## Yann Algan, Alberto Bisin, Alan Manning, and Thierry Verdier

Print publication date: 2012

Print ISBN-13: 9780199660094

Published to Oxford Scholarship Online: January 2013

DOI: 10.1093/acprof:oso/9780199660094.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 02 July 2020

# Cultural Integration in France

Chapter:
(p.49) 2 Cultural Integration in France
Source:
Cultural Integration of Immigrants in Europe
Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780199660094.003.0002

# Abstract and Keywords

The French Republican model appears as a polar case among the different cultural integration models. Dating back to the French Revolution and the Third Republic, France has a long secular tradition imposing restrictive attitudes on the expression of religious and cultural identity in the public sphere. There are, however, growing concerns that this model, despite its claimed egalitarianism and universalism, fails to integrate the new immigrant minorities. The most illustrative example is the 2004 ruling against the display of conspicuous religious symbols in school, mainly targeted at Muslim schoolgirls who wished to wear the hijab. The main consequence of this refusal to acknowledge any minorities has been an inability to know whether the reality of equality matches the rhetoric of perfect cultural integration. While views on national identity and the integration model are very strongly held in France, the evidence base is rather weak. The goal of this chapter is to fill this gap.

# 2.1 Introduction

Immigration has a very long history in France since the late nineteenth century (Noiriel, 1988). In the 1920s, France ranked second, just after USA, as the country with the highest share of immigrants, reaching seven per cent of total population. In the early 2000s, as many as 25 per cent of the population had some immigrant background, from the first, the second or the third generation.

Table 2.1 reports on the composition of the immigration population according to the most recent data set, the French Labour Force Survey, for the period 2005–2007. It distinguishes the sample proportions of native French, first-generation immigrants and second-generation immigrants. Around 90.2 per cent of the sample consists of natives, 6.5 per cent are first-generation immigrants and around 3.3 per cent are second-generation immigrants.

First-generation immigrants mostly come from Maghreb (44.1 per cent), Southern Europe (24.8 per cent), and Africa (11.3 per cent). These percentages are slightly modified for second-generation immigrants, the share of immigrants from Southern Europe is higher (37.4) and those from Africa (5.0) and Maghreb (40.7) is lower.

Table 2.1 also shows an evolution over time in the composition of the population of immigrants. During the first half of the twentieth century, immigration in France was mainly driven by inflows from Southern Europe, in particular from Italy and Spain, with some peaks, such as the inflows of Spanish immigrants during the Spanish Civil War. (p.50)

Table 2.1 Origins of immigrants in the French Labour Force Survey 2005–2007.

Country of origin

First generation

Second generation

Natives

90.2

Immigrants of which (%)

6.5

3.3

Maghreb

44.1

40.7

Southern Europe

24.8

37.4

Africa

11.3

5.0

Northern Europe

6.6

3.7

Eastern Europe

5.9

7.5

Turkey

4.1

3.6

Asia

3.2

2.2

Note: Data source is the French Labour Force Survey (LFS) 2005–2007. Proportions are computed using individual sampling weights.

A second wave of immigration from Southern Europe took place in the 1960s and the 1970s, with ongoing inflows from Portugal now.

Immigration from Maghreb dates back to as early as the First World War, driven by the replacement of the labour force in farms and arms industry. But the main wave of immigration from this region took place after the Second World War. Immigration inflows come from three main countries: Algeria, Tunisia, and Morocco. Immigration from Algeria boomed after the Second World War until 1958 and the Algerian civil war. Immigration from Morocco and Tunisia took place later, during the 1970s.

Immigration from Sub-Saharan Africa is a more recent phenomenon. The immigrants from this region are mainly from the French ex-colonies: Cameroon, Ivory Coast, Mali, and Senegal. The most recent wave of immigration is from Eastern Europe and Turkey, with ongoing inflows from Turkey since the 1970s. The smallest group of immigrants come from Asia. Most of those immigrants originate from the ex-French colonies in South-East Asia: Cambodia, Laos, and Vietnam.

French immigration policy is rather pragmatic and dictated by the labour market conditions. First-generation immigrants are now accorded permits of various tenures ranging from one to ten years. Regarding citizenship, Weil (2002, 2005) documents that France is one of the most open countries in Europe. For second-generation immigrants, naturalization comes from the right of birth. Any immigrant born in France is granted French citizenship, but this right becomes effective mostly when children are older than 18 years.

Despite France’s long immigration tradition, and the growing concerns about persistent cultural differences with immigrants from (p.51) Maghreb and Africa, very few studies have provided a quantitative assessment on the cultural integration path. Most studies have rather looked at economic outcomes. Silberman and Fournier (1999, 2007) look at job outcomes and show the persistent employment penalty for second-generation Maghrebis compared to French natives and other immigrant groups. Aeberhardt and Pouget (2007) estimate national wage origin differential by matching employer-employee data. They typically find that earning differentials mostly reflect differences in the type of jobs, suggesting the existence of occupational segregation rather than mere wage discrimination. Besides, it has been well documented that second-generation immigrants from Maghreb face the highest penalty on the French labour market among the different immigrant groups (see Algan et al., 2010). Recent audit studies show that this labour market penalty is partly driven by pure cultural discrimination (Adida et al., 2010).

In other social sciences a strong debate opposes the supporters of the Republican model, stressing that ethnic origin does not have to interfere with the public sphere (Schnapper, 1991) and those who call for a civil society are more open to multiculturalism (Wieviorka, 1996). But few economic studies have tried to quantify the evolution process of cultural attitudes by waves of immigration and birth cohorts. Yet, there is growing evidence of a strong interplay between cultural and economic integration in France. In particular, Algan et al. (2011) focus on the transmission of Arabic name versus non-Arabic name in the French society. They show that parents do take into account the expected economic cost that they inflict on their child by choosing a culturally distinctive name in order to maintain their cultural trait.

This chapter tries to fill this gap by providing a quantitative assessment of the path of cultural and economic integration of immigrants in France.

# 2.2 Data and methods

## 2.2.1 Data

We investigate the patterns of integration in France by using three main surveys. We measure labour market and educational outcomes with the French Labour Force Survey (FLFS), which cover the years 2005–2007. In addition to the traditional information on country of birth of the respondent, the FLFS has, since 2005, provided information on the country of birth of the parents. The FLFS contains information on country of birth for first-generation immigrants at a very detailed (p.52) level. It distinguishes between 29 countries or regions.2 The FLFS also reports the country of parental birth for the second generation but at a more aggregate level. There are nine categories: France, Northern Europe, Southern Europe, Eastern Europe, the Maghreb (Arab North Africa), Turkey (Middle East), (Sub-Saharan) Africa, Asia, and other countries. We exclude the last category as it comprises very heterogeneous populations. This leaves us with seven immigrant groups for our analysis. To facilitate the comparison of the results between first-generation and second-generation immigrants, we aggregate the more detailed countries of birth of first-generation immigrants into the seven broader immigrant categories. The native reference group consists of individuals who have lived in France for at least two generations, that is, those who are born in the country and whose two parents were also born in France. First-generation immigrants are individuals born abroad and whose parents were also born abroad and from the same country of origin. Second-generation immigrants are individuals who are born in France but whose parents were both born abroad. We exclude individuals born abroad with at least one parent born in France and individuals born in France with either one parent born in France and the other born abroad or both parents born abroad but in different countries.

We measure fertility rates based on the 1999 French Family Survey ‘Enquete Histoire Familiale’ (1999). This survey was conducted in parallel with the Population Census and aimed at analysing the evolution of family structures. It consists of a sub-sample of 380,000 adults, and the survey includes several questions about family status and family relationships, country of birth of the respondent, of his/her relatives (parents, husband/wife), language spoken at home, with children, with parents, etc. In particular this survey is extensively used to compute reliable completed fertility rates.

The French family survey displays three types of information concerning the origins of the respondent. It provides information on the respondent’s country of birth: the recorded countries are broken down into 16 categories.3 The survey also records the country of birth of the father, of the mother, and of the spouse. The countries that are recorded are exactly the same as for the survey respondent. The survey (p.53) also gives information on the nationality (citizenship) of the respondent (at the time the survey was conducted, and at birth). The list of citizenship is exactly the same as the list of country of birth. To compute homogenous regions of origins, we cluster the countries: (1) France: France; (2) Northern and Eastern European Countries; (3) Southern Europe: Italy, Spain, Portugal; (4) Maghreb: Morocco, Algeria, Tunisia; (5) Africa: Sub-Saharian African countries; (6) Asia: Vietnam, Laos, Cambodia, and other Asian countries; (7) Others: mainly African countries. Contrary to the Labour Force Survey used for the analysis of economic integration, we cannot make a distinction here between individuals from Northern European countries and those from Eastern European countries.

In order to explore subjective attitudes of immigrants, we also a survey ‘Histoires de Vies’, conducted in 2003 by the French national statistical office (INSEE). The sample of the survey includes 8403 adults living in France, with a deliberate over-sampling of immigrants of the first and second generation. The survey includes many questions pertaining to subjective identity, gender issues and work values. It contains information about the country of birth of surveyed persons, their parents and their living partner (if any). Due to the small size of the sample, we only distinguish four main categories of ethnic origin, aggregating countries into large regions as follows: (1) France; (2) Southern Europe: Italy, Spain, Portugal; (3) North Africa or Maghreb: Algeria, Morocco, Tunisia; and (4) rest of the World (foreign country, but not Southern Europe or Maghreb). We chose to distinguish Maghreb and South Europe as these are the most important sources of immigration in France. For instance, in the 1999 French census, those two groups accounted for 62 per cent of foreign immigrants.

## 2.2.2 Specification

We compare the economic and cultural outcomes between first and second generation of immigrants by using the following specification:

$Display mathematics$

where $β j$ and $γ j$ measures the impact of being a first-generation immigrant and a second-generation immigrant from country j relative to (p.54) natives.4 First-generation immigrant and second-generation immigrant are dummies equal to 1 if the individual belongs to either group and 0 otherwise. First-generation immigrants are defined as individuals who are foreign-born from country j and whose two parents are foreign-born. Second-generation immigrants are defined as immigrants who are born in the host country France, but whose parents are both foreign-born from country j. The reference group is represented by the natives in the host country, that is individuals who are born in France and who have both parents also born in France. The natives are always considered as the omitted group.

# 2.3 Fertility and marriage

## 2.3.1 Fertility and age at first child

We look at two different outcomes in terms of fertility: completed fertility rates and age at first child.

To investigate the impact of ethnicity on completed fertility rates, we restricted the sample to women older than 40 to avoid censoring issues due to younger women not having completed their fertility. An alternative solution would have been to include all women regardless of their age and to include a polynomial in the age of the woman as explanatory variables.5

Table 2.2 reports the coefficient estimates associated with completed fertility rates of immigrants relative to natives. Positive coefficients on first-generation migrants in the first column of Table 2.2 mean that, regardless of their region of origin, immigrants have a greater completed fertility rate on average than native women. Among all immigrants, immigrants from Maghreb, Asia, and Africa exhibit the highest fertility rates. First-generation immigrant women from Maghreb have on average 0.56 more children than natives, and immigrants from Asia and Africa have 0.32 more children than natives. However, this discrepancy (p.55)

Table 2.2 OLS estimates of completed fertility rates by country of origin and immigration generation.

Country of origin

First generation

Second generation

France

Reference

Africa

0.328***

0.126

(0.054)

(0.136)

Asia

0.329***

−0.065

(0.053)

(0.111)

Europe

0.019

−0.036

(0.030)

(0.025)

Southern Europe

0.043

−0.249***

(0.024)

(0.023)

Maghreb

0.566***

0.166***

(0.020)

(0.047)

Controls

Age, education, occupation

N

135,025

R2

0.090

Note: EHF 1999. The sample is for all women over 40. Specification is that of model (1). Standard errors clustered at the country of origin level. * Denotes statistical significance at the 10% level, ** at the 5% level, and

(***) at the 1% level.

seems to be greatly reduced for the second generation of immigrants. Second-generation women from Maghreb have only 0.16 more children during their lives than natives. For second-generation women from Asian origins, the difference with natives vanishes completely and is not significantly different from 0. Women born from parents from Southern Europe have 0.24 less children on average than French natives.

To estimate age at first child, we use all women aged 40 or younger and use a censored model to control for women without children at their current age.6 Results are displayed in Table 2.3 and show that first-generation immigrants from Africa, Southern Europe, and Maghreb tend to have children earlier than natives. Median age at first birth is one year earlier for first-generation immigrants from Africa, and 0.23 years and 0.35 years earlier for women from Southern Europe and Maghreb, respectively. Note that these differences tend to persist among second-generation women from Africa and Maghreb who still have their first child 0.35 and 0.33 years earlier, respectively, than native women.

## 2.3.2 Marriage and divorce rate

(p.56)

Table 2.3 Estimates of the age of the mother at first birth by country of origin and generation of immigration.

Country of origin

First generation

Second generation

France

Reference

Africa

−1.082***

−0.351*

(0.133)

(0.333)

Asia

0.921***

−0.249*

(0.141)

(0.116)

Europe

0.329***

−0.332**

(0.083)

(0.070)

Southern Europe

−0.232***

0.649***

(0.065)

(0.062)

Maghreb

−0.351***

−0.329***

(0.055)

(0.103)

Controls

Age, education, occupation

N

88,449

R2

0.039

Note: EHF 1999. Censored median regression estimates (Chernozukhov) to deal with censoring of women not having children at their current age.

(*) Denotes statistical significance at the 10% level,

(**) at the 5% level, and

(***) at the 1% level.

We next consider marriage patterns. We compare marriage rates at age 257 for natives and first and second-generation of immigrants. We restrict the sample to all men and women aged between 25 and 40. Table 2.4 displays the results for men and women, and then breaks down the results by gender. Marginal effects at the mean of a probability model of being or having been married at age 25 are reported.

Results show that first-generation immigrants tend to marry more and earlier than native individuals. This difference is especially large for individuals coming from Europe and Southern Europe, and for individuals coming from Maghreb. The probability of being married at age 25 is 7.9 percentage points higher for European immigrants, 7.2 percentage points higher for immigrants from Southern Europe and 1 percentage point higher for immigrants from Maghreb. This can be compared with an average probability of being married at age 25 of 27 per cent in our estimation sample. The difference between immigrants and natives is greatly reduced for the second generation. It is even reversed for second-generation immigrants from Maghreb, who have (p.57)

Table 2.4 Estimates of the probability of being married at age 25 by country of origin and generation of immigration.

Country of origin

All

Men

Women

First gen.

Second gen.

First gen.

Second gen.

First gen.

Second gen.

France

Reference

Africa

−0.038**

−0.077***

−0.030

−0.093***

−0.044**

−0.063**

(0.015)

(0.017)

(0.022)

(0.020)

(0.020)

(0.028)

Asia

−0.010

−0.059**

−0.055**

−0.063**

0.038

−0.054*

(0.017)

(0.021)

(0.022)

(0.028)

(0.026)

(0.031)

Europe

0.079***

0.005

0.086***

0.008

0.070***

0.002

(0.013)

(0.011)

(0.020)

(0.017)

(0.017)

(0.016)

Southern Europe

0.072***

0.009

0.084***

0.009

0.058***

0.011

(0.014)

(0.007)

(0.021)

(0.010)

(0.017)

(0.010)

Maghreb

0.010

−0.025***

−0.003

−0.037***

0.026*

−0.010

(0.010)

(0.006)

(0.014)

(0.008)

(0.014)

(0.008)

Controls

Age, education, occupation

N

88,449

40,029

61,570

R2

0.095

0.059

0.086

Note: EHF 1999. Logit estimates: marginal effects at the mean. Sample: individuals under 40.

(*) Denotes statistical significance at the 10% level,

(**) at the 5% level,

and (***) at the 1% level.

a slightly smaller probability of being married at age 25 than native individuals (minus two percentage points). The next columns in Table 2.4 investigate the same probability model for men and women separately. The main result is that men and women from the same region of origin do not seem to differ significantly in their marriage behaviours. Both men and women migrating from Europe and Southern Europe have a higher probability of being married at age 25 than native French, but second-generation men and women from these same regions do not have significantly different marriage behaviours from native French.

Among immigrants from Maghreb, only women seem to be more likely to be married when they are young (with a higher probability of 2.6 percentage points), whereas men seem to marry later. This may reflect the different nature of immigration between men and women from Maghreb, men coming younger and for working purposes and women coming for family reasons along the policy of ‘family gathering’.

(p.58)

Table 2.5 Estimates of the probability of being or having been divorced by country of origin and generation of immigration.

Country of origin

All

Men

Women

First gen.

Second gen.

First gen.

Second gen.

First gen.

Second gen.

France

Reference

Africa

0.003

−0.0307

0.064*

−0.081***

−0.039**

0.000

(0.017)

(0.021)

(0.035)

(0.009)

(0.016)

(0.031)

Asia

−0.044***

0.080**

−0.033**

0.095

−0.054***

0.067

(0.014)

(0.039)

(0.024)

(0.068)

(0.017)

(0.044)

Europe

−0.009

0.042***

−0.001

0.032

−0.015

0.0476**

(0.010)

(0.015)

(0.018)

(0.024)

(0.012)

(0.020)

Southern Europe

−0.000

0.006

0.005

0.010

−0.003

0.002

(0.013)

(0.008)

(0.022)

(0.013)

(0.016)

(0.010)

Maghreb

−0.001

0.045***

0.005

0.049***

−0.010

0.041

(0.010)

(0.008)

(0.015)

(0.015)

(0.013)

(0.010)

Controls

Age, education, occupation

N

51,087

17,628

33,459

R2

0.032

0.026

0.038

Note: EHF 1999. Logit estimates: marginal effects at the mean. Sample: all individuals being or having been married.

(*) Denotes statistical significance at the 10% level,

(**) at the 5% level, and

(***) at the 1% level.

We then look at divorce patterns. We consider the fraction of individuals who got divorced.8 Table 2.5 shows that divorce rates among first-generation immigrants are very close to that of natives. But interestingly, it seems that among second-generation individuals, divorce rates are greater than that of French natives. For second-generation immigrants from Magrheb, for instance, men have a 4.9 percentage point more probability of being divorced once married than native French, and this probability is 4.2 higher for women. Along with the evidence of high endogamy rates among second-generation immigrants from Maghreb, this may suggest the existence of some cultural tension in the marriage model of Maghrebian communities, with some conservative elements (high marriage and endogamy rates) being challenged by elements of high cultural integration (educational gap, etc.), which may explain higher divorce rates.

## (p.59) 2.3.3 Inter-ethnic marriages

This section explores the frequency of inter-ethnic marriage. Table 2.6 reports the fraction of each community that is married to someone of a different immigration backgrounds. We distinguish three categories: a marriage with a native spouse, a marriage with a spouse who comes from the same country of origin, grouping together spouses from first and second generation, and marriages with non-native spouses coming from a different country of origin. We distinguish the exogamy rates among first and second-generation respondents.

The proportion of immigrants whose spouse or partner comes from the same country of origin (either first or second generation) is naturally higher for first-generation immigrants. The endogamy rates are equal to 74 per cent for first-generation Maghrebin, 69 per cent for first-generation African, 85 per cent for first-generation immigrants from Turkey or Middle East, and 79 per cent for first-generation immigrants from Asia. When we turn to immigrants from other European countries, the endogamous marriage rate is also higher than marriage rates with natives.

But as Table 2.6 shows, this endogamy is strongly reduced in the second generation: 23.4 per cent for South Europeans, 39.3 per cent

Table 2.6 Inter-ethnic marriages.

Country of origin

French native

Non-French Native—same origins

Non-French natives—different origins

First generation

Maghreb

21.67

74.29

4.05

Africa

26.83

69.16

4.01

Southern Europe

30.34

65.84

3.82

Northern Europe

45.21

44.25

10.54

Eastern Europe

38.89

53.88

77.23

Turkey

9.72

85.35

4.92

Asia

18.63

78.59

2.78

Second generation

Maghreb

41.06

53.40

5.54

Africa

52.40

39.35

8.24

Southern Europe

71.21

23.42

5.37

Northern Europe

85.27

6.16

8.57

Eastern Europe

72.48

16.16

11.36

Turkey

36.41

51.76

11.83

Note: Data source is the French Labour Force Survey (LFS) 2005–2007. Proportions are computed using individual sampling weights.

(p.60) for Africans, 51 per cent for Turkish, and 53 per cent for Maghrebins. Maghrebin immigrants of the first and second generation remain particularly endogamous, as compared to other groups. This is confirmed by regression analysis controlling for the individual characteristics aforementioned.

## 2.3.4 Spousal age gap

Table 2.7 reports estimates for the age gap between the spouse, which could capture a gender inequality. Immigrant women of the first and second generations do not seem to get married younger than French natives. Their age at the first child is not significantly lower than that of French natives, except for the first-generation immigrants from Maghreb, where the age gap is on average 2 years older than for native couples; and up to 3.6 years higher when both spouses share the same origin. The age difference between spouses is statistically different for first-generation immigrants from Maghreb, but not for the second generation. However, when one distinguishes endogamous couples (where both spouses come from the same country) from exogamous ones, the difference is persistent and statistically significant, even for second-generation immigrants (the age difference is about two years older than for French native couples)

Table 2.7 OLS estimates of the age gap between husband and wife for all individuals by ethnicity and place of birth.

Country of origin

All

Spouses of same origin

Spouses of different origin

First gen.

Second gen.

First gen.

Second gen.

First gen.

Second gen.

France

Reference

Southern Europe

−0.10

0.31

0.40

1.91

−0.80

−0.07

(0.28)

(0.29)

(0.37)

(1.24)

(0.54)

(0.47)

Maghreb

1.88***

−0.07

3.55***

1.81**

0.18

−0.59

(0.22)

(0.34)

(0.31)

(0.90)

(0.47)

(0.54)

Controls

Age, education, occupation

N

5,905

4,212

1,690

R2

0.032

0.05

0.02

Note: Histoire de Vie, INSEE 2003.

* Denotes statistical significance at the 10% level,

(**) at the 5% level, and

(***) at the 1% level.

.

# (p.61) 2.4 Educational attainment and gender gap in education

Another way immigrants are thought to be different from the French native is the level of education and the attitudes towards gender equality in education. We document these education patterns, focusing on the sample of individuals older than 26 years and who have left education. Table 2.8 reports education distribution and the gender gap in education for natives and immigrants.

## 2.4.1 Educational attainment

We first measure the gap in educational attainment of immigrants relative to French natives. We measure the evolution of this gap between different birth cohorts of immigrants and waves of immigration. We start by regressing the age they left full-time education on dummies for the country of origin of first and second generations. Native French are the reference group. The controls are a quadratic in year of birth, time dummies for the different waves of the survey, and region dummies.

Table 2.9 reports the educational gap for immigrant men relative to natives. The x-axis reports the coefficients for second-generation immigrants and the y-axis reports the coefficients for the first-generation immigrants. First-generation immigrant men from Africa, Northern Europe and Eastern Europe are one or two years older when leaving full-time education than their native counterparts, who themselves leave education when they are on average around 18.3 years old. First-generation immigrant men from Southern Europe and Turkey are on average three years and one year younger than native men, respectively, when they leave education, while immigrants from the Maghreb and Asia are of about the same age.

Table 2.8 Gender gap in age left full-time education.

Country of origin

Whole

First generation

Second generation

Natives

0.13

Maghreb

0.73

1.2

−0.3

Africa

2.46

2.3

2.0

Southern Europe

0.41

0.3

0.4

Northern Europe

0.95

0.8

0.6

Eastern Europe

0.82

1.30

0.9

Turkey

1.61

1.60

1.2

Asia

2.61

2.72

−1.9

Note: Data source is the French Labour Force Survey (LFS) 2005–2007. Proportions are computed using individual sampling weights. * Denotes statistical significance at the 10% level, ** at the 5% level, and *** at the 1% level.

(p.62)

Table 2.9 Age left full-time education.

Men

Women

First generation

Second generation

First generation

Second generation

Maghreb

−0.491***

−0.476***

−1.241***

−0.390***

(0.103)

(0.161)

(0.106)

(0.145)

Southern Europe

−3.285***

−0.733***

−3.084***

−0.731***

(0.128)

(0.134)

(0.119)

(0.128)

Africa

2.441***

3.252***

−0.443**

0.812

(0.207)

(0.891)

(0.195)

(0.744)

Northern Europe

2.083***

−0.166

1.439***

−0.254***

(0.248)

(0.454)

(0.210)

(0.380)

Eastern Europe

1.378***

−0.673**

0.066

−0.582**

(0.299)

(0.303)

(0.224)

(0.255)

Turkey

−3.172***

−0.396

−3.579***

−0.680

(0.311)

(0.586)

(0.325)

(0.567)

Asia

0.296

0.750

−0.905**

2.581*

(0.365)

(1.016)

(0.359)

(1.052)

N

51,219

56,311

50,446

54,603

Note: Data source is the Labour Force Survey (LFS) 2005–2007, the sample is all individuals aged 26 and above. These are the coefficients on dummy variables in a censored linear regression. The outcome variable is age left full-time education. The other covariates included are a polynomial in year of birth, region dummies, and time dummies. Sample aged 16–64 including students for which the dependent variable is top-coded at the current age. Reported standard errors are robust.

(*) Denotes statistical significance at the 10% level,

(**) at the 5% level, and

(***) at the 1% level.

From the first to the second generation, the gap in educational attainment relative to natives becomes negative for most immigrant groups. For instance, second-generation immigrants from Maghreb and Africa are 0.3 and 0.4 years younger when they leave the education system. Note, however, that the negative gap for Southern European men decreases from −2.9 years to −0.2 years from the first to the second generation.

Table 2.9 shows that only first-generation women from Northern and Eastern Europe are at least as old as native women when they complete their full-time education. All other groups are significantly younger than both native women and their male immigrant counterparts. Immigrants from Maghreb are almost one year younger, and immigrants from Southern Europe are three years younger. But there is an important improvement from the first to the second generation in terms of educational attainment, in particular among the groups which were the most disadvantaged in the first generation. Second-generation Asian women are performing outstandingly well, with an edge of 1.4 years of education relative to native French women. Second-generation women (p.63)

Table 2.10 OLS estimates of the gender gap in age left continuous full-time education by country of origin, wave of immigration and birth cohort.

Maghreb

Africa

Southern Europe

Northern Europe Caribbean

Eastern Europe

Turkey

French-born

Born before 1970

−0.24

−2.27***

−0.20*

−0.38

−0.98***

−1.28

(0.18)

(0.30)

(0.11)

(0.49)

(0.24)

(0.87)

Born after 1970

0.47***

−1.05***

0.50***

−0.14

0.90**

−0.27

(0.18)

(0.39)

(0.18)

(1.68)

(0.43)

(0.95)

Foreign-born

Born before 1970

−0.72***

−0.06***

−1.19***

−0.54***

−0.76*

(0.099)

(0.08)

(0.21)

(0.25)

(0.40)

Born after 1970

−0.47***

0.64***

−0.82

1.81***

−0.72

(0.30)

(0.24)

(0.57)

(0.83)

(0.49)

R2

0.075

0.100

0.034

0.037

0.073

Observations

11,963

2,209

10,594

2,206

2,361

1,381

Note: Data source is the Labour Force Survey (LFS) 2005–2007, the sample is all individuals aged 26 and above. Clustered standard errors at the individual level in parentheses.

(*) Denotes statistical significance at the 10% level,

(**) at the 5% level, and

(***) at the 1% level.

from Maghreb and Southern Europe also almost catch up their educational lag.

Table 2.10 provides a complementary picture of the evolution of the educational gap by distinguishing immigrants by birth cohorts. We focus on second-generation immigrants and compare the educational gap relative to natives among the young generation, born after 1970, and the old generation born before 1970. We run two separate regressions for the two different cohorts, taking the native as the reference group for each generation. Among natives, the average age they left full-time education is 20.67 years for the young generation against 17.83 years for the old generation, which represents a significant increase of almost three years between the two cohorts.

Relative to natives, the young second-generation immigrants are sometimes performing worse than the older cohort. Take the case of immigrants from Maghreb, who have an edge of 0.11 years among the old generation, and trail back by −0.45 years among the young generation. Naturally, this evolution does not mean that the younger cohort is less educated than the old one (in the particular case of immigrants from Maghreb, the younger cohort is educated for one year more than than the old one), but the gap relative to the natives has increased. The same is true for immigrants from Turkey. The evolution of the pattern of (p.64) female education by birth cohort is slightly different to their male counterparts. In general, the gap narrows among the young cohort, or remains fairly similar.

# 2.5 Female employment

We now turn to the analysis of female employment rate. The sample is made up of prime-age women between 25 years and 59 years old. For almost all ethnic groups, the employment rate is much lower relative to the native women, whose employment rate reaches 74.4 per cent. The employment gap is the most significant for foreign-born women from Maghreb, Africa, and Turkey, whose employment rate is 43.0 per cent, 53.9 per cent, and 20.0 per cent respectively. The difference is greater among married women with children than with single women.

The female employment rate increases significantly from the first to the second generation of immigrants. The employment rate of second-generation women immigrants from Maghreb increases by 16.6 points relative to first-generation immigrants. With married women immigrants from Maghreb with dependent children, the employment rate increases by 20 points from the first to the second generation.

Table 2.11 shows the estimates for the evolution of the employment gap between first and second-generation immigrants, controlling for age and education. The coefficients are the marginal effects from probit estimates on employment. The regressions are run on the whole female prime-age population between 25 and 59 years old, where French-native women are taken as the reference group. Among the first-generation immigrant, there is a statistically significant employment gap of female immigrants relative to natives. The gap reaches around 23–24 percentage points for female immigrants from Africa and Maghreb, and 41.5 for female immigrants from Turkey. The female employment gap remains sizeable and statistically significant among the second-generation female immigrants from those countries of origin.

# 2.6 Values and beliefs

## 2.6.1 National identity

In the survey ‘Histoire de Vies’, a series of questions were asked concerning the elements of the respondents’ identity. Table 2.12 (p.65)

Table 2.11 Estimates of the probability of being employed for women.

Country of origin

First generation

Second generation

France natives

Reference

Maghreb

−0.232***

−0.172***

(0.017)

(0.022)

Africa

−0.245***

−0.193***

(0.030)

(0.106)

Southern Europe

0.026

0.023

(0.018)

(0.022)

Northern Europe

−0.164***

−0.047

(0.040)

(0.091)

Eastern Europe

−0.218***

−0.015

(0.045)

(0.062)

Turkey

−0.415***

−0.334***

(0.042)

(0.110)

Asia

−0.198***

−0.114

(0.062)

(0.156)

Controls

Age, education

N

86,059

Note: Data source is the Labour Force Survey (LFS) 2005–2007. The sample is all female prime-age population between 25 and 59 years old. The coefficients are the marginal probit estimate, relative to native women. Clustered standard errors at the individual level in parentheses.

* Denotes statistical significance at the 10% level,

** at the 5% level, and

*** at the 1% level.

documents the result for national identity. If first-generation immigrants tend to have different attitudes and values, compared to French natives, this difference is largely attenuated for in second-generation immigrants. For example, the respondents are asked about their attachment to a particular country or continent: ‘Overall, do you feel mostly: from a French region, French, European, from another country, from another continent?’ Second-generation immigrants are more likely to declare that they feel French than the first generation. First-generation immigrants from Southern Europe are 50 per cent less likely to declare that they feel French than are French natives, controlling for age, gender, and education. This is particularly true of those who were born after 1970 (where the probability is reduced by 77 per cent). In the second generation, immigrants from Southern Europe are still 16 per cent less likely to declare that they feel French than native French. Those who were born after 1970 are three times less likely to declare that they feel French. By contrast, if first generation-immigrants from Maghreb are 28 per cent less likely to ‘feel French’, this effect is not statistically significant for second-generation immigrants from this region.

## (p.66) 2.6.2 Language and religion

Table 2.12 Estimates of the feeling of French identity.

Country of origin

All

Born after 1970

Born before 1970

First gen.

Second gen.

First gen.

Second gen.

First gen.

Second gen.

France

Reference

Southern Europe

0.133***

0.030

0.200

0.090**

0.127***

0.010

(0.022)

(0.017)

(0.154)

(0.044)

(0.022)

(0.018)

Maghreb

0.129***

0.002

0.138***

0.010

0.128***

0.001

(0.017)

(0.017)

(0.056)

(0.026)

(0.018)

(0.024)

Controls

Age, education, occupation

N

8,403

1,626

6,777

Note: Data source is Histoire de Vie, INSEE 2003. Marginal probit effects. Clustered standard errors at the individual level in parentheses.

* Denotes statistical significance at the 10% level,

** at the 5% level, and

*** at the 1% level.

Another key dimension of integration and identity is language. In the survey, the following question is asked: What language(s) did your parents usually speak when you were a child (around five years old)? The possible answers are: only French, another language, French and another language, two other languages. If the respondent answers that his parents spoke another language (including French), he is asked about this language, and whether he speaks in this language with his spouse, his children (who live in France), other adults living in the household and other adults living in the neighbourhood.

Table 2.13 shows the probit estimates of speaking in one’s foreign mother tongue with their relatives, controlling for age, gender, and education. Even among the second-generation immigrants, around 30 per cent of immigrants declare that they speak in their foreign mother tongue with their spouse, children, family, or their neighbours. The differences shown in Table 2.12 remain statistically significant for all migrants from South Europe and Maghreb, of the two considered cohorts.

Immigrants attach a high importance to the transmission of religion to their children. This religious attachment does not decrease from the first to the second generation of Maghrebin immigrants. Surprisingly, this attachment to religious transmission is more pronounced in the younger cohort of Maghrebins born after 1970. The proportion of (p.67)

Table 2.13 Estimates of the probability of speaking in one’s foreign mother tongue with spouse, children, family or neighbours.

Country of origin

All

Born after 1970

Born before 1970

First gen.

Second gen.

First gen.

Second gen.

First gen.

Second gen.

France

Reference

Southern Europe

0.672***

0.282***

0.485***

0.1653***

0.227***

(0.019)

(0.027)

(0.024)

(0.020)

(0.030)

Maghreb

0.431***

0.250***

0.846***

0.413***

0.364***

0.117***

(0.021)

(0.031)

(0.024)

(0.042)

(0.023)

(0.043)

Controls

Age, education, occupation

N

8,403

1,626

6,777

Note: Data source is Histoire de Vie, INSEE 2003. Marginal probit effects. Clustered standard errors at the individual level in parentheses.

* Denotes statistical significance at the 10% level,

** at the 5% level, and

*** at the 1% level.

immigrants of the first generation who declare that they have a religious practice is higher than that of French natives. This difference almost disappears for the second generation, except for Maghrebins, for whom this attitude remains statistically more pronounced, even in the younger generation of those born after 1970.

# 2.7 Conclusion

This chapter has compared a wide range of outcomes for immigrants relative to the natives in France. We have looked at fertility, marriage and divorce rates, inter-ethnic marriage, spousal age gaps, the gender gap in education, employment rates, national identity, religiosity, and language use. We find substantial heterogeneity across communities but also evidence that in almost all dimensions and for all groups, there is a fast integration process between first and second-generation immigrants. The rate of cultural and economic integration is faster for some variables than others. Religion, family arrangements, and endogamy show the slowest rate of convergence, in particular among immigrants from Maghreb. Second-generation immigrants from Maghreb also display a persistent employment penalty. Yet this slower assimilation process in religious and family arrangements does not go against a strong (p.68) feeling of French identity among the second-generation immigrants from Maghreb.

References

Bibliography references:

Adida, C., Laitin, D., and Valfort, M.A. (2010) Identifying Barriers to Muslim Integration in France. Proceedings of the National Academy of Sciences of the United States of America, 107(52), 384–390.

Aeberhardt, R. and Pouget, R. (2007) National Origin Wage Differentials in France: Evidence from Matched Employer-Employee Data. Crest Working Paper.

Algan, Y., Dustmann, D., Glitz, A., and Manning, A. (2010) The Economic Situation of First and Second-Generation Immigrants in France, Germany and the United Kingdom. Economic Journal. Royal Economic Society, 120(542), 4–30.

Algan, Y., Mayer, T., and Thoenig, M. (2011) The economic incentives of cultural transmission: Spatial Evidence from Naming Patterns across France. Sciences Po Working Paper.

Noiriel, G. (1988) Le Creuset Français: Histoire de l’immigration au XIX–XX ème siècle. Paris, Seuil.

Schnapper, D. (1991) La France de l’intégration. Paris, Editions Gallimard.

Silberman, R. and Fournier, I. (2007) Is French Society Truly Assimilative? Immigrant Parents and Offspring on the French Labor Market. In: A.F. Heath and S.Y. Cheung. Unequal Chances Ethnic Minorities in Western Labour Markets. Oxford, Oxford University Press.

Silberman, R. and Fournier, I. (1999) Les enfants d’immigrés sur le marché du travail: les mécanismes d’une discrimination sélective. Formation et Emploi, 65, 31–55.

Weil, P. (2005) La République et sa diversit é. Immigration, Intégration, Discriminations. La R épublique des idées. Paris,Seuil.

Weil, P. (2002) Qu’est-ce qu’un immigré? Histoire de la Nationalit é Française depuis la R évolution. Paris, Grasset.

Wieviorka, M. (1996) Une société fragmentée: Le multiculturalisme en débat. Paris, La Découverte, pp. 11–60.

## Notes:

(1) The authors would like to thank Vincent Tiberj (Sciences Po) for his helpful comments.

(2) France, Algeria, Tunisia, Morocco, rest of Africa, Asia (including Vietnam, Laos, Cambodia), Italy, Germany, Belgium, Netherlands, Luxembourg, Ireland, Denmark, Great Britain, Greece, Spain, Portugal, Switzerland, Austria, Poland, Yugoslavia, Turkey, Norway, Sweden, Eastern Europe, United States or Canada, Latin America, and other countries.

(3) France, Algeria, Tunisia, Morocco, Africa, Vietnam, Laos, Cambodia, Italy, Spain, Portugal, other northern and eastern European countries, Turkey, other Asian countries, America, all other countries.

(4) Note that this specification assumes that the birth cohorts and other regressors have the same effect for all country of origins.

(5) Each solution has its assets and its drawbacks. In the first case, we are compelled to look at older generations of immigrants, but we have a perfect picture of completed fertility rates. In the second case, we rely on functional form assumptions to control for the evolution of fertility with respect to age, but one can investigate more recent trends because of the inclusion of younger women. The reason we chose the first specification is that the EHF survey is specifically made for giving an accurate picture of completed fertility, whereas the use of Labour Force Surveys (such as in the UK study for instance) makes it difficult (not to say impossible) to observe completed fertility accurately.

(6) In our cross-sectional setting, the censoring point varies across observations. To deal with this issue, we use a censored median regression described by Chernozukhov.

(7) Marriage rate at age 25 is defined as the fraction of individuals being or having been married at age 25.

(8) Note that we therefore restrict the sample to individuals married or having been married. To control for possible censoring of younger individuals who may finally get divorced, we include a polynomial in age.