality and country of birth, this is not always true of their descendants, especially if they are born to a mixed-origin couple consisting of two migrant

20

C. Beauchemin et al.

Table 2.3  Origin of immigrants and descendants of immigrants Country or birth region of immigrants and parents of descendants of immigrants Algeria Morocco and Tunisia Sahelian Africa Guinean and Central Africa Southeast Asia Turkey Portugal Spain and Italy Other EU27 countries Other countries Total

Immigrants (aged 18–60) Numbers Weighted numbers Unweighted (thousands) 889 481 1194 679

Descendants of immigrants (aged 18–50) Numbers Weighted numbers Weighted Weighted percentages Unweighted (thousands) percentages 13 1306 617 20 19 1122 474 15

665

137

4

480

76

2

736

238

7

333

52

2

774

116

3

573

82

3

830 847 485

212 414 216

6 12 6

447 933 1692

63 418 777

2 14 25

754 1282

381 710

11 20

649 575

276 246

9 8

8456

3583

100

8110

3080

100

Coverage: persons aged 18–60; descendants aged 18–50. Population living in ordinary households in metropolitan France Interpretation: 889 immigrants born in Algeria were surveyed. They represent 481,000 Algerianborn immigrants living in metropolitan France in an ordinary household in 2008, making up 13% of France’s immigrants Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

parents of different origins. In that relatively rare scenario (7% of descendants with two immigrant parents), for the sake of simplicity, they were assigned the father’s origin. This limit having been set, the results in Table 2.3 suggest that there is a relationship between the number of immigrants (18–60 year olds) of a particular origin and the number of descendants of immigrants of the same origin (here, 18–50 year olds). The relationship depends, inter alia, on the age of the migration wave, the ways in which immigrant families were formed (frequency and intensity of family reunification, propensity to form mixed unions, fertility), or the intensity of their return to their country of origin (about which very little information exists). Broadly speaking, the descendants of immigrants of a certain origin are much more numerous if the wave of immigration is old and large. This is a fairly obvious observation, as the children of the most recently arrived immigrants are not yet 18 years old at the time of the survey (2008). That explains why over half of descendants aged 18–50 are of European origin, with descendants of immigrants(s) in Spain and Italy being the largest groups in numerical

2  Migration Histories and Socioeconomic Profiles

21

terms (27%). Their parents belong to the immigration waves of the late 1930s and the immediate post-war period. Following this chronological order, people with at least one parent from the Maghreb are fewer in number than descendants of Europeans, and the numbers decrease once again for descendants of migrants from sub-Saharan Africa (4% of all descendants), Turkey (2%) or Southeast Asia (2%). Note that this distribution by origin of descendants of immigrants is very much dependent on the age group observed: descendants of African immigrants (including Maghreb) are proportionally more numerous in the younger age groups (48% of descendants aged 18–25  years), while those from European immigration are older (42% of 36–50 year-olds) (Borrel and Lhommeau 2010; Breem 2010; Lhommeau and Simon 2010). Logically, the breakdown by origin of the population of descendants of immigrants will evolve over the coming years following changes in the composition of migration flows, with a couple of decades’ delay.

2.2.3  T  he Uneven Spatial Distribution of Migrants and Descendants of Migrants Migrants and their descendants are not uniformly distributed across the country. While they represent on average 22% of 18–50 year olds in metropolitan France, their proportion varies from 6% to 39% depending on the region (teo.site.ined.fr/ annexes). The Paris region (Île-de-France) traditionally concentrates a greater proportion of immigrants (INSEE 2012). Immigrant presence has tended to grow in this region, where – according to censuses – the estimated proportion of immigrants increased from 14% to 17% between 1990 and 2006, while that of their minor children born in France increased from 25% to 33% (Sagot and Dupoizat 2011). Within this region (the only one for which results can be produced at the département level using TeO survey data), areas of spectacular concentration are observed. For example, immigrants and their descendants represent two thirds of the population of Seine-Saint-Denis, certainly the département with the highest proportion of immigrants in metropolitan France. (Annex 3, teo.site.ined.fr/annexes). Île-de-France is characterized by a very high concentration of visible minorities. Its inhabitants notably include 61% of immigrants from sub-Saharan Africa and 65% of descendants of the same origin, 54% of DOM native-borns and 58% of their children, 49% of immigrants from Southeast Asia and 47% of descendants aged 18–50 (Lhommeau and Simon 2010). Traces of the oldest migration from southern Europe are found in other regions: 53% of immigrants from Italy and Spain live in the South of France. Migration from Turkey is contiguous with the epicentre of Turkish settlement in Germany, in the border regions of eastern France (one fifth in Alsace, Lorraine, Franche-Comté) and in the Rhone corridor (a quarter in Rhône-­ Alpes and Auvergne).

22

C. Beauchemin et al.

Furthermore, one immigrant in five and almost one descendant in seven lives in a sensitive urban area (ZUS). Again, these averages conceal significant differences. While 32% of migrants from Turkey live in a ZUS, this is the case for just 6% of those born in EU27 countries (excluding Spain, Italy and Portugal). Among descendants of immigrants, the proportion is highest for the descendants of sub-Saharan immigrants (28%) and  – again  – lowest (6%) for the descendants of European immigrants. People with a migrant background are significantly overrepresented in sensitive urban areas: more than one in two ZUS residents is an immigrant or a descendant of an immigrant, though these categories represent only one in five persons on the scale of the whole country (Onzus 2011). These results clearly show that the immigrant presence in France is not uniform and that there are areas of high concentration. Whereas most of the results described in this book address the situation of immigrants and descendants of immigrants at the national level, it is essential to keep in mind that they present average situations and that  – in all domains  – they mask deep disparities which remain largely unexplored.

2.3  M  igrants in Metropolitan France: A Diverse Population Upon Arrival Immigration in France is not a new phenomenon. Despite coming into the spotlight fairly recently,5 immigration has had a profound impact on the country’s history since the mid-nineteenth century (Noiriel 1988). Migration history is thus constituted by, to use a visually evocative term, “waves” of migrants; demographers refer to “migrant cohorts” in a more technical language. Though not completely homogeneous, these cohorts are classified within a given time period according to socio-­ demographic characteristics (geographical and social origins, age, sex ratio, etc.). The goal of this section is to identify the socio-demographic characteristics of the survey populations. Throughout the chapter we will be looking at the differences between origin groups – the various spheres of economic and social life will therefore be interpreted against the backdrop of these socio-demographic structures. We will first discuss conditions of arrival (age and period), distinguishing between first-­ generation migrants (immigrants who came as adults) and generation 1.5 (immigrants who arrived in France before age 16 and who were educated and socialized primarily in France); the different origin groups’ socio-demographic profiles (social origin, sex ratio and age at the time of the survey) will subsequently be outlined.

5  Only towards the late 1980s did immigration begin to take shape as a research field in various disciplines: history, sociology, demography (Simon 1999b).

2  Migration Histories and Socioeconomic Profiles

23

2.3.1  P  eriods and Age on Arrival: Heterogeneous Origin Groups The TeO survey opens a relatively narrow window onto the history of immigration: since respondents were between 18 and 60  years old in 2008, only persons who entered metropolitan France between 1948 and 2007 are represented in its sample. The boundary of 1948 is mechanically defined by the “60 years old in 2008” criterion (for those born in 1948). No respondent could therefore have immigrated before 1948. The 2007 boundary corresponds to the date of the annual census from which the survey sample was drawn. This is illustrated by the Lexis diagrams in Fig. 2.2, in which the dark area – representing all immigrants interviewed – is confined on the horizontal axis to those years. The figure also highlights the relationship between date of arrival, age of respondents at their first entry into France, as well as their age at the time of the survey (right-hand side of the figure, for year 2008). The upper limit of “60 years old in 2008” acts on the maximum age on arrival by date of entry. Thus, the far-left side of the graph shows that persons who arrived in 1948 were necessarily less than one year old. The further along we advance in time (moving across the figure from left to right), the higher the maximum age on arrival: persons who arrived in 1958 were 10 years old at most; those who came in 1968, 20 years old at most, and so on. Conversely, the boundary of 18 years in 2008 truncates the minimum age on arrival. Migrants who arrived in 2007 could not have been under 17 years old; those who arrived in 1997 were at most 7 years old on arrival, and so on. Such demographic mechanics have varying effects on the survey sample depending on respondents’ origins, since each group has a unique history. By construction, the oldest immigration waves (Spanish, Italian) are composed of young arrivals, whereas people who are young at the time arrival are quite rare in the most recent waves of immigrants (sub-Saharan, for instance). Clearly, these age structure differences need to be taken into account when considering integration outcomes. In fact, each origin group is characterized by a unique profile when illustrated in a Lexis diagram (Fig. 2.2). By comparing the different scatter plots formed by the various groups’ dots, we discern the succession of waves which have punctuated immigration history in France. The early arrival of Southern Europeans (Spain, Italy, Portugal) is manifested by the cluster of respondents in the graphic’s lower left-hand side. The vast majority of migrants still present in 2008 arrived before 1974, when the French government restricted immigration to family reunification and specific employers’ requests; 71% of migrants of Spanish or Italian origin and 59% of those born in Portugal aged 18–60 in 2008 moved before the “closure” of borders (teo.site.ined.fr/annexes). By a mechanical effect, these immigrants are set apart by their young age on arrival: Spanish and Italian immigrants, were on average 11 years old when they entered France, either with their parents or through deferred family reunification. In the later wave from Portugal, immigrants were on average 4 years older (Appendix 4, teo. site.ined.fr/annexes). However, renewed southern European immigration is discern-

24

C. Beauchemin et al.

Fig. 2.2  Lexis diagrams representing immigrants’ ages and years of first entry into France by origin Other French born abroad

DOM

Algeria

Morocco, Tunisia

60 40 20 0 Sahelian Africa

Guinean and Central Africa

Southeast Asia

Spain, Italy

Other EU-27

Turkey

Age at first admission

60 40 20 0 Portugal

Rest of world

60

Graphs by immie2

40 20 0 1940

1960

1980

2000

2020 1940

1960

1980

2000

2020 1940

1960

1980

2000

2020 1940

1960

1980

2000

2020

Year of first admission

Coverage: Migrants aged 18 to 60 in 2008 Interpretation: Each dot represents at least one surveyed individual, with dot size determined by the number of individuals Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

ible in the upper right side of the charts: since the beginning of the 1980s, many new arrivals have been young adults (20 to 40  years). These new arrivals represent approximately one fifth of the immigrant stock from Spain or Italy aged 18–60 in 2008, and a quarter of the stock of those from Portugal (Appendix 4, teo.site.ined. fr/annexes). These new immigrants attest to the integration of the European labour market. North African immigration, especially from Algeria, is also a dated wave: around one fifth of the immigrant stock present in 2008 had already arrived prior to 1974 (Appendix  4, teo.site.ined.fr/annexes). What makes this migration pattern unique compared to that of Southern Europe is that it has largely continued beyond 1974: the scatter plot occupies the whole spectrum of age and possible periods of arrival (Fig.  2.2). Indeed, 57% of immigrants from North Africa arrived after the mid-­ 1970s (Appendix  4, teo.site.ined.fr/annexes). The steadiness of this flow can be attributed to the rise of family reunification, migration for studies, as well as the exile of many Algerians during the 1990s civil war. Immigrants from other parts of the world have come in more recent waves. The post-1974 “closure of borders” period actually appears to be characterized by a greater diversity of migration flows and migrant profiles. During the period, immigrants

2  Migration Histories and Socioeconomic Profiles

25

Fig. 2.3  Generation 1.5: Proportion of migrants having arrived in France at age 16 at the latest (%) Sahelian Africa Other EU-27 countries Other countriies Guinean and Central Africa Morocco, Tunisia Algeria Turkey Southeast Asia Portugal Spain, Italy DOM Other French born abroad 0

10

20

30

40

50

60

70

80

90

100 Percentage

Coverage: Population aged 18–60 living in ordinary households in metropolitan France Interpretation: 70% of immigrants born in Spain or Italy arrived before age 16 Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

began arriving from Southeast Asia, sub-Saharan Africa, Turkey, new European countries and the rest of the world, as well as from countries which had sent smaller numbers of migrants to metropolitan France in previous periods. More than 90% of immigrants arrived after 1974 (Appendix 4, teo.site.ined.fr/annexes), as shown by the near-empty bottom-left corners in the charts of Fig. 2.2. Immigrants were most often adults, as shown by the low dot density in the diagrams’ lower bands. In the latter half of the 1970s, exile migration began to pick up, in particular from former Indochina (Southeast Asia) and Central Africa (former Zaire, Portuguese-­ speaking African countries affected by civil war, Côte d’Ivoire, etc.). Migration flows from Southeast Asia (Cambodia, Laos and Vietnam) have a unique profile in several respects. They span all age categories and are tightly concentrated in time, with 63% of immigrants arriving between 1975 and 1983 (Appendix  4, teo.site. ined.fr/annexes), during the political crises that caused upheaval in the South Asian peninsula. Turkish immigration followed shortly thereafter (50% of immigrants arrived after 1989), and included large numbers of children, especially after 1974. Immigrants from sub-Saharan Africa form an even more recent flow, with 80% of immigrants surveyed having entered France after 1984. The low dot density below “20 years old on arrival” (Fig. 2.2) speaks to this being essentially a migration of adults. The average age on arrival among all sub-Saharans is 22 years, one of the highest ages among all groups (Appendix  4, teo.site.ined.fr/annexes). Child migration from the Central African states and the Gulf of Guinea (29% of migrants arrived before age 16), is more frequent than from the Sahel (15% arrived before age 16) (Fig.  2.3). This contrasted finding clearly illustrates the heterogeneity of sub-Saharan immigration. Such immigration diversity is the result of very varied

26

C. Beauchemin et al.

national contexts. Outflows from Sahel countries began with labour migration, followed later by migration for family reunification. Emigration from the countries south of the Sahel, on the other hand, is made up largely of refugee and asylum-­ seeker families escaping protracted political and military crises (Côte d’Ivoire, Congo-Kinshasa, Congo-Brazzaville, etc.). The profiles of French nationals born outside metropolitan France are heterogeneous. Flows from the DOMs appear to mainly comprise very young adults; these flows have gained momentum since the 1960s, as shown by the horizontal band of high density dots around “20 years old at arrival” (Fig. 2.2). Moreover, the immigration profile of other French nationals born outside metropolitan France indicates a wide diversity of situations. Repatriates from the former colonial empire are concentrated in the lower left corner of the diagram (Fig. 2.2): these are repatriates who arrived as children, whose parents are excluded from the survey scope due to their age. A quarter of these repatriates arrived between 1961 and 1963, immediately after Algerian independence. But 30% of French nationals born outside metropolitan France arrived before 1961 and 43% after 1963 (around 418,000 persons). Repatriates from Algeria therefore constitute a minority within the group “other French nationals born outside metropolitan France’. The range of the scatter plot also illustrates the extent of migration of “French citizens abroad” throughout the period. In short, migrants of working age (18–60 years) living in France in 2008 are a very diverse population, in terms not only of origins, but also of migration timing and ages on arrival. The respective timing of flows – and hence the potential presence of predecessors and of reception networks – also influences the conditions of integration. Duration of stay affects numerous integration parameters, including language skills, knowledge of norms and codes, the extent of social networks, decisions concerning plans to return and investment in the host society. Age on arrival significantly influences integration into the host society; this holds true over several generations. To simplify, one might say that the younger immigrants arrive in France, the easier their socio-economic integration: early socialization in the country, including through schooling, facilitates adaptation. For France, the results of the Geographical mobility and social integration survey (Mobilité géographique et insertion sociale, MGIS) already highlighted in 1992 the potentially significant differences between immigrants who arrived as adults and those who arrived as children (Tribalat 1996). Age on arrival is not a neutral variable for analysing the integration process. It is important to bear in mind that certain groups are essentially made up of persons who arrived before age of 16 (the so-called generation 1.5). These persons represent 70% of immigrants from Spain or Italy but only 15% of immigrants from Sahelian Africa (Fig. 2.3).

2  Migration Histories and Socioeconomic Profiles

27

2.3.2  Socio-Demographic Contrasts in 2008 2.3.2.1  Age and Sex of Migrants in 2008 The demographic structure used to describe migrants’ arrival conditions (age and period of entry), also has an impact on the age composition of migrants at the time of the survey. The groups, like the Italians and Spanish, which are characterized by early entry in terms of period and age are necessarily represented in the survey by older persons. Likewise, groups made up of newly arrived migrants mainly consist of young adults. We therefore cannot observe the entire amplitude of ageing among migrants who arrived in the inter-war and immediate post-war periods (Attias-­ Donfut 2006). Within this constrained interval, however, there are striking differences between immigrants from Spain or Italy, three-quarters of whom are over 45, and those from Turkey, over half of whom are under age 35 (Lhommeau and Simon 2010). The other groups fall somewhere between these two extremes: DOM native-­ borns, migrants from sub-Saharan Africa, Morocco, Tunisia and Algeria are proportionally the youngest (average ages ranging from 37 to 41 years, Fig. 2.4), while immigrants from Portugal or Southeast Asia are mainly spread across the over-35 age groups (average ages of 46 and 43  years). On average, migrants are slightly Fig. 2.4  Mean age in 2008 of immigrants and descendants of immigrants, in years Spain, Italy Portugal Southeast Asia

Immigrants

Other EU-27 countries

Immigrants with children aged over 18 Descendants of immigrants

Algeria Morocco, Tunisia Other countries Sahelian Africa Guinean and Central Africa Turkey 0

10

20

30

40

50

60

70

80

90

100 Age

Coverage: persons aged 18–60; descendants of immigrants aged 18–50. Population living in ordinary households in metropolitan France Interpretation: descendants of immigrants born in Turkey are 24 years old on average; immigrants from Turkey are 36  years old on average; and immigrants with children aged 18 or above (i.e. parents of descendants who may have been surveyed) are 48 years old on average Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

28

C. Beauchemin et al.

Fig. 2.5  Proportion of immigrants who have had a child in France who was at least 18 years old at the time of the survey (%) Guinean and Central Africa Other countries Sahelian Africa Other EU-27 countries Turkey Algeria Morocco, Tunisia Southeast Asia Portugal Spain, Italy Overall 0

10

20

30

40

50

60

70

80

90 100 Percentage

Coverage: Population aged 18–60 Interpretation: 27% of immigrants born in Morocco and Tunisia have had a child in France and were at least 18 years old at the time of the survey Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

older than the mainstream population (respective average ages of 40 and 41 years for persons aged 18–60). Only European and Asian migrants have higher average ages than the mainstream population (Fig. 2.4). When analysing the age structures of the different groups we must be bear in mind that the immigrants surveyed only very partially correspond to the generation of parents of descendants of immigrants interviewed. In other words, when we compare immigrants to the descendants of immigrants, we are comparing direct and indirect linkages with migration, not parents’ and children’s respective generations. In fact, only 26% of immigrants surveyed are parents of children born in France over age 18. This proportion varies depending on the source: only 11% of Guinean or Central African immigrants are parents of descendants aged 18 and over in 2008. The proportion is 59% for immigrants from Spain or Italy (Fig. 2.5). The origin groups not only differ in age, but also with regard to the sex ratio. From the 1970s and throughout the 1980s and 1990s, women accounted for an ever-­ increasing proportion of immigrants (Beauchemin et al. 2013). The highly imbalanced sex ratios of the early years of work-related immigration have practically disappeared. In 2008, all origins combined, there were more women than men among 18–60-year-old immigrants (48% men compared to 52% women). This process of feminization is particularly marked among migrants from Algeria, and even more so among Guinean or Central African migrants, where women represent around 60% of all migrants, as well as among migrants from EU-27 countries (except Portugal, Spain and Italy).

2  Migration Histories and Socioeconomic Profiles

29

2.3.2.2  Social Origin of Migrants Migration is a selective process: not only do destination countries have legal mechanisms used to “profile” migrants, choosing those they wish to take in (though with fairly limited success), but above all else, migration decisions are not uniformly distributed across populations in countries of origin. The propensity to migrate tends to be higher in people with a medium or high level of education, and with the cultural, social and economic resources needed to fund a plan to immigrate and, especially, to carry it through (Massey et al. 1998). This selection at departure and arrival is visible in immigrants’ social origins, which may in turn determine some of the trajectories followed after migration. The TeO survey provides information on the occupational categories of respondents’ parents when the respondents were 15 years old, making it possible to study these social origins. However, responses to these questions cannot be analysed precisely as it is so difficult to collect comparable information from very diverse contexts, both from a geographical standpoint (respondents’ parents could reside in almost any country in the world) and historically (respondents provide information about when they were 15, so depending on their age at the time of the survey, about any time between 1963 and 2005). The information gathered does, nevertheless, enable researchers to create a broad-brush “social profile” of each origin group. Although immigrants taken as a whole do not differ much from the mainstream population, the series of charts in Fig. 2.6 shows a large diversity of profiles. On these “radars”, the solid shape, which represents the mainstream population, is juxtaposed with the black outline representing each origin group, providing a metric of differences in profiles. Comparison of the charts brings to light three main social types: 1. The first type corresponds to immigrant groups which differ from the mainstream population, in that their parents are more often “manual/clerical workers”. This type covers all populations from the Mediterranean rim (Southern Europe, Turkey, the Maghreb), with a percentage of manual/clerical workers ranging from 61% to 74% (47% for the mainstream population). In total, Turkey and Portugal are differentiated by the relatively large population of farming origin (18% and 17% respectively, compared with 9% in the mainstream population). 2. Persons from families of self-employed workers are overrepresented in two specific groups. Immigrants from Sahelian Africa by and large come from farming families (22% compared with 9% in the mainstream population), while immigrants from Southeast Asia are twice as often from families of self-employed workers than persons from the mainstream population (30% versus 15%). 3. The third type comprises origins where managers/professionals are overrepresented among parents of immigrants. This type includes migrants from the rest of the world (‘other EU-27 countries’ and ‘other countries’) and from Guinea or Central Africa. Among the latter, a quarter comes from families of managers/ professionals versus only 15% among the mainstream population. This finding

Fig. 2.6  Social background of immigrants and DOM native-borns (parents’ occupational category) Overall Manager/professional

Mainstream population

Turkey

Southeast Asia Manager/professional

Manager/professional

50 % 40

Unskilled manualclerical

30 20

Interm. occ.

Unskilled manualclerical

Interm. occ.

Unskilled manualclerical

Interm. occ.

Farmer

Skilled manualclerical

Farmer

Skilled manualclerical

Farmer

10 0

Skilled manualclerical

Self-employed

Self-employed

Portugal

Algeria Manager/professional

Unskilled manualclerical Skilled manualclerical

Interm. occ.

Unskilled manualclerical

Farmer

Skilled manualclerical

Manager/professional Interm. occ.

Unskilled manualclerical

Interm. occ.

Farmer

Skilled manualclerical

Farmer

Self-employed

Spain, Italy

Self-employed

Morocco, Tunisia

Manager/professional Unskilled manualclerical

Farmer

Unskilled manualclerical

Manager/professional Interm. occ.

Skilled manualclerical

Self-employed

Farmer

Unskilled manualclerical

DOM

Manager/professional Interm. occ.

Unskilled manualclerical

Farmer

Skilled manualclerical Self-employed

Farmer

Self-employed

Other countries

Manager/professional

Interm. occ.

Skilled manualclerical

Self-employed

Other EU-27 countries

Self-employed

Sahelian Africa

Manager/professional Interm. occ.

Skilled manualclerical

Skilled manualclerical

Guinean and Central Africa Manager/professional

Self-employed

Unskilled manualclerical

Self-employed

Manager/professional Interm. occ.

Unskilled manualclerical

Interm. occ.

Farmer

Skilled manualclerical

Farmer

Self-employed

Coverage: Individuals aged 18–60; Population living in ordinary households in metropolitan France Definition: Social origin is based on the occupational category of both parents when the respondent was 15 years old. For cases in which the parents belong to different categories, the surveyed individual is classified in the higher of the two categories. The figure is plotted according to the following precise definitions: Manager/professional: at least one manager/professional parent; Farmer: at least one (independent) farmer parent; Self-employed: at least one (independent) self-employed parent; Intermediate occupation: at least one parent in an intermediate occupation; Skilled manual/clerical worker: at least one skilled manual/clerical worker parent; Unskilled manual/clerical worker: at least one unskilled manual/ clerical worker parent and other situations (one unskilled and one inactive; both inactive; info missing) Interpretation: 29% of parents of immigrants, taken as a whole (“overall” figure), belonged to the category of qualified manual/clerical workers, 23% to the category of unskilled manual/clerical workers, 17% to self-employed, 10% were farmers, 10% were in intermediate occupations and 12% were managers/professionals Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

2  Migration Histories and Socioeconomic Profiles

31

corroborates the idea that sub-Saharan migrants are not homogeneous groups. It also shows that African emigration, especially in conflict countries, is extremely selective: the socio-economically advantaged are much more likely to emigrate to the Global North than others (Flahaux et al. 2010). This diversity of immigrant social backgrounds appears to be relatively independent of the structure of national economies in the sending countries. In fact, the proportion of farmers among parents of immigrants never exceeds 22%, even though industrialization is at such an early stage of advancement in certain countries of origin that their populations chiefly work in the agricultural sector. The variety of immigrants’ social origins reflects the differentiated selection processes for migration at departure. In other words, they reflect the wide variety of reasons for migrating, which can be clearly discerned by comparing the two sub-Saharan groups. While the groups themselves are heterogeneous, one (Sahel) is composed of migrants from families with low education migrate in order to work in unskilled labour-intensive industries, while the other (Central Africa and Gulf of Guinea) is composed of migrants from wealthy families able to finance a long-distance migration to pursue their education in France and/or flee political instability or civil war.

2.4  D  escendants of Immigrants: Structural Differences by Origin As shown above, the descendants of immigrant(s) eligible for the TeO survey cannot be – in the vast majority of cases – children of surveyed migrants. Comparing these two groups does not enable to analyse intergenerational developments in the narrowest sense.6 The extent to which the first and second generations living in metropolitan France in 2008 have similar experiences and positions in the different areas of social and economic life nonetheless remains a matter of primary interest. Comparisons should be made prudently, however. First, they should be made between comparably-aged populations. Let us recall that,, the sample of immigrants spans the 18-to-60 age group, while the bulk of the sample of descendants of immigrants(s) covers the 18–50 age group.7 This is why the analyses aimed at comparing immigrants and their descendants chiefly focus on the 18–50 age group. The second precaution for those comparing descriptive results: immigrants and their descendants present different demographic and socio-economic structures, in particular if the descendants of immigrants have only one single immigrant parent.

6  Certain intergenerational comparisons are still possible in areas where the questionnaire contains questions on the respondent’s parents. See in particular the comparison of occupations between the children of immigrants and their fathers (Okba 2012). 7  The survey sample consists of 51 descendants of immigrants above age 50; their numbers are too small for a detailed analysis by origin for this age bracket. See the methodological note for details on sampling.

32

C. Beauchemin et al.

The purpose of this section is to outline the characteristics of the different origin groups to which the sons and daughters of immigrants belong.

2.4.1  Half of Descendants are Born to Mixed-Origin Couples As will be seen throughout this book, having one or two immigrant parents often produces significant differences in education, access to employment and in social and cultural practices, precisely because immigrants who form a union with a person of the mainstream population are generally among the most qualified. Some 90% of descendants of immigrants(s) from EU-27 have parents of mixed origin (excluding Spain, Italy and Portugal) versus only 10% of descendants of immigrants from Turkey (Fig. 2.7). For most origins, around a third of descendants are of mixed parentage and the other two thirds have two immigrant parents. For Spanish and Italian descendants of immigrants the proportions are reversed: two thirds are born to mixed couples and one third have two immigrant parents. On average, all origins taken together, almost half of descendants have a single immigrant parent, 20% having an immigrant mother and a French father and 30% the reverse. Fig. 2.7  Distribution of descendants of immigrants, born to one or two immigrant parents (%) Other countries Other EU-27 countries Spain, Italy Portugal Turkey Southeast Asia Guinean and Central Africa Sahelian Africa Morocco, Tunisia Algeria 0

10

20

30

Two immigrant parents

40

50

60

Mixed (immigrant mother)

70

80

90

100 Percentage

Mixed (immigrant father)

Coverage: descendants of immigrants aged 18–50 Interpretation: 67% of descendants of immigrants from Algeria have two immigrant parents, 8% have an immigrant mother and a French-born father, 25% have an immigrant father and a nonimmigrant mother Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

2  Migration Histories and Socioeconomic Profiles

33

The formation of mixed origin parental couples depends mainly on two factors. The first is migrants’ age on arrival. Though such an explanation might seem ­obvious, it is often overlooked: migrants who arrived as young, still-single people are much more likely to form a mixed union than migrants who arrive after having already begun their married lives. The likelihood that they will form a couple with a partner in the country of installation is even greater if they arrived as children. This is the case for immigrants from Spain and Italy and largely explains why the proportion of mixed parental couples is high among their descendants. But the prevalence of mixed couples also depends upon the degree of openness of French society and of the immigrant populations, which depends in turn on the social, cultural and religious proximity between the groups (see Chap. 5).

2.4.2  Social Origins of Descendants of Immigrants Diversity among parental couples also bears upon the socio-economic context in which the sons and daughters of immigrants(s) develop during their youth. As a matter of fact, persons who have one immigrant parent have a much more similar profile to that of the mainstream population than those whose both parents are migrants (Fig.  2.8). Being born of two immigrants significantly increases the chances of coming from a modest social background. Among descendants with two immigrant parents, 33% come from a family of unskilled manual workers, whereas the proportion drops to 15% and 13% respectively for children born of only one or no immigrants. Conversely, descendants with one or no immigrant parents come much more often from families of managers/professionals (11% and 12% respectively) than do the sons and daughters of two immigrants (3%). The sons and daughters of immigrants(s), like immigrants themselves, present a great diversity of social origins, depending on the geographical origin of their parents (Fig.  2.9). At the lower end of the socioeconomic scale, the proportion of unskilled manual-clerical workers is particularly high among parents of descendants from Sahelian Africa (40% versus 12% in the mainstream population), Algeria (36%), and the rest of the Maghreb and Turkey (29%). They also differ, however, by their relatively large number of self-employed parents (26% compared with 14% in the mainstream population). Skilled manual-clerical workers are particularly numerous among parents from Southern Europe (54% and 44% respectively for descendants of persons from Italy, Spain and Portugal) and the DOMs (53%). At the other end of the socioeconomic scale, we observe almost twice as many manager/ professional parents among descendants of Africans from Gulf of Guinea and Central Africa than in mainstream population (25% versus 13%). There is a lesser over-representation among parents from Asia or other countries (Europe and the rest of the world). Despite all these specificities, the comparison of these different profiles reveals two common features: an over-representation of manual-clerical worker parents (66% for all descendants compared with 46% for people in the mainstream population) and an under-representation of parents in manager/professional and intermediate occupations (20% and 32%, respectively).

34

C. Beauchemin et al.

Fig. 2.8  Social background of individuals by parentage (descendants and mainstream population aged 18–50) Manager/professional

Unskilled manualclerical

Intermediate occupation

10

20

30

40

50 %

Farmer

Skilled manualclerical

No immigrant parent 1 immigrant parent Self-employed

2 immigrant parents

Coverage: Individuals aged 18–50 Interpretation: among parents of persons belonging to the mainstream community (no immigrant parents), 34% belonged to the category of qualified manual workers, 12% to unskilled manual workers, 14% to clerical/sales workers, 7% farmers, 19% intermediate professions and 13% managers/professionals Source: TeO survey, INED-INSEE, 2008

2.4.3  D  escendants of Migrants at Different Stages in their Life Cycle Despite being commonly referred to as children of immigrants in France, the sons and daughters of immigrants surveyed in TeO are all adults between 18 and 50 years old. As the survey excludes minors, we prefer in this book to use the term “(direct) descendants of immigrants” to designate the second generation, even though “descendant” more commonly refers to an entire lineage (see Box 2.1). The descendants of immigrants surveyed, all of them adults, are at different stages of their life cycle. Descendants of Turkish origin, with an average age of 24, are the youngest group. At the other end of the spectrum, Spanish and Italian descendants of immigrants are on average 13  years older (Appendix  5, teo.site.ined.fr/annexes and Lhommeau and Simon 2010). In fact, the average age of descendants of immigrants roughly reflects the timing of migration flows. Logically, migrants from the most long-established groups in metropolitan France began to have children in France earlier than the most recent

2  Migration Histories and Socioeconomic Profiles

35

Fig. 2.9  Social background of immigrants and DOM native-borns (occupational category of parents)

Coverage: Individuals aged 18–50 years. Population living in ordinary households in metropolitan France Definition: See Fig. 2.6 Interpretation: among parents of immigrant descendants, taken as a whole, 42% are in the category qualified manual-clerical workers, 24% unskilled manual-clerical workers, 14% self-employed, 1% farmers, 12% intermediate occupations and 8% managers/professionals Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

36

C. Beauchemin et al.

groups. This is why over half of the descendants of Spanish or Italian immigrants are over age 35 (Appendix 5, teo.site.ined.fr/annexes), it being understood that this group also includes persons over 50 who were not overrepresented in the TeO ­sample (see the methodological note). By contrast, around two-thirds of descendants of Turks and Sahelian(s) are below 26 years old, given that – beyond the TeO sample – they are also well-represented among minor descendants (Breem 2010). On average, in the 18–50 age range in 2008, descendants of immigrants(s) are slightly younger than persons from the mainstream population (32% versus 35%). As will be seen throughout the rest of the book, disparities in age translate, of course, into varying accomplishments in terms of educational, family, and employment trajectories.

2.4.4  A Third Generation? Though descendants of immigrants can be clearly distinguished according to whether they have one or two immigrant parents, consideration must also be given to the immigrant parent(s)’ migration profile. “Immigrants” refer here to all parents born as foreigners outside France; they may have immigrated at any point in their lives and a significant proportion of these parents arrived as children, accompanied by their own parents. In this case, they are in an intermediate position, which we have called ‘Generation 1.5’, having been largely socialized and educated in France. These people are more akin to descendants of immigrants than immigrants themselves. The survey provides information to identify the descendants of immigrants whose parents came as children. Where both parents are migrants, we have used the younger parent’s age on arrival. Figure 2.10 shows the proportion of children whose parents arrived before age 10. While just over 20% of descendants of immigrants have parents who grew up in France, this situation is much more common among descendants with mixed parentage. Mixed unions are much more prevalent among immigrants who arrived young, and even more so among the second generation. The result is that descendants of mixed parentage almost constitute a “third generation”. Such is the case for nearly half of descendants of immigrants from southern Europe, less so for the other groups. Again, this shows that the “descendants of immigrants” category is highly heterogeneous. Differences observed between groups throughout the book reflect their socio-demographic characteristics as much as their origin per se.

2.5  Conclusion Migrants and their descendants living in France have a long history, which we partially truncated in the TeO survey to focus on the 18–60 age group for migrants, and the 18–50 age group for descendants of migrants. The uniform vision of “immigrants” imposed by the use of categories conceals a huge variety of migration

2  Migration Histories and Socioeconomic Profiles

37

Fig. 2.10  Proportion of parents of descendants of immigrants who came to France before age 10 by detailed origin and mixed parentage (%) Overall Other countries

Parents in mixed unions

Other EU-27 countries*

Two immigrant parents

Spain, Italy

Overall

Portugal Turkey* Southeast Asia Sub-Saharan Africa Morocco, Tunisia

* numbers too small for parents in mixed unions

Algeria 0

10

20

30

40

50

60

70

80

90

100 Percentage

Coverage: descendants of immigrants aged 18–50. Population living in ordinary households in metropolitan France Interpretation: 23% of parents of descendants of immigrants came to France before age 10, 11% of those with two immigrant parents and 33% of those with one immigrant parent Note: The list of countries in each geographical entity is given in Table 2.2 Source: TeO survey, INED-INSEE, 2008

conditions, outcomes and trajectories following arrival. These differences are also found among the descendants of immigrants who, though born in France, grew up in very diverse families in terms of social and cultural capital, and were socialized in different ways. The very objective of the TeO survey is to reconstruct the variety of trajectories followed by migrants and descendants of migrants by looking at them through the lens of their family background (origin) and their collective affiliation (immigrant group with its unique history). In order to conduct detailed analyses by spheres of social life, it is first necessary to know the socio-demographic profile of the groups examined; social morphology forms a framework that significantly influences the rules for participation in social life. Age structure, gender, time of arrival in France and social origin are partly responsible for what we observe in terms of education, employment or housing conditions. These structures also influence many of the practices and attitudes recorded in the survey. These elements remind us that comparisons between groups are still limited by the unique socio-demographic characteristics of each one. How can Turkish immigrants who arrived as adults in the 1980s be compared with Spaniards who immigrated as children alongside their parents in the 1960s? Similarly, how can one divine what directions descendants of sub-Saharan African migrants will take at age 50 when the vast majority of them are still under 35? Will they pursue the avenues opened up by the descendants of European immigrants? Readers should bear these

38

C. Beauchemin et al.

differences in mind as they make their way through the ensuing chapters. For example, traits which, at first sight, appear attributable to differences of origin may in fact be a consequence of age differences. History has moulded and shaped these groups, whose paths we must now reconstruct.

References Attias-Donfut, C. (2006). L’enracinement. Enquête sur le vieillissement des immigrés en France. Paris: Armand Colin. Barou, J. (2011). De l’Afrique à la France: d’une génération à l’autre (pp. 13–37). Paris: Armand Colin. Blanc-Chaléard, M.-C. (2001). Histoire de l’immigration. Paris: La Découverte. Borrel, C. (2006). Près de 5 millions d’immigrés à la mi-2004. Insee première, 1098, 1–4. Borrel, C., & Lhommeau, B. (2010). Être né en France d’un parent immigré. Insee première, 1287, 1–4. Breem, Y. (2010). Les descendants d’immigrés. Infos migrations, 15, 1–6. Breuil-Genier, P., Borrel, C., & Lhommeau, B. (2012). Les immigrés, les descendants d’immigrés et leurs enfants. In France portrait social 2012, Insee Référence (pp. 33–38). Paris: Insee. Beauchemin, C., Borrel, C., & Régnard, C. (2013). Immigrants in France: a female majority. Population & Societies, 502, 1–4. Flahaux, M.-L., Beauchemin, C. et al. (2010). Partir, revenir. Tendances et facteurs des migrations africaines intra et extra-continentales. MAFE Working Paper, 7. Insee. (2012). Immigrés et descendants d’immigrés en France. Paris: Insee. Lhommeau, B., & Simon, P. (2010). « Les populations enquêtées ». In C. Beauchemin, C. Hamel, & P.  Simon, (Eds.), Trajectoires et Origines. Enquête sur la diversité des populations en France. Premiers résultats, Paris, Ined, Document de travail n° 168, pp. 11–18. Massey, D. S., Arango, J., et al. (1998). Worlds in motion. Understanding international migration at the end of the millennium. Oxford/New York/Toronto: Clarendon Press/Oxford University Press. Noiriel, G. (1988). Le creuset français : histoire de l’immigration, xixe-xxe siècle. Paris: Le Seuil. Okba, M. (2012). Métiers des pères et des descendants d’immigrés : une mobilité sociale davantage liée à l’origine sociale qu’à l’origine géographique. Dares Analyses, 58, 1–9. Onzus. (2011). Situation socioéconomique des immigrés, intégration et discriminations. Rapport 2011 (pp. 76–99). Paris: Onzus. Poulain, M., Perrin, N., & Singleton, A. (2006). Towards harmonised European statistics on international migration. Louvain-La-Neuve: Presses universitaires de Louvain. Sagot, M., & DuPoizAt, J.  (2011). Les descendants d’immigrés vivant en Île-de-France. Note rapide Société, Institut d’aménagement et d’urbanisme Île-de-France, 531, pp. 1–4. Simon, P. (1999a). Nationality and origins in French statistics: ambiguous categories. Population, an English Selection, 11, 193–220. Simon, P. (1999b). L’immigration et l’intégration dans les sciences sociales en France depuis 1945. In P. Dewitte (Ed.), Immigration et intégration. État des savoirs (pp. 82–95). Paris: La Découverte. Simon, P.-J. (2006). Pour une sociologie des relations interethniques et des minorités. Rennes: PUR. Tribalat, M. (Ed.). (1996). De l’immigration à l’assimilation. Enquête sur les populations d’origine étrangère en France. Paris: La Découverte/INED.

Chapter 3

Educational Trajectories and Transition to Employment of the Second Generation Jean-Luc Primon, Yaël Brinbaum, and Laure Moguérou

3.1  Introduction The purpose of this chapter is to provide information about educational inequalities by migration origin. It tracks the diversity of educational trajectories among descendants of immigrants1 and of DOM native-borns in the French education system, from primary school to the start of working life. The TeO survey2 data are retrospective, not longitudinal, but the indicators they provide allow us to reconstruct a few key stages in people’s educational careers and the school-to-work transition.3 The analyses concern people who were aged 18–35  in 2008, who had been schooled entirely in metropolitan France and were no longer in education, distinguishing between men and women and taking into consideration their migration background in terms of their parents’ countries of origin and nationalities. Other factors liable to  See glossary in Chap. 2 (Box 2.1) for definitions of the population categories mentioned in this chapter. 2  The TeO survey methodology is presented in an Appendix at the end of the book. 3  The available indicators are as follows: repeating a year in elementary school; the track taken at the end of lower secondary school; secondary school certificates obtained or not; access to higher education and whether or not a degree or diploma was obtained; whether a job was found in the first year after completing initial education; experience of unemployment or uncertain employment status at the start of working life. 1

J.-L. Primon (*) Université Nice Sophia Antipolis, Urmis, Nice, France e-mail: [email protected] Y. Brinbaum Conservatoire national des arts et métiers (CNAM), Paris, France e-mail: [email protected] L. Moguérou Université Paris Nanterre, Nanterre, France e-mail: [email protected] © Springer International Publishing AG, part of Springer Nature 2018 C. Beauchemin et al. (eds.), Trajectories and Origins: Survey on the Diversity of the French Population, INED Population Studies 8, https://doi.org/10.1007/978-3-319-76638-6_3

39

40

J.-L. Primon et al.

affect educational and occupational outcomes were also controlled for, such as parents’ social origin and educational level, family structure, number of siblings, school subject speciality, and level of education when they ended their schooling.4 The chapter analyses the educational trajectories of descendants of immigrants or of DOM native-borns5 and their transition to working life. Their trajectories are compared with each other and with those of the mainstream population so as to reveal any specific features related to origin. As with other aspects of their social trajectories, the educational careers and school-to-work transitions of descendants of immigrants are very diverse. Several immigrant origin groups leave school with few qualifications, while others achieve the same level as the mainstream population or higher. In this chapter we aim to find out whether the educational difficulties of immigrants’ descendants reflect social inequalities due to their working-class family backgrounds, or are attributable to other factors connected with migration origin and differential treatment at school. In the last part we analyse the transition to employment and examine whether some origin groups are particularly penalized at the start of their working lives.

3.2  Difficulties from the Start of Primary School As shown by the results of national student panels6 (Vallet and Caille 1996; Brinbaum and Kieffer 2009; Ichou 2013), inequalities among immigrants’ descendants by origin group and gender are evident from the first years of school. Whereas only a quarter of the mainstream population repeat a grade of primary school, the figures are much higher for descendants of immigrants from Sahelian Africa (36%), Algeria (33%), Morocco and Tunisia (34%), Portugal (33%) and Turkey (44%) (Fig. 3.1). Far fewer descendants of immigrants from Southeast Asia (15%) and Guinean and Central Africa (22%) repeat a year. Gender inequality also varies by origin (Fig. 3.1). Among descendants of Moroccan, Tunisian and Sahelian African immigrants, more boys than girls repeat a year, but the situation is reversed for those of Turkish origin, with almost half of the girls repeating a year (48%, compared to 40% of the boys). Repeat years are fewer among descendants of Southeast Asian immigrants of either sex than among the mainstream population, and the figure for girls is especially low (8.5% versus 20% for boys).

 The survey does not give any data on levels of general and academic knowledge.  “Descendants of immigrants” means individuals born in France with at least one immigrant parent who was born abroad with foreign nationality at birth. “Descendants of DOM native-borns” means descendants of individuals born in the French overseas départements and who are living in Metropolitan France. No one born outside Metropolitan France was included in the analysis. 6  The educational trajectory data are self-reported (and retrospective). Figures for grade repetition should be regarded as orders of magnitude; they are useful for showing inter-group differences, but they can be measured more precisely from educational statistics. 4 5

3  Educational Trajectories and Transition to Employment of the Second Generation

41

Fig. 3.1  Repeated years in primary school, by parents’ country of birth and by gender (%)

Coverage: Individuals aged 18–35, schooled in France and no longer in initial education Interpretation: In the mainstream population, 26% of boys and 24% of girls repeated at least one year in primary school Note: The list of countries in each geographical entity is given in Chap. 2, Table 2.2 Source: TeO survey, INED-INSEE, 2008

These rates of grade repetition in primary school reflect both the children’s learning problems and the failure of the education system to teach educational basics to certain student categories. These inequalities appear from the very first year, after which 17% of descendants of Turkish immigrants and 11% of descendants of North African immigrants repeat the year, versus only 2% of Southeast Asian descendants.

3.3  Different Tracks in Secondary School Despite the difficulties encountered in primary school, children rarely drop out at the end of primary schooling. Overall, less than 1% does so – although the share is 5% among descendants of Turkish immigrants, all of them boys (10% of the boys). Half go into apprenticeships. After lower secondary school, students are channelled into different tracks: longor short-track general, technical or vocational.7 The track joined by students depends on their own and their family’s preferences, the decision of the “class council” (conseil de classe) and institutional practices. On average, 59% of the descendants of 7  The survey data do not give details on the courses followed, only the type of school: general lycée, technical lycée, vocational lycée or apprentice training centre (CFA). In the figures, these are simply grouped into “long track” (general and technical tracks) and “short track” (vocational and apprenticeship tracks).

42

J.-L. Primon et al.

Fig. 3.2  Tracks followed after lower secondary school by sons of immigrants and DOM native-­borns, by parents’ country or département of birth

Coverage: Men aged 18–35, schooled in France Interpretation: 50% of young men in the mainstream population aged 18–35 took a long track on leaving lower secondary school; 47% took a short (vocational) track and 3% dropped out of school Note: The list of countries in each geographical entity is given in Chap. 2, Table 2.2 Source: TeO survey, INED-INSEE, 2008

immigrants in our sample of young adults had taken a long track in upper secondary school (general or technical lycée), but with marked differences by gender (47% of boys, 57% of girls) and origin (Figs. 3.2 and 3.3). Some groups are much more likely to take the general or technical long tracks, particularly descendants of Southeast Asian immigrants (61% of boys and 66% of girls) and descendants of immigrants from European countries (except Southern Europe). Girls with DOM native-born or Central or Guinean African parents are also over-represented in the long tracks while boys with parents from Central or Guinean Africa, Portugal or the DOMs, and both sons and daughters of Turkish immigrants are over-represented in the vocational tracks. A majority of Portuguese-origin students take the vocational track and 14% take apprenticeships (the same figure as for the mainstream population), though this trend mainly concerns boys. Most of the girls take the long upper secondary track; their educational aspirations have changed over the generations (Brinbaum and Kieffer 2005). A majority of both sons and daughters of Turkish parents take the vocational track, and almost as many go into apprenticeships as with Portuguese-­ origin boys. Like the Portuguese, Turkish families seem to value vocational training, especially apprenticeships. But among the Portuguese this applies only to boys. Rather than seeing school qualifications as a means of social mobility (as North African families do), Turkish families seem to favour an occupation-oriented model in

3  Educational Trajectories and Transition to Employment of the Second Generation

43

Fig. 3.3  Tracks followed after lower secondary school by daughters of immigrants and DOM native-borns, by parents’ country or département of birth

Coverage: Women aged 18–35, schooled in France Interpretation: 57% of young women in the mainstream population, aged 18–35, took a long course on leaving lower secondary school, 39% took a short (vocational) course and 4% dropped out of school Note: The list of countries in each geographical entity is given in Chap. 2, Table 2.2 Source: TeO survey, INED-INSEE, 2008

which boys join the labour force at an early age. This model of social reproduction is visible in the high proportion of artisans and shopkeepers in the Turkish immigrant population (see Chap. 2). As shown by the indicators on the transition to working life (Table 3.7), a high proportion of male descendants of Turkish immigrants are in work from their first year in the labour market, despite their low average level of education. On the other hand, few of the girls go out to work, with family life taking priority.

3.4  M  any Young People Leave School with No Upper Secondary Qualification Under European Union education policy, leaving school with no upper secondary qualification (with or without the lower secondary certificate, the brevet) is considered as “early school leaving”. In the French education system, a significant proportion of early school leavers quit at the end of lower secondary or at the start of upper secondary in the short CAP and BEP vocational tracks (see Appendix Table A). Many children in these tracks fail or drop out, especially among those who are not in the track or speciality they would have chosen (Brinbaum and Guégnard 2013).

44

J.-L. Primon et al.

While leaving school with no qualifications is seen as a poor outcome in academic terms, the experience actually covers a variety of situations. Youngsters may leave school early because they have performed poorly, owing to learning problems or inadequate teaching, but they may also leave because they have a different view of school, for example a preference for other modes of social reproduction. Among the 18–35 age group in 2008, 15% had left school with no upper secondary qualification,8 and 9% had left without even the brevet (Table 3.1). Descendants of immigrants, taken as a whole, were much more likely to have left school than the mainstream population (21% versus 14%). But the figures vary widely by parents’ country of origin, gender and, of course, family and social background. Immigrants’ sons are more likely to leave the education system with no upper secondary qualification than are the mainstream population (24% versus 16%). The proportion is particularly high (30%) among the sons of immigrants from Algeria, Morocco and Tunisia, Sahelian Africa and Guinean and Central Africa. For sons of Turkish immigrants the proportion is 35%, and 27% do not even have the lower secondary brevet. The proportion leaving school with no brevet or equivalent is also high (24%) among sons of Guinean and Central African immigrants. Some 20% of Portuguese immigrants’ sons leave school with no upper secondary qualification, but many take an apprenticeship to access the labour market. By contrast, the percentage of sons of Southeast Asian immigrants leaving with no upper secondary qualification is close to that for the mainstream population. Thus, in some migrant origin groups, the sons of immigrants are particularly likely to leave compulsory education with no qualifications at all or just the lower secondary brevet. This is obviously a worrying sign for their future, given the importance of academic qualifications in the French labour market. The daughters of immigrants do better at school than the sons and are less likely to leave school with no upper secondary qualifications.9 Their academic achievements are closer to those of their mainstream population counterparts than are the boys’. Daughters of Algerian and Turkish immigrants are exceptions, with 26% and 39% respectively leaving with no upper secondary qualification. Daughters of Turkish immigrants are particularly vulnerable; this is the only migration origin for which more women than men have no upper secondary qualification (39% versus 35%). Leaving school with no upper secondary qualification may be the result of learning problems or lack of personal investment, but for Turkish girls it may also be due to early marriage. This is quite frequent, the spouses being of Turkish origin and met in France or Turkey (see Chap. 10). This practice reflects the traditional model of family and marriage whereby marriage rather than education is the path to

8  The TeO survey provided two different ways to identify the stage at which individuals left the education system, and they do not always concord. One was the path taken at the end of each stage (elementary, lower secondary, upper secondary), the other was the highest qualification obtained at the end of initial education. In the results we have favoured the “highest qualification” data. 9  On average, the boys are 1.4 times more likely to leave school with no qualifications than the girls (and 1.5 times more likely if both their parents are immigrants).

8 18 15 13 16 11 27 11 9 6 8 13

8 9

14 29 24 23 23 17 37 16 17 14 13 21

14 15

Total No incl. no qualifications brevet

1 137 5 293

284 314 541 597 190 223 3 797

359 627 633 272 116

Unweighted numbers

Source: TeO survey, INED-INSEE, 2008 Note: The list of countries in each geographical entity is given in Chap. 2, Table 2.2 Coverage: Individuals aged 18–35 who attended school in France and no longer in initial education at the time of the survey Interpretation: 29% of descendants of Algerian immigrants left the education system with no qualifications, including 18% without the lower secondary certificate (brevet)

Men Women No incl. no Unweighted No incl. no Unweighted qualifications brevet numbers qualifications brevet numbers Country or département of birth of parents of descendants of immigrants or of DOM native-borns DOM 18 13 182 10 2 177 Algeria 32 20 296 26 16 331 Morocco and Tunisia 32 19 290 16 10 343 Sahelian Africa 29 19 129 17 6 143 Guinean and Central 30 24 53 18 11 63 Africa Southeast Asia 18 14 147 16 9 137 Turkey 35 27 152 39 26 162 Portugal 20 14 294 11 7 247 Italy and Spain 18 10 297 17 8 300 Other EU-27 countries 10 5 92 18 7 98 Other countries 14 9 132 11 7 91 24 15 1 882 18 10 1 915 All descendants of immigrants / DOM native borns Mainstream population 16 8 559 13 7 578 All aged 18–35 17 9 2 623 13 8 2 670

Table 3.1  Leaving school with no qualification above the lower secondary school certificate (brevet), by parents’ country or département of birth and by gender (%)

3  Educational Trajectories and Transition to Employment of the Second Generation 45

46

J.-L. Primon et al.

social mobility for women, and parents play a major role in its perpetuation. It is also a way for them to maintain links with their country of origin. Given that a high proportion of immigrant families are working-class families, and given the strong social selection built into the French system (Baudelot and Establet 2009), it is fair to surmise that the educational difficulties encountered by the children of immigrants may be linked to the effects of socio-demographic structures, social or ethnic origins, or institutional practices. To address this question, we performed multivariate analyses to compare the different factors “all other things being equal” (Box 3.1). We ran separate models for the male and female populations because of the differences between the two in their educational trajectories. Box 3.1: Logistic Regression Models Applied to Educational Trajectories Statistical models were constructed on three indicators of the educational career: –– the probability having no upper secondary qualifications (with or without the lower secondary certificate) –– the probability of having an upper secondary qualifications (baccalauréat) –– the probability of having left higher education without a degree. For each indicator, nested logistic regression models were applied to explain the differences in educational achievement between immigrants’ descendants and the mainstream population. The first model (M1) only took account of migration origin (parents’ country or département of birth) to measure inequality between origins. As well as migration origin, the second model (M2) took account of social background using four categories: (1) farmers, artisans and other self-­ employed, (2) unskilled manual and clerical workers, (3) skilled manual and clerical workers and (4) managers, professionals and intermediary; and parents’ level of education (taking the more educated of the two parents), as this is known to be an important factor in educational achievement. Material and socio-economic conditions (whether or not the family had financial difficulties during the respondent’s childhood and youth) were also included, along with housing conditions (whether there was a room where the child could be alone to do homework) and family structure (number of siblings). For the last model, educational characteristics were added: –– variables concerning earlier school trajectory. In the models for secondary education, repeating a year in primary school was taken as an indicator of earlier difficulties (M3); in the models for higher education, the track taken after lower secondary school (long or short track), the field of study (academic or tertiary versus industrial) and age at baccalauréat (whether delayed), were added to Model 2. (continued)

3  Educational Trajectories and Transition to Employment of the Second Generation

47

Box 3.1 (continued) –– variables for assessing parents’ involvement in the child’s education (private tuition) and school context (perceived segregation in lower secondary school) were included in the models for secondary education. In the analysis of higher education, contextual variables were not used, and the variables for material conditions, housing and family structure were included in Model 4 whereas social origin was included in Model 3. No weighting was used in the models.

3.4.1  I n the Male Population, Gaps Are Narrower Where Social Characteristics Are Similar In the raw model, sons of immigrants are more likely than boys in the mainstream population to leave school with no qualifications (see Model 1, Table 3.2)10. This is true for all migration origins except Europe (other than Portugal) and Southeast Asia. Once social and family characteristics are controlled for (Model 2), sons of Portuguese immigrants and DOM native-borns no longer differ from the mainstream population. For most other origins (North Africa, Turkey, Guinean and Central Africa) the gap narrows but persists. In other words, differences in social and educational heritage have an impact on inequalities between males of different migration origins but, at this stage, do not explain all the differences in trajectory. The parents’ educational capital and social level correlate closely with whether or not the child leaves school without qualifications or only the brevet: a boy whose parents have no educational qualifications is almost twice as likely to leave school with no upper secondary qualifications than one whose parents have a vocational qualification (lower secondary level). The probability of gaining a qualification is almost doubled when one or both parents have completed higher education. Social origin also has an impact. Those with parents in higher-level or intermediate occupations have a low risk of leaving school with no qualifications. These results reveal the persistence of social inequality in education and the inability of the French school system to reduce family-based inequalities. Family structure also has an effect on chances of obtaining an upper secondary school qualification (Moguérou and Santelli 2012). Children and adolescents who have always lived with both parents are less likely to leave school with no qualifications than those who grew up in other types of family. The number of siblings has  Odds ratios of 3 for sons of Sahelian African and Turkish immigrants, between 2 and 3 for sons of North and Central African immigrants, and 1.5 for sons of Portuguese immigrants or DOM native-borns.

10

48

J.-L. Primon et al.

no effect on boys’ chances, however. Financial problems in the family increase the probability of leaving school with no qualifications. But controlling for all these factors does not eliminate the inequalities between origins. Model 3 (see Table 3.2) shows that educational inequalities across migration origins are also linked to earlier school experience, starting in primary school. The gaps between the mainstream population and immigrants’ descendants narrow when previous school trajectory is controlled for. Inequalities at this level are apparently linked not only to early difficulties (which in turn correlate strongly with social background) but also to educational segregation. The groups that most often drop out of initial education with no qualifications are more likely to have repeated a year at the start of primary school, and are later unable to catch up: those who repeated a grade are 2.4 times more likely to leave school with no qualifications. This probability is slightly greater when there is strong educational segregation; in that situation, educational levels are lower and initial difficulties are greater. Some groups, such as descendants of immigrants from Turkey and sub-Saharan Africa, are more often in situations of ethnic segregation at school and this has a negative impact on their education (Appendix Figure A) (Felouzis 2003; Felouzis et  al. 2005). All other things being equal, private tuition – considered as an educational strategy but also an indication of financial resources – reduces the likelihood of leaving school with no upper secondary qualification. Leaving school with no qualifications or only the lower secondary school brevet can also be the result of differences between origin groups in the tracks taken at upper secondary level, as we saw above. We tested the impact of upper secondary school track assignment.11 The results show that boys who take the technical and vocational pathways are more likely to leave with no qualifications than those who take the general lycée track, and those who go into apprenticeships are even more likely to do so. These large proportions of dropouts reflect the high frequency of academic failure in the vocational tracks (often because students are assigned to a track or speciality they would not have chosen), although they may also be the consequence of early labour market entry, for example after an apprenticeship. However, after controlling for school track assignment and other educational and social background factors (see Table  3.2,  Model 3), the migration origin effect remains. Variables not included in the models partly explain academic failure among the sons of immigrants from North and sub-Saharan Africa and Turkey. These factors may be to do with the young people themselves, their relationship with the school and their lack of personal investment, but it may also be that the educational institution is treating these descendants of immigrants differently from the rest. The possibility of discrimination against boys from origin groups that are often stigmatized in social life cannot be excluded. It is former students in these groups who most often report arbitrary and discriminatory treatment in track assignment or in their relationship with teachers (Brinbaum and Primon 2013; Brinbaum et al. 2013).

11

 Model not communicated.

3  Educational Trajectories and Transition to Employment of the Second Generation

49

Table 3.2  Factors associated with leaving school with no upper secondary school qualification, by gender (odds ratios) Men Ref. Mainstream population

Country or département of birth of parents of descendants of immigrants or of DOM native-borns

DOM Algeria Morocco and Tunisia Sahelian Africa Guinean and Central Africa Southeast Asia Turkey Portugal Italy and Spain Other EU-27 countries Other countries No qualifications

Parents' qualifications

Social background

Financial conditions Housing conditions Family structure

Primary or secondary school certificate Ref. Vocational lower secondary qualification (CAP-BEP)

* *** *** *** ** *** *

1.1 1.6 1.6 1.8 2.3 1.1 1.8 1 1.3 1.1 0.8 1.4

*** ** ** ** ***

**

Women Model 3 1 1 1.5 1.4 1.5 2 1.1 1.6 0.9 1.2 1 0.8 1.3

** * * **

**

Model 1 1 0.8 2.4 1.3 1.9 1.2 1.4 5.2 0.9 1.5 0.8 0.9

*** **

***

Model 2 1 0.6 1.2 0.7 0.8 0.8 1 2.9 0.7 1.4 0.9 1 2.1

*

*** 0.1 0.1

***

Model 3 1 0.5 1.1 0.7 0.7 0.8 1 2.4 0.6 1.3 0.8 0.9 2

0.8

0.8

1.3

1.3

1

1

1

1

0.5

0.8

0.9

0.9

0.5

0.5

Managers, professionals, intermediate

0.8

Ref. Skilled manual or clerical workers Unskilled manual or clerical workers Farmers. artisans or other selfemployed

1

1

1

1

1.2

1.2

1

1

1

1

0.8

0.9

Ref. No problems Financial problems Ref. Separate room No separate room

1 1.3 1 1.3

1 1.3 1 1.2

1 1 1 1.2

1 1 1 1.2

Ref. Parents living together / other situations

1

Number of siblings

Ref. 2 siblings 3 or more siblings 0 or 1 sibling

Earlier school career

Ref. No grade repeats Grade repeat in primary school

Parental involvement

1.5 2.8 2.4 3 2.3 1.5 3 1.4 1.4 0.8 1

Model 2 1

At least one parent has upper secondary or higher qualification

One-parent family

Concentration of immigrant-origin children in the school

Model 1 1

1.5 1 1.1 1

***

** **

** *

1 ***

1.4 1 1 1

***

*

1 ***

1.8 1 1.6 1.1

0.6

**

*** *

***

**

1 *** ***

1.8 1 1.5 1.1

*** ***

***

1 2.1 1

***

Ref. Low concentration

1 1.8 1

High concentration

1.5

***

1.3

**

Ref. No private tuition Private tuition

1 0.7

**

1 0.7

**

Source: TeO survey, INED-INSEE, 2008 Note: The list of countries in each geographical entity is given in Chap. 2, Table 2.2 Coverage: Individuals aged 18–35 who attended school in France and were no longer in initial education Interpretation: In the male population, the son of an Algerian immigrant has a higher probability (2.8) than a boy from the mainstream population of leaving school with no upper secondary qualification (Model 1) Significance levels: *** p < 0.01; ** p < 0.05; * p < 0.1

50

J.-L. Primon et al.

3.4.2  G  irls’ Perform Better, Excepting Daughters of Turkish Immigrants Compared to the mainstream population, daughters of immigrants from Sahelian Africa and Algeria are twice as likely to leave school with no qualifications, and daughters of Turkish immigrants are five times as likely (see Table 3.2, Model 1). For other origins, however, girls’ achievements do not differ significantly from those of their mainstream population counterparts. In this respect there are fewer inequalities in the female population than in the male one. Controlling for the parents’ social position and educational capital (Model 2), the only significant effect is that of Turkish origin (with an OR of 2.9). Controlling for social and educational background, it even emerges that daughters of Portuguese immigrants and DOM native-borns are less likely than girls in the mainstream population to leave school with no qualifications or only a lower secondary certificate. Model 3 shows a significant family structure effect (of both family type and size). Controlling for school-related factors (Model 3), the results are the same (except for a slightly lower odds ratio for daughters of Turkish immigrants). The models show that both sons and daughters of Turkish immigrants are least successful in school, but also show differences in explanatory factors between boys and girls who leave school with no qualifications. The results suggest that boys face greater difficulty, but also that gendered socialization patterns may have an effect, as may the particular investments made by girls to achieve independence (Brinbaum et al. 2011). Girls from certain non-European migrant groups, may also be treated better in school (Lorcerie 2011).

3.5  Upper Secondary Education Based on the “speciality” chosen after lower secondary school,12 we find that at this stage many students from the mainstream population opt for arts, maths and science or humanities. Among boys, sons of Southeast Asian immigrants are more likely than boys in the mainstream population to take maths and science (27% versus 17%). This is an exception. Very few immigrants’ sons of other origins take this option: only 6% among Algerian immigrants, 5% among Sahelian African immigrants and 6% among Guinean and Central African immigrants. The latter, like the sons of European immigrants, do sometimes take literature, arts, maths, science or humanities, but especially humanities (14% versus 5% of boys in the mainstream  In the TeO survey, subject specialization in school was identified using the official education and training nomenclature (CNCP, NSF 1994) and not the baccalauréat subjects. The terms used are speciality, field and course. The “academic” field includes maths and science, arts, humanities or law. The other specialities are technical or vocational courses, divided into three fields: production (engineering, construction, materials etc.), services (management, communication, information, transport, trade, etc.) and personal development (personal services, sport, domestic activities, etc.).

12

3  Educational Trajectories and Transition to Employment of the Second Generation

51

population boys). Boys of this group are less frequently found in technical or vocational training for industry (31% versus 55%). Sons of immigrants from Sahelian Africa (42%), Morocco and Tunisia (34%) are more likely than the average to opt for technical or vocational training in the highly feminized services branch (versus 15% of the mainstream population) and are broadly concentrated in technical and vocational courses. Sons of Portuguese and Turkish immigrants are the only groups to opt mostly for technical and vocational training for industry (63% and 65% respectively). The male population is sharply divided between academic courses (arts, maths and science, humanities) and technical and vocational training for industry and services, with a very sharp divide between services (including human development) and industry. There are fewer disparities in the female population. Daughters of immigrants from Sahelian Africa, Algeria and Turkey are over-represented in technical/vocational courses, mainly highly feminized service specialities (68%, 63% and 60%, respectively) but also in far more male-dominated specialities leading to jobs in industry.

3.6  Obtaining the Baccalauréat Access to the baccalauréat upper secondary diploma depends on the track taken after lower secondary school. Students in technical and vocational tracks more often leave school with a lower-secondary vocational certificate (CAP or BEP) than the baccalauréat. For example, sons of Portuguese and Turkish immigrants are over-­ represented in industrial courses: 40% and 39%, respectively, have a CAP or BEP as their highest qualification, compared with 26% of men in the mainstream population. Thus the sons of Portuguese or Turkish immigrants conform to the traditional model for reproducing the working class (Baudelot and Establet 1971), in line with family aspirations (Brinbaum and Kieffer 2005), whereas other origin groups seem to prefer tertiary-sector or academic courses. As a consequence of the education policies of the 1990s, the aspiration to take the baccalauréat is now the norm (Beaud 2002), including among immigrant families (Brinbaum and Kieffer 2005). Immigrants’ descendants have benefited to some extent from more widespread access to the baccalauréat: 55% of them, on average, obtain the certificate (compared with 62% of the mainstream population; see Table 3.3). The proportion of boys who get the baccalauréat is markedly lower among sons of immigrants than in the mainstream population, and lowest of all among sons of Turkish immigrants (26%). Only sons of immigrants from Southeast Asia (59%) and EU-27 countries (81%) achieve success rates as high as or higher than the mainstream population. Some, as we have seen, compensate by getting vocational qualifications. Regression analysis of percentages of boys obtaining the baccalauréat (Table 3.4) shows that when social characteristics are controlled for (Model 2), the disadvan-

52

J.-L. Primon et al.

Table 3.3  Percentage having obtained the baccalauréat, by gender and origins (%) Unweighted Men Women Total numbers Country or département of birth of parents of descendants of immigrants or of DOM native-borns DOM 48 70 59 359 Algeria 41 51 46 627 Morocco and Tunisia 45 63 55 633 Sahelian Africa 40 58 48 272 Guinean and Central Africa 46 69 60 116 Southeast Asia 58 70 63 284 Turkey 26 38 31 314 Portugal 41 64 51 541 Italy and Spain 52 66 58 597 Other EU-27 countries 81 75 78 190 Other countries 65 81 72 223 All descendants of immigrants 48 62 55 3 797 Mainstream population 59 65 62 1 137 57 65 61 5 293 Source: TeO survey, INED-INSEE, 2008 Note: The list of countries in each geographical entity is given in Chap. 2, Table 2.2 Coverage: Individuals aged 18–35 who attended school in France and are no longer in initial education at the time of the survey Interpretation: 41% of sons of Algerian immigrants obtained the baccalauréat, versus 51% of girls of the same origin

tages still hold for sons of DOM native-borns, Sahelian and Central Africans and Turkish immigrants, but disappear for sons of North African immigrants. When family structure and financial circumstances (Model 2) and previous educational trajectory and schooling conditions (Model 3) are taken into account, only the sons of Turkish immigrants are less likely than boys from the mainstream population to get the baccalauréat. Girls are proportionally more likely than boys to get the baccalauréat (65% versus 57%). The percentages for immigrants’ daughters match that for daughters of mainstream population parents and, for some origins, exceed it. Only daughters of Turkish immigrants show a lower success rate than mainstream population daughters of comparable socioeconomic status (Model 2, Table  3.4). Daughters of Moroccan, Tunisian and Portuguese immigrants are all more likely to get the baccalauréat than are mainstream population girls after controlling for social and family characteristics. This advantage remains, and is even stronger, when educational trajectory and family involvement are controlled for (Model 3), confirming that girls of immigrant origin often do better at school (Brinbaum and Kieffer 2009; Brinbaum et al. 2011). But the disadvantage remains for Turkish-origin girls. Daughters of immigrants from Algeria and Sahelian Africa have the same likelihood of getting the baccalauréat as mainstream population girls with the same social and family characteristics. When family type and size are controlled for, daughters of Sahelian African immigrants are more likely than mainstream population girls to get the baccalauréat (whatever the speciality). Controlling for school-

At least one parent has baccalauréat or higher qualification Higher-level or intermediate occupations Ref. Skilled manual or clerical workers Unskilled manual or clerical workers Farmers, artisans or other self-employed Ref. No problems Financial problems Ref. Separate room No separate room Ref. Parents living together/other situations Single-parent family Ref. Two siblings 3 or more siblings 0 or 1 sibling Ref. No grade repeats Grade repeat in primary school Ref. Low concentration High concentration Ref. No private tuition Private tuition

Ref. Vocational secondary qualification (CAP-BEP)

Men Model 2 1 0.8 0.9 0.9 0.7 0.6 1.4 * 0.5 *** 0.8 0.8 1.6 * 1.4 * 0.7 ** 1.4 *** 1 2.2 *** 2 *** 1 0.9 1.1 1 0.9 1 0.8 ** 1 0.6 *** 1 0.9 1.1 Model 3 1 0.8 0.9 1 0.8 0.7 1.4 0.6 ** 0.9 0.8 1.8 ** 1.6 ** 0.8 * 1.4 ** 1 1.9 *** 1.6 *** 1 1 1 1 1 1 0.8 1 0.7 *** 1 1 1 1 0.3 *** 1 0.6 *** 1 2.1 ***

Model 1 1 1 0.6 *** 0.9 0.6 *** 1.2 1 0.3 *** 0.9 0.9 1.8 ** 1.9 **

Women Model2 1 1.5 ** 1.3 2 *** 1.5 * 1.7 * 1.5 * 0.5 *** 1.4 * 1 1.7 * 1.9 ** 0.6 *** 1.1 1 2.5 *** 2.1 *** 1 1.1 1.4 ** 1 1 1 0.7 *** 1 0.5 *** 1 0.7 *** 1.3 ** 0.8 1 0.5 1 0.7 1.3 1 0.3 1 0.7 1 1.5

***

***

***

** *

***

*

Model 3 1 1.7 ** 1.4 * 2.1 *** 1.7 ** 1.8 * 1.4 0.7 * 1.5 ** 1.1 1.9 ** 2.1 ** 0.7 *** 1.1 1 2.4 *** 1.7 *** 1 1.1 1.2 1 1

Source: TeO survey, INED-INSEE, 2008 Note: The list of countries in each geographical entity is given in Chap. 2, Table 2.2 Coverage: Individuals aged 18–35 who attended school in France and were no longer in initial education at the time of the survey Interpretation: In the male population, a son of a Portuguese immigrant is 0.6 times as likely to obtain the baccalauréat rather than not obtain it as a boy from the mainstream population (Model 1). The results are presented in the form of odds ratios Significance levels: *** p