1On 1 January 2020, France had a population of slightly above 67 million people, of whom 14.4 million were minors (under age 18). The annual increase was again smaller than in 2019 but still mainly driven by natural growth (an excess of births over deaths) rather than positive net migration. Natural growth nonetheless fell to its lowest level in 40 years (+141,000).
2In 2018, the inflow of foreigners from third countries—whose nationals must hold a residence permit to remain in France—increased by 4.9% over 2017, reaching 249,474, the highest number since 2000. Roughly 1 in 10 incoming migrants were minors. The proportion of minors was highest among applicants for residence permits from Europe (they represent almost one-quarter of the total). Half of the minors entering France were from Africa, the region of origin that accounts for the majority of applicants for residence permits (almost 6 in 10 arrivals, on a slight upward trend since 2017). The share of migrants from Asia decreased slightly but remained high. Between them, these two world regions accounted for almost 85% of inflows in 2018. As in 2017, men and women were represented in almost equal numbers (50.8% men), and more than one-third of immigrants came to France for family reasons (37.6%). However, entries for employment reasons are increasing, accounting for 10% of admissions in 2018. Unlike previous years, the number of asylum seekers fell between 2017 and 2018. While most admissions in this category were for humanitarian reasons (65.3%), the proportion dropped by 3.6 percentage points in 1 year.
3The annual number of registered births (753,400) fell for the fifth consecutive year in 2019, due mainly to a decrease in the population of women of reproductive age. Fertility remained stable, with a total fertility rate of 1.87 children per woman in 2019, and France maintained its leading position in Europe. Mean age at childbearing continued to increase, reaching 30.7 years for all births and 28.8 years for first births. The pattern of age-specific fertility rates also reflects this trend towards postponement of births. Fertility behaviours also appear to be increasingly polarized, with a widening of the fertility differential between the most and least educated women over the last 15 years. A total of 4,111 infants (0.5% of total births) were born to women aged under 18 at the time of the birth. It was in the overseas departments that adolescent births were most frequent (1 in 20 births in French Guiana and Mayotte).
4The number of induced abortions increased in 2019 (230,000 vs. 224,000 in 2018). The total abortion rate was 0.58 abortions per woman in 2019, up from 0.56 in 2018. In 2019, there was one abortion for every three births, compared with one for every four births in 2018. However, while the number of induced abortions is increasing, a steady downtrend is observed for women under 18, who accounted for just 3.5% of the total in 2019.
5The numbers of marriages (234,735) and PACS unions (pacte civil de solidarité) (208,871) increased in 2018, although the number of PACS is gradually converging with that of marriages (a difference of 25,854). In 2018, the proportion of same-sex marriages (2.6%) fell to its lowest level since 2013, while a reverse tendency was observed for PACS unions (4.1% in 2018). Age at marriage continues to increase in France (35.4 years for women and 37.9 years for men). For same-sex couples, however, who marry later than different-sex couples, the mean age is decreasing. With the steady decrease across divorce cohorts in the probability of remarrying after a divorce, and the decline in numbers of widows and widowers, remarriage is becoming increasingly rare. In 2018, almost 1 in 3 marriages were between partners who already had at least one common child. Each divorce pronounced in France affects 1.33 children, of which 0.91 minors.
6Family situations vary considerably by age and sex, including for children, whose circumstances depend upon the partnership and residential situation of the parent(s) they live with. It is at around age 10 that children most often live with (half-) brothers and sisters.
7According to estimates from the French National Institute of Statistics and Economic Studies (INSEE), the number of deaths rose to 612,000 in 2019. This increase over 2018 is explained mainly by population ageing and by increased mortality among the large post-war baby-boom cohorts now reaching advanced ages. Life expectancy continued its steady increase, reaching 79.7 years for men (+0.2 years) and 85.6 years for women (+0.1 years). Cancers are now the leading cause of death, ahead of cardiovascular diseases, whose intensity is much lower in France than in the other countries of Europe. That said, standardized cancer mortality rates are also continuing to fall, for men especially, and for all types of cancer. For women, the trend is less favourable due to a steady increase in lung cancer deaths. These contrasting trends for men and women are the main reason behind the narrowing of the life expectancy gender gap. Under-15 mortality is very low in France. Deaths in this age group are concentrated in the first year of life, and even the first month. Contrary to many other European countries, infant mortality has stagnated in France for the last 10 years.
I – General trends and population age structure
1 – Natural increase at a historic low
8On 1 January 2020, the population of France was slightly above 67 million (67,063,703), including 2.16 million in the overseas departments and regions (i.e. 3.2% of the total, equivalent to the population of central Paris) (Papon and Beaumel, 2020). The population continued to increase in 2019 but at a historically low rate of +1.3 per 1,000 overall and +1.2 per 1,000 in metropolitan France (mainland France and Corsica),  down from 4.9 per 1,000 a decade earlier (Table 1, Appendix Table A.1 ). France gained just 86,000 inhabitants in a year, equivalent to a medium-sized town such as Versailles or Saint-Pierre de La Réunion. This gain contrasts with an average annual increase of 347,000 between 2005 and 2015, equivalent to the city of Nice. Total population change varies considerably across geographical regions. It is in the Caribbean departments and the Grand Est region of eastern France that the proportional decrease is greatest (−13.2 per 1,000 in Martinique, −11.7 per 1,000 in Guadeloupe, −12.7 per 1,000 in Haute-Marne, −11.0 per 1,000 in Meuse, and −10 per 1,000 in Ardennes). The departments with the most rapid growth are Corse-du-Sud (+10.2 per 1,000), Gironde (+9.9 per 1,000), Loire-Atlantique (+9.6 per 1,000), and Seine-Saint-Denis (+9.1 per 1,000), well behind Mayotte (+35.8 per 1,000) and French Guiana (+24.6 per 1,000) where the demographic transition is still ongoing. 
9Natural increase (the difference between births and deaths) is almost 3 times greater than net migration (difference between migrant inflows and outflows) and in 2019 still accounted for the largest share of population increase, although it decreased again in that year and is now well below 150,000 (+141,000). This historic low is due to both a decrease in births (crude birth rate of 11.2 per 1,000, the lowest rate ever recorded) and an increase in deaths (stable crude mortality rate of 9.1 per 1,000), although the former contributed 3 times more to the decline than the latter.
Indicators of population change
Indicators of population changen/a: not available.
* Provisional results, end 2020.
Coverage: Whole of France (including Mayotte since 2014).
10Net migration is difficult to estimate. ‘Apparent’ net migration, the difference between total population change and natural increase, has been negative since 2015 (−55,000  in 2019), while ‘adjusted’ net migration is positive (+46,000, Table 1). This adjustment has been made annually by INSEE since 2015 to ensure comparability between population numbers for successive years (Brutel, 2015; INSEE, 2020). It takes account of changes in the census dwelling form made in 2018 to facilitate identification of individuals with more than one residence and thus avoid double counts. This improvement, referred to as the ‘questionnaire effect’ or ‘adjustment’ by INSEE, has affected the population count. It will have an increasing impact until the results of the 2020 census are published (annual census surveys 2018 to 2022) (INSEE, 2020). As a result, the contribution of immigrants to French population growth in a given year changes over time. Each year, the adjustments made in previous years are corrected, sometimes quite considerably, until the final estimates are obtained after 5 years (duration of a complete census cycle). For example, net migration in 2013 was estimated at +47,000 in 2015 but was adjusted upwards to +100,000 in 2016 (Figure 1). In 2019, apparent net migration was +46,000 but was adjusted downwards by −101,000 to achieve consistency with the population estimates for 1 January based on census data (Papon and Beaumel, 2019).
Estimated annual net migration to France published since 2009 as adjusted by INSEE
Estimated annual net migration to France published since 2009 as adjusted by INSEECoverage: Whole of France (including Mayotte).
2 – A new ‘depleted cohort’ in the population pyramid?
11The base of the population pyramid began to narrow in 2011, and the process is continuing, albeit more slowly, with births decreasing annually by 15,000 on average between 2014 and 2018 but by only 5,600 between 2018 and 2019 (Figure 2). The year 2019 is probably the low point of a new ‘depleted cohort’ almost symmetrical to that of the 1990s that occurred 3 decades earlier. On 1 January 2020, the under-20s represented just 24.0% of the total population (23.7% in metropolitan France, Appendix Table A.2). The steady numerical decline in this age group again illustrates the process of population ageing and the narrowing of the pyramid at its base. The top is widening simultaneously, mainly because of increasing life expectancy: 26.6% of the total population is now over 60 and 9.6% over 75 (Appendix Table A.2).
Population pyramid of France on 1 January 2020 (numbers)
Population pyramid of France on 1 January 2020 (numbers)Coverage: Whole of France (including Mayotte).
12On 1 January 2020, France had 21,860 centenarians, most of whom were women (83%). With 32.6 centenarians per 100,000 population, France has the highest ratio in Europe. As the depleted cohorts born during the First World War reached their 100s during the 2010s, the ratio has remained relatively stable since 2011 (Teixera et al., 2017), but the number of centenarians is set to double over the next 10 years (Blanpain and Buisson, 2016).
13The COVID-19-related excess mortality in the second half of 2020 will affect the shape of the population pyramid, but without altering its overall structure. Under the worst-case scenario, the share of over-75s, which had been increasing steadily, may drop by 0.5 percentage points in 2020. 
3 – Family situations that vary across age groups
14Using census data on family relationships between people living in the same dwelling, we can determine the family structure of all French households (Figure 3). This distribution of family types is shaped by the family behaviours of adults, and it varies over the life course differently for women and men. Read from bottom to top, the pyramid reveals the ages when young adults leave the parental home (from age 18), form a union (up to age 28), have their children, in most cases with a partner (up to age 45), separate from their partner, experience their children’s decohabitation (up to age 60), and last, enter residential care (from age 80). Read horizontally, it reveals the differences between males and females. The pyramid is almost perfectly symmetrical before age 18, but as women leave home and form a union at younger ages than men, it becomes asymmetrical between ages 19 and 29. Beyond age 60, excess male mortality and the large share of women living alone lead to further dissymmetry.
Family structure pyramid in 2017
Family structure pyramid in 2017Coverage: Whole of France (including Mayotte).
15On 1 January 2020, children under 18 numbered 14.4 million. As children grow older, the share of those living in lone-parent families increases due to parental separation. But most minors live in households of the type ‘couple with child(ren)’ that include children living with both parents and those living in stepfamilies. Until 2018, these two categories could not be distinguished. The census bulletin has since been changed to make this distinction possible, but data from five annual census waves will be needed to obtain a final estimate of the proportions of each family type. According to provisional INSEE figures, 72% of minors living in a family in 2018 lived with both parents, 21% with a lone parent, and 7% with one parent and his or her partner (Algava et al., 2020). In the 1990 census, these proportions were respectively 85%, 10%, and 5% (Festy, 1994). Over time, the proportion of children living with both parents has fallen, while that of children living in a stepfamily and, even more so, in a lone-parent family, has increased. Shared-custody arrangements have become increasingly prevalent across cohorts, as shown by a recent analysis of tax returns: at age 10, the proportion of children concerned increased from 2.0% to 3.7% between 2000 and 2006 (Algava et al., 2019). Of course, these family arrangements and their evolution by sex and age vary across time and over geographical and social space, giving rise to divergent family norms, depending on whether the family lives in metropolitan France or an overseas department, for example, or on the social background of the household reference person. 
16Due to family recomposition and subsequent births, 80% of ‘children in a family’  live with a sibling, a half-sibling, or a stepsibling (Figure 4).  The size of this cohabiting sibship in the broad sense is greatest at age 10, when children are reaching the end of their primary school years. The position of this pivotal age is linked to birth timing (birth intervals of around 2–3 years), family size (rarely more than three children), and age at leaving the parental home.
Distribution (%) of children in families by age, family structure of the household, and number of children in families in 2017
Distribution (%) of children in families by age, family structure of the household, and number of children in families in 2017Coverage: Whole of France (including Mayotte).
II – Immigration from third countries
17This section describes recent trends in immigration  from so-called ‘third countries’ whose adult nationals must obtain a residence permit to live in France. It does not concern inflows from the countries of the European Economic Area (EEA) and Switzerland.  To ensure consistency of comparisons over time, the statistics are established for constant geographical areas. We do not count people of nationalities formerly required to hold a residence permit but now exempt. 
18Flows of third-country nationals arriving legally in France to establish residence in the country are estimated here from the statistics on long-term residence permits and visas valid as residence permits. They are based on data from the system used by the French Ministry of the Interior to track the status of foreigners residing in France (AGDREF) and transmitted annually to INED. The method developed by d’Albis and Boubtane (2015) is used to construct these flows. It applies the basic principle whereby people are counted in the flows of the year in which they receive their first residence permit valid for 1 year or more.  This year is generally the same as the year of entry, although in some cases it may be later (e.g. because the person previously held a more short-term residence permit). It is thus the entry into permanent migrant status (i.e. long-term legal residence) that is measured, rather than the physical entry into France. The inflows considered here cover the entire French territory, although large disparities are known to exist across different departments (Breton et al., 2017; d’Albis and Boubtane, 2018b).
19Because certain residence permits take some time to process, migration is only analysed to 2018 and does not cover the likely disruption caused by the COVID-19 pandemic.
1 – An upward trend in inflows
20Table 2 shows inflow data for the years 2013 to 2018. A total of 249,474 people received a residence permit in 2018, the highest since 2000 (Appendix Table A.3). Inflows in 2018 were 5% higher than in 2017 and 30% higher than in 2013.
Inflows of third-country nationals by first year of validity and period of validity of first residence permit of 1 year or more
Inflows of third-country nationals by first year of validity and period of validity of first residence permit of 1 year or moreCoverage: Permits granted in France and abroad to foreign nationals excluding citizens of the EEA and Switzerland (constant geographical area from 2013 to 2018). Permits granted in year n and recorded in the data extracted in July of the year n + 2. Permits of less than 10 years are valid for between 364 and 3,649 days; permits of 10 years or more are valid for more than 3,649 days.
21Among the individuals counted, the share of immigrants receiving a residence permit valid for 10 years or more remained low, at slightly above 11% in 2018. A residence permit valid for 10 years or more, typically a resident’s card, is generally granted after one or more permits of less than 10 years.
22Inflows of foreigners can also be estimated using other statistical sources. INSEE uses population censuses and notably a question on the place of residence in the previous year. According to Eurostat, which disseminates INSEE data, 255,185 foreign nationals entered France in 2018.  This total also includes nationalities not required to hold a residence permit (i.e. EEA countries and Switzerland) and, potentially, undocumented third-country immigrants. Applying the same geographical scope as that of Table 2 brings the total down to 172,057, a figure well below that obtained using AGDREF data. This underestimation is difficult to explain because no information is available on the method used to construct the series. Students may have been excluded by INSEE, even if they stayed in France for more than a year.
2 – Almost 60% of inflows from third countries are Africans
23Recent immigrants are young. People aged 18–34 represented 64% of all arrivals (Table 3) and 72% of all immigrants who were adults when they received their first residence permit. The share of minors increased slightly in 2018 (10.8%), reaching 27,059. Only minors receiving a residence permit are counted.  Foreign minors do not have to hold a residence permit but may need to obtain one if, for example, they wish to travel outside France. Minors born in France to foreign parents are not counted in the inflows. The first row of Table 3 thus only includes minors born abroad who hold a residence permit. These arrivals are detailed below.
Distribution (%) of inflows by age group, by first year of validity of first residence permit of 1 year or more
Distribution (%) of inflows by age group, by first year of validity of first residence permit of 1 year or moreCoverage: See Table 2.
24Figure 5 details flows by age and sex in 2018. The peak at age 18 is explained by individuals who arrived as minors and waited until age 18 to apply for a residence permit. Students also contribute to the large share of young people in the distribution. The distributions of women and men are similar overall, although the male distribution is more markedly bimodal. In 2018, the mean age at entry into France was slightly lower than in 2017, at 29.0 years for women and 28.2 years for men.
Distribution of inflows by age and sex in 2018
Distribution of inflows by age and sex in 2018Coverage: See Table 2.
25A large majority of incoming migrants (excluding those from the EEA and Switzerland) are African nationals. Their share has increased slightly, accounting for 59.2% of the total in 2018, the highest percentage since 2006 (Table 4; d’Albis and Boubtane, 2015). The share of arrivals from America increased slightly in 2018, while that of immigrants from Asia and Europe decreased. 
Distribution (%) of inflows by continent of origin, by first year of validity of first residence permit of 1 year or more
Distribution (%) of inflows by continent of origin, by first year of validity of first residence permit of 1 year or more*The total does not necessarily sum to 100 due to rounding and missing values.
Coverage: See Table 2. Turkey is included in Asia. Europe includes all European countries outside the EEA and Switzerland.
26As of 2017, women no longer account for the majority of arrivals. In 2018, they represented 49.2% of total inflows, down from 52.3% in 2014 (Table 5), marking a return to the levels observed in the early 2000s (Beauchemin et al., 2013; d’Albis and Boubtane, 2015). In 2018, there were fewer women than men among inflows from Africa and Asia, although they formed a large majority among inflows from all other continents, from the Americas (57.7%) and Europe (58.3%) especially. As observed below, the declining share of women from Africa coincides with a change in the reasons for granting residence permits, with a shift towards admissions for employment reasons.
Share of women in inflows (%) by continent of origin, by first year of validity of the first residence permit of 1 year or more
Share of women in inflows (%) by continent of origin, by first year of validity of the first residence permit of 1 year or moreCoverage: See Tables 2 and 4.
3 – A sharp increase in the share of admissions for employment reasons
27In 2018, 93,814 people were admitted to France for family reasons—still the main reason for admission—an increase of almost 5% in 1 year. As a proportion of total flows (37.6%), their level has stabilized, however, and remains well below that of 2013 (46.7%; see Table 6). The share of admissions for educational reasons (26.4% in 2018) remains stable with respect to 2017. While admissions for employment reasons account for just 10.5% of total flows (26,280 people), its share increased sharply over the observation period. In two-thirds of cases, the people concerned were in wage employment or non-wage employment. The remaining third were Passeport Talent  permit holders (6,543 people in 2018), whose numbers have increased substantially since 2016, and seasonal workers (2,177 people). Admissions for humanitarian reasons fell sharply in 2018, accounting for 12.8% of total admissions. They mainly concern two types of immigrants: foreigners who are ill (3,865 people in 2018) or admitted as refugees, stateless persons, or beneficiaries of territorial asylum or subsidiary protection (27,887 people).  Admissions for this second set of reasons fell by almost 8% in 2018, after rising rapidly for 2 years due to increased inflows of asylum seekers from 2014 (d’Albis and Boubtane, 2018a).
28Predictably, the reasons for admission vary by sex. Women are still over-represented among immigrants admitted for family reasons and under-represented among those admitted for humanitarian and, above all, employment reasons (Table 7). Among students, women are only slightly outnumbered by men.
Breakdown and distribution (%) of inflows by reason for granting first residence permit valid for 1 year or more, by first year of permit validity
Breakdown and distribution (%) of inflows by reason for granting first residence permit valid for 1 year or more, by first year of permit validityNote: Tables 6, 7, 8, and 9 have been updated from Breton et al. (2019), following clarification of the reasons provided by the Ministry of the Interior.
Coverage: See Tables 2 and 4. The ‘refugee’ line covers permits granted on the following grounds: refugee and stateless, territorial asylum, and subsidiary protection.
Share (%) of women among inflows by reason for granting first residence permit valid for 1 year or more, by first year of permit validity
Share (%) of women among inflows by reason for granting first residence permit valid for 1 year or more, by first year of permit validityCoverage: See Table 2 and the note in Table 6.
29Reasons for admission are distributed very differently from one continent of origin to another (Table 8). Family reasons are still the main reason for admission for immigrants from all continents. Admissions for this reason are over-represented among Africans (43.4% of permits in 2018 compared to 37.6% on average) and Europeans (38.4%) and under-represented among Asians (24.6%). Educational reasons are over-represented among Asians (29.2% vs. 26.4% on average) and Americans (29.6%) and under-represented among Europeans (10.4%). Humanitarian reasons account for a large share of permits granted to Asians (22.7% vs. 12.8% on average), notably due to the impact of the Syrian war, and to Europeans (20.4%), and a slight share among Americans (3.4%), for whom employment-related reasons are over-represented (13.4% vs. 10.5% on average).
Breakdown and distribution (%) of inflows by reason for granting first residence permit valid for 1 year or more, by continent of origin and first year of permit validity
Breakdown and distribution (%) of inflows by reason for granting first residence permit valid for 1 year or more, by continent of origin and first year of permit validityCoverage: See Tables 2 and 4, and the note in Table 6.
30The share of Africans admitted for family reasons remained practically stable in 2018 (−0.8 percentage points), while admissions for employment reasons increased sharply. Among Americans, the share of admissions for employment reasons increased, and admissions for other reasons remained stable. Among Asians, the share admitted for humanitarian reasons levelled off. Migration of Europeans for family reasons increased in 2018, while admissions for educational and humanitarian reasons (especially high in 2016 and 2017) declined.
4 – Slight decrease in asylum seekers admitted for residence in 2018
31Asylum seekers may be admitted for residence in France (i.e. receive a residence permit of 1 year or more) in several ways. If their application is accepted, they obtain a permit on humanitarian grounds and are counted as ‘refugees’ (Table 6, Row 4). Some of those whose application is rejected are admitted for residence on different grounds, most often family reasons. The rates of admission for residence by submission date of the asylum application are provided in d’Albis and Boubtane (2018a). The perspective here is different. Table 9 shows the annual inflows of people having submitted an application to the French asylum authority (Office français de protection des réfugiés et apatrides [OFPRA]). A total of 44,470 people applied for asylum in 2018, representing 17.8% of inflows, a slight drop from the peak observed in 2017 caused by the rise in inflows from Asia.
Inflows of asylum seekers by first year of validity of first residence permit valid for 1 year or more, and distribution (%) by sex, continent of origin, and reason for admission
Inflows of asylum seekers by first year of validity of first residence permit valid for 1 year or more, and distribution (%) by sex, continent of origin, and reason for admissionCoverage: Permits granted in France and abroad to foreign nationals who applied for asylum between 1985 and the first year of validity of the first residence permit valid for 1 year or more. Permits granted in year n and recorded in the data extracted in July of the year n + 2; see Table 3. See note in Table 6.
32Asylum seekers are not admitted solely on humanitarian grounds, so the total number of admissions is higher than the number admitted for this reason alone; 24% of asylum seekers in 2018 were admitted for family reasons. The share of women among incoming asylum seekers is lower than among overall inflows; they represented 36.5% of the total in 2018, a proportion similar to 2017. Since 2013, the largest share of incoming asylum seekers has been of Asian origin, representing more than 46% of the total in 2018. Africans accounted for 38.8%.
5 – Almost 1 in 4 Europeans entering France in 2018 were minors
33As indicated in Table 3, 10.8% of arrivals were minors (27,059) and received a first residence permit valid for more than 1 year in 2018. Among them, 47.5% were girls, a much lower proportion than the overall share of women in inflows (see Table 5). The AGDREF database does not specify whether these minors were accompanied by adults. 
34The majority of minors arriving in 2018 came from Africa, but their proportion among all minors (Figure 6A) was lower than the proportion of Africans among all inflows (Table 4). This was also the case for arrivals from America, while minors from Asia and Europe were over-represented. Almost 13% of all incoming minors were from Europe, yet Europeans represented less than 6% of total flows. These divergences are found in the share of minors in the flow from each continent (Figure 6B); the share of minors in European flows (23.4%) is much higher than that of Asians (12.4%) or Africans (9.3%).
Share of minors from each continent in overall flows of minors (A) and share of minors in overall flows from that same continent (B) in 2018
Share of minors from each continent in overall flows of minors (A) and share of minors in overall flows from that same continent (B) in 2018Interpretation: Figure 6A: 50.7% of minors who entered France in 2018 came from Africa. Figure 6B: 9.3% of Africans who entered France in 2018 were minors.
Coverage: See Tables 2 and 4.
35The geographical distribution of inflows of minors in 2018 is shown in Table 10. The departments receiving the largest numbers of minors were Paris and Seine-Saint-Denis. The Paris region (Île-de-France) alone received 33% of all minors entering France (and 35% of those entering metropolitan France). Elsewhere, the departments receiving most minors were Rhône and Bas-Rhin, while Lozère and Creuse received the fewest. In the overseas territories, most minors were received in Mayotte and French Guiana.
36The share of minors in inflows varies considerably across different geographical regions. While the national average was 10.8%, proportions were much higher (above 25%) in Haute-Marne, Gers, and Vosges, and lower (around 7.5%) in Paris, Hérault, and Gironde. The share was higher in overseas departments (14.2%) than in metropolitan France (10.7%). These differences are partly explained by the uneven distribution of inflows of foreign students. Because students are less likely to be accompanied by minors, the departments with large numbers of incoming students have smaller proportions of minors in overall flows.
Inflows of foreign minors by department of residence, 2018
Inflows of foreign minors by department of residence, 2018
III – Births and fertility
1 – A decline in births in 2019, set to continue in 2020
37The decline in births continued in 2019 (753,400 vs. 758,600 in 2018), and the estimates for the first 6 months of 2020 show no signs of a trend reversal. The year 2020 is likely to be the first in more than a quarter of a century with fewer than 750,000 births and only the fourth since the Second World War, alongside 1976, 1993, and 1994.  The annual number of births depends both on fertility rates (behaviour) and the number of women at various reproductive ages (structure). By summing age-specific fertility rates, we obtain the total fertility rate (TFR), which corresponds to the number of children a woman would have throughout her reproductive life if she behaved in the same way as all women in a given year (Table 11; Appendix Table A.4). This TFR has decreased very slightly (1.88 in 2018 and 1.87 in 2019) (Papon and Beaumel, 2020). The decrease in births is thus due mainly to variations in the number of women at each age between 15 and 50, and especially between 20 and 40, ages at which most births occur (84.4%). While the TFR is almost stable, mean age at childbearing is still increasingly slowly, at an annual rate of +0.1 years since 2010, and reached 30.7 years in 2019.  Since the mid-1970s, mean age at childbearing has been increasing in parallel with mean age at first birth (Volant, 2017), now close to 28.8 years.
Fertility by age group from 2013 to 2018
Fertility by age group from 2013 to 2018* TFR: total fertility rate (sum of age-specific fertility rates) expressed as a mean number of children per 1,000 women. Due to rounding, the total may differ slightly from the sum, and the variations may not correspond exactly to apparent differences.
(p) Provisional data.
Coverage: Whole of France (including Mayotte since 2014).
38In 2018, France still had Europe’s highest TFR, ahead of Sweden (1.76) and Ireland (1.75), where fertility has traditionally been quite high. Romania (1.76) and the Czech Republic (1.71), two countries that exemplify the recent fertility rebound in Eastern Europe (Pison, 2020) are reaching similar levels. At the same time, age at first birth in France remains lower than in other countries of Western and Southern Europe (Breton et al., 2019).
39The question is whether the current decrease in the French TFR marks the beginning of a decline in cohort fertility or a new and lasting change in fertility timing. The steady TFR downtrend over the last 10 years (Figure 7) reflects a decrease in rates before age 30 (as already observed up to the mid-1990s) that is not offset by an increase in rates at older ages, unlike the period before 2010.
40For the TFR to reach replacement levels (around 2.06 children per woman), fertility rates over age 30 would need to start increasing again, or more improbably, the fertility downtrend before age 30 would need to be radically reversed. Yet, except for Ireland and to a lesser extent Switzerland, the French fertility rate over age 30 is among the highest of all OECD countries (Figure 8). Ireland has a very atypical profile. With the rise in long-term unemployment observed since the economic crisis of 2008, fertility rates at young ages are now very low (Matysiak et al., 2020), while third births at later ages remain stable. In all these countries, age at first birth is considerably higher than in France (31.2 years in Italy, 31.0 in Spain, 30.9 in Switzerland and 30.5 in Ireland).  The high fertility rates at older ages in these countries are due to a larger rise in age at first birth.
2 – Slightly more than 4,000 births to adolescent mothers, one-third of which in the French overseas departments and territories
41In 2018, 1.5% of births (11,650) were to women who had not yet reached their 20th birthday (Papon, 2019). This proportion rises to 2.1% if exact age is used rather than age reached in the year (15,580 births). Among these adolescent births, 4,111 (0.54%, or 1 in 200 births)  were to women aged under 18.  While relatively low, this is well above the number of births to women aged 45 and over (2,291). As is the case in France, adolescent births are declining in absolute and relative terms across cohorts in most European countries but with different levels and rates of decline (Tomkinson, 2019). A large share of French adolescent births are registered in the overseas departments and territories (which account for 32.3% of adolescent births in France but just 5.1% of all births), in Mayotte and French Guiana especially (473 and 450 adolescent births, respectively, representing 1 in 20 births), but also Réunion (1 in 50 births, 300 in total). 
Age-specific fertility rates and TFR in France, 1990–2019
Age-specific fertility rates and TFR in France, 1990–2019Coverage: Whole of France from 1994 (including Mayotte since 2014) and metropolitan France (1990–1993).
Age-specific fertility rates in France and in 5 OECD countries with the highest mean age at childbearing (over 30.5 years) in 2018
Age-specific fertility rates in France and in 5 OECD countries with the highest mean age at childbearing (over 30.5 years) in 2018
42By applying a specific methodology to census data (Tomkinson and Breton, 2016), two main profiles of French adolescent mothers can be identified. In the first group, which represents a majority of cases, the mothers are inactive and have a low educational level. There is a strong correlation at regional level, in metropolitan France at least, between the proportion of minors not in education, employment, or training (NEET) among 16- to 17-year-olds and the frequency of adolescent births (R² = 0.82, Figure 9). Determining the direction of causality in this relationship is not straightforward, however. Are pregnancy and childbirth the cause or the consequence of dropping out of school? The second main group comprises adolescent mothers whose life events at entry into adulthood follow much the same pattern as that observed among most young people but are more concentrated in time: the birth occurs very shortly after leaving school, forming a union, and/or moving out of the parental home (Tomkinson et al., 2017).
3 – Cohort fertility: signs of a downturn?
43The women born in 1980, aged 40 in 2020, will have a completed fertility of slightly more than 2.0 children. France is a European outlier in this respect, although Ireland, Sweden, the United Kingdom, Belgium, and Denmark are close to this level, with completed fertility of between 1.9 and 2.0 children per woman (Breton et al., 2019). Cumulative fertility at different ages reveals the first signs of a further shift to later childbearing from the 1980s cohorts (Figure 10; Appendix Table A.5), but the very slight dip at age 32 for the 1985 cohort probably marks the end of the catch-up mechanism observed in previous cohorts.
Proportion of 17-year-olds not in education or employment and adolescent birth rates in the French regions, 2017
Proportion of 17-year-olds not in education or employment and adolescent birth rates in the French regions, 2017Note: The values for Réunion (982; 450), French Guiana (1,776; 1820) and Mayotte (2,012; 1,639) are outliers and not represented in the figure. Corsica is also omitted because of the very small number of registered adolescent births. Adolescent childbearing rates: for each region, ratio of adolescent births to the number of girls aged 17 (the majority of adolescent births occur in the year of the mother’s 17th birthday).
Coverage: Whole of France.
Cumulative fertility at different ages, 1960–1995 cohorts, metropolitan France
Cumulative fertility at different ages, 1960–1995 cohorts, metropolitan FranceInterpretation: For the 1985 cohort, cumulative fertility at age 32 is 135 children per 100 women versus 153 for the 1965 cohort at the same age.
Coverage: Metropolitan France.
4 – More large families among the least educated women, fewer among the most educated
44The number of children per woman is relatively stable across cohorts, but how has the distribution of women according to their completed fertility varied over time? Estimates based on census data (Box 1) show that while this distribution varies little between the 1967 and 1977 cohorts (Figure 12B), it conceals increasing social differentiation of behaviours (Figure 11). In France, as in many countries with low or lowest-low fertility, family size is linked to educational level (Davie and Mazuy, 2010; Van Bavel, 2012). Between the 1967 and 1977 cohorts, the share of low-educated women decreased (Figure 11, circle size); at the same time, education’s effect on the proportion of women with three or more children in the same household increased. With the expansion of education, the proportion of low-educated women has fallen, but their family size has increased. The share of low-educated women in the 1977 cohort with three children is 5.7 points higher than that of women with the same characteristics in the 1967 cohort. At the other end of the educational spectrum, the share of women with at least 2 years of higher education is increasing, while the proportion in this group with at least three children is falling (−3.7 points). These trends indicate greater polarization regarding fertility behaviour by level of education.
Percentage of women aged 40 living with at least 3 children in the 1967, 1972, and 1977 cohorts in France
Percentage of women aged 40 living with at least 3 children in the 1967, 1972, and 1977 cohorts in FranceInterpretation: In the 1977 cohort, 40% of low-educated women aged 40 were living with at least three children versus 34.3% in the 1967 cohort. In this 1977 cohort, 9.3% of women living with at least three children were low-educated (circle size) versus 17% (circle size) in the 1967 cohort.
Coverage: Population of women living in ordinary households and aged 40 in the census year.
Box 1. Using census data to estimate cumulative fertility at age x
Applying the same principle, the distribution of women of age x according to the number of cohabiting children provides an estimate of cumulative fertility at age x based on the standard method. The values obtained underestimate actual cumulative fertility (Figure 12A) because mothers’ and children’s deaths and the under-reporting of newborns in censuses are not factored in. Moreover, not all children live with their mother because of parental separation, placement in care, migration, or departure from the parental home (Tomkinson and Breton, 2016). This last factor explains why this indicator starts dropping off after age 40. The underestimation remains consistent from one census to the next and is lowest at ages 36–40 (between 9% and 12%). The term ‘children’ should be understood as ‘children in a family’ (see Note 7), i.e. the children whose mother is recorded as ‘family reference person’ or ‘partner of family reference person’. As some mothers are actually stepmothers (stepfamilies in which the father has custody of his children), the number of children attributed to a woman does not correspond exactly to her actual fertility. Likewise, women can be distributed by age and number of children in the family to obtain an estimate of net cumulative fertility (taking account of mortality and migration) at age x in a given birth cohort (Figure 12B). However, without reliable statistics on birth order, this indicator’s accuracy cannot be verified. The distribution is nonetheless consistent with the estimates obtained via the 2011 Family and Housing survey (Breton et al., 2019).
Cumulative fertility and number of cohabiting children in the household by age across cohorts, 2007, 2012, and 2017
Cumulative fertility and number of cohabiting children in the household by age across cohorts, 2007, 2012, and 2017
Distribution of women aged 40 by number of children in the household across cohorts, 2007, 2012, and 2017
Distribution of women aged 40 by number of children in the household across cohorts, 2007, 2012, and 2017Coverage: Metropolitan France.
5 – A situation at birth that varies across French regions
45Most children are recognized by both parents at birth, but less and less frequently through marriage (Toulemon, 2013). In 2019, 61% of children were born to unmarried parents,  setting a record for France, and for Europe, up from 53.7% 10 years ago and 42.7% 20 years ago. This trend may reflect a weakening of the norm of birth within marriage, still very strong and widespread in many societies. In France, couples now most often marry after the birth of one or more children. In the 2017 census, the proportion of children in families where the reference person (generally the father or mother) is married increases from 43.4% when the child is age 1 to 57.1% at age 10. By the time they finish primary school, the majority of children live with married parents.
46While the link between marriage and childbearing has weakened, other norms are more robust, such as giving the father’s surname to the child. For the last 15 years in France (law of 2005), parents have been free to choose either the father’s name only, as in the past, or the mother’s name only, or both names in any order (father’s followed by mother’s or vice-versa). All children born to the same parents must have the same surname (Mazuy et al., 2013). Very few children have the mother’s name only or the mother’s name followed by the father’s. In 2018, 91.1% of children were given the father’s name either alone or before that of the mother. This proportion has remained stable since 2014 (Table 12) and is even slightly higher than in 2012 (90.3%), the first year for which these figures were published (Mazuy et al., 2013). Double surnames are becoming more popular, however (9.0% in 2012, 10.3% in 2014, and 11.6% in 2018).
47Children born outside marriage were those who most often had a double surname in 2018 (15.6% vs. 5.4% of those born to married parents). Practically all children with only the mother’s surname are born to unmarried parents (1 in 10 children born outside marriage vs. 1 in 300 of those born within marriage) (Table 12).
48Given the link between parents’ marital status and the child’s name, the regions with the highest proportions of non-marital births should be those with the lowest proportions of children with the father’s name only. This is not the case, however, either in metropolitan France (R² = 0.05) or in the overseas territories (R² = 0.15) (Figure 13). The Paris region (Île-de-France), for example, has a low proportion of non-marital births and a high proportion of children bearing the father’s name only, while in Brittany, where the share of non-marital births is relatively high, the share of children with only the father’s name is identical to that of the Paris region.  The overseas departments and territories are outliers, with a very high share of non-marital births linked to specific local factors. In Mayotte, ‘customary’ marriage remains the norm (Marie et al., 2017); the French Caribbean has a long tradition of non-marital births (Mulot, 2013) in line with the ‘Caribbean model’; and Réunion has a situation between that of the Caribbean and that of metropolitan France. Except for Mayotte, the overseas departments and territories also have a low proportion of children with the father’s name only, perhaps because children with unmarried mothers are often born and raised in a lone-parent family. Such situations are extremely rare in metropolitan France (Marie et al., 2011).
Distribution of surname choices for children born in 2014 and 2018 by parents’ marital status
Distribution of surname choices for children born in 2014 and 2018 by parents’ marital statusCoverage: Whole of France.
Children’s surnames and non-marital births in France in 2018
Children’s surnames and non-marital births in France in 2018Coverage: Whole of France.
IV – Induced abortion
1 – More induced abortions in 2019 and new data sources
49Fewer births occurred in 2019 due to a reduction in the number of women of reproductive age and a slight fertility decline over the year (Section III). However, the number of induced abortions increased, reaching 230,000 for the whole of France (217,000 in metropolitan France, Appendix Table A.8), up from 224,000 and 209,000, respectively, in 2018. Abortion frequency (number of abortions per 1,000 women aged 15–49) therefore increased, from 15.6 per 1,000 to 16.1 per 1,000. The total abortion rate (mean number of abortions per woman, calculated as the sum of age-specific rates for a given year) was 0.58, up from 0.56 in 2018 (Breton et al., 2019; Vilain, 2019).
50The increase is observed primarily among women aged 25–29, the ages at which abortion is most frequent, so the impact on the total number of abortions is considerable. It also rose among women aged 30 and over, while it decreased among the under-25s. This increase in abortions does not appear to signal an upsurge in unwanted pregnancies, but rather a more frequent propensity to terminate a pregnancy than in previous years. This is shown by the abortion ratio (ratio of abortions to births), which reached almost 1 abortion per 3 births in 2019, up from 1 abortion per 4 births the previous year. This increase suggests that women and/or couples who have conceived but for whom the conditions for having a child (or another child) are unmet, more frequently decide to have an abortion, perhaps with the aim of bringing a future pregnancy to term (thus contributing to the secular increase in mean age at childbearing). According to data published by DREES,  abortion frequency is higher when living standards are low (Vilain et al., 2020). While the reasons behind this trend are multiple, the increasing hardship and insecurity affecting the most disadvantaged households may well be a contributing factor (Delmas et Guillaneuf, 2020). The number of people unemployed or in precarious employment  has fluctuated around 8 million since 2015 (Observatoire des inégalités, 2019). The DREES data we mention concerns 2016, and the trend may have amplified over the last 3 years. The issue is not so much reduced contraceptive coverage or less careful contraceptive behaviours, but rather an increasing propensity to terminate pregnancies in a more uncertain social climate.
51This increase should be viewed with caution, however, as new sources are now used to count the number of abortions in France (Box 2). The absolute number of abortions has nonetheless been trending upward since the 1990s, doubtless in response to an initial rise in the number of women of reproductive age, a lengthening of the period of sexual youth (Amsellem-Mainguy, 2019), and the growing complexity of emotional, sexual, and partnership trajectories involving more periods of transition in women’s sexual and emotional lives.
Box 2. More exhaustive abortion data
- Since 2005: reimbursement statistics (doctor’s fees or drug prescriptions) for non-hospital medical abortions, drawn from the general health insurance scheme (régime général) and the inter-scheme datamart (DCIR and DCIRS).
- Since 2009: medical abortions performed in health centres and family-planning or education centres.
- Since 2010: abortions covered by the health insurance schemes, Mutuelle sociale agricole and the Régime social des indépendants.
- Since 2014: abortions performed in healthcare settings (medical and surgical abortions performed in public and private hospitals) and entered in the PMSI medical statistics database (programme de médicalisation des systèmes d’information).
(a) These codes called GHM (groupes homogènes de malades, homogeneous patient groups) form part of the medical and economic classification system for medical, surgical, obstetric, and odontological hospitalizations.
(b)In 2019, for example, almost 2,800 abortions were potential double counts, one-third having been performed outside hospital then recorded in a hospital the following month.
52In the most recent years, the propensity to abort has increased, with an increase in the proportion of pregnancies that do not lead to a birth. From the 1990s, the proportion remained stable at 27 abortions per 100 births, but by 2019 it had risen 31 (Vilain et al., 2020).
2 – Low numbers of adolescent abortions
53The absolute number of abortions among women under age 18 is decreasing, from nearly 10,000 per year in the early 2010s to fewer than 8,000 in 2019, which may reflect improvements in contraceptive coverage and more effective use of contraceptive methods among young people. The proportion of sexually active adolescent girls using no contraception is minute (2.3% at ages 15–19, according to Baromètre Santé) (Rahib et al., 2017).
54The number of abortions among this population still exceeds the number of births (see below). Both are trending downwards, which suggests a decrease in the number of pregnancies and a lower frequency of abortion, perhaps due partly to the growing use of emergency contraception. The ‘day-after pill’ is most frequently used at ages 15–19. In 2016, more than 1 in 5 women in this age group reported having used it in the last 12 months (Rahib et al., 2018). In other age groups, demand for emergency contraception is generally low. A broader range of professionals (pharmacists, school nurses, and more recently, university medical services) are now authorized to prescribe it. The day-after pill is provided free of charge and anonymously in pharmacies for girls under 18.
55After an uptrend in the early 2000s, adolescent abortions have now fallen back to the levels observed in the 1990s. 
3 – More than 1 in 5 non-hospital abortions performed by a community midwife
56Under the health system modernization act of 26 January 2016 (Article 127), midwives are now authorized to provide abortion services. The law came into force on 6 June 2016, after the implementing decree detailing the practical rules of application was published in the Official Journal of 5 June 2016. Consequently, obtaining an abortion is becoming easier, and geographical inequalities are narrowing. In 2018, abortion services were provided by 248 midwives (Vilain, 2019). In 2019, almost 11,000 non-hospital abortions were provided by midwives, representing 5% of all abortions and 17% of non-hospital abortions.  Only medical abortions are provided by midwives; an extension to surgical abortions is under discussion. 
57The share of abortions provided by midwives varies considerably across geographical regions, ranging from 0% to more than 50%. The regions where the share is very low may be those where very few non-hospital abortions are provided, where community midwives are scarce, and/or where few community midwives are trained to provide abortions and are able to team up with a healthcare facility. This heterogeneity likely also reflects geographical differences in health- and social-care provision and in services to meet women’s sexual, contraceptive, and reproductive needs.
V – Unions
1 – Registered unions
Stable marriage numbers in 2018 and a continued increase in civil unions
58There were 234,735 marriages in the whole of France in 2018, 820 more than in 2017. This slight increase was temporary, however, as numbers started dropping again in 2019, falling to a record low of 227,000  (Papon et Beaumel, 2020), 7,700 fewer than in 2018 (−3.3%).
59Since 2011, the numbers of marriages and civil unions (pacte civil de solidarité [PACS]) have been converging,  and the continued increase in PACS unions in 2018 has narrowed the gap to below 30,000 (Figure 14A). The provisional estimate for the number of PACS unions is 209,000, up nearly 7% from 2017 (195,569). The number of PACS unions in 2019 is still unknown because the various entities that register them (municipal registry offices since November 2017 and notaries since 2011) have been slow to communicate the relevant statistics.
60For same-sex couples, marriages and PACS unions have been trending in opposite directions since 2014. The number of marriages fell to 6,386 in 2018 and 6,000 in 2019 (provisional data), while that of PACS unions rose sharply to 8,589 in 2018, up 17% from 2017 (Figure 14B). This rise continues the trend observed in November and December 2017 (Breton et al., 2019) and may be a temporary phenomenon linked to a change in the organization of PACS registration.  For the first time since 2013 and the legalization of same-sex marriage, PACS unions largely outnumbered marriages among same-sex couples (2,200 more PACS unions than marriages in 2018).
Annual numbers of marriages and PACS unions since 2000 by partners’ sex
Annual numbers of marriages and PACS unions since 2000 by partners’ sexNote: Provisional data for marriages in 2019 and PACS unions in 2017 and 2018.
Coverage: Whole of France (including Mayotte since 2014).
61The share of same-sex marriages among total marriages has continued to fall since 2014 and has reached an all-time low (2.6% in 2019). Conversely, same-sex PACS unions are trending upwards (Table 13), and their share has been increasing steadily since 2013; 4.1% of PACS unions were between same-sex couples in 2018. This is still below the level of 2012 (4.3%), the last year the PACS was the only form of legalized union open to same-sex couples. Clearly, freedom to marry has not led to a massive preference for marriage over the PACS among same-sex couples.
Number of unions (PACS or marriage) registered between 2013 and 2019 by partners’ sex
Number of unions (PACS or marriage) registered between 2013 and 2019 by partners’ sexNote: Provisional data for marriages in 2019 and PACS unions in 2017 and 2018.
Coverage: Whole of France, excluding Mayotte up to 2013 and including Mayotte from 2014.
62The total marriage rate  has been increasing slightly since 2016. It stood at 576 marriages per 1,000 men and 565 per 1,000 women in 2018 (Table 14). This increase is due solely to the slight rise in first marriages of different-sex couples (459 first marriages for men and 454 for women in 2018). The total marriage rate for remarriages has been trending downward since 2014, as has that of same-sex marriages for both men and women. The pronounced increase in PACS unions in 2018 presages an increase in the total PACS rate, which probably exceeded 500 per 1,000 individuals for men in 2018, and approached that level for women.
Number of unions (marriage, PACS) per 1,000 people by partners’ sex and marital status at union formation, 2014–2018
Number of unions (marriage, PACS) per 1,000 people by partners’ sex and marital status at union formation, 2014–2018Note: Dashes indicate data were not available for PACS unions in 2018.
Coverage: Whole of France.
An ongoing increase in age at marriage, except for same-sex couples
63While total marriage rates and first marriage rates  varied little between 2016 and 2018 (Table 14) due to the stability of marriage numbers over the period, age at marriage continued to increase. In 2018, the mean age at marriage for all marriages (including remarriages and marriages after age 50) was 35.4 years for women and 37.9 for men, up by 0.4 and 0.2 years, respectively, with respect to 2017. This increase is closely linked to the rise in mean age at first marriage of different-sex partners (+0.4 years for men and +0.3 years for women between 2017 and 2018). Conversely, age at first marriage of same-sex couples is moving in the opposite direction, at 36.7 years on average for women and 41.8 years for men (down 0.5 years between 2017 and 2018 for both sexes), and converging towards the ages at marriage of different-sex couples (Figure 15).
Mean age at marriage by spouses’ sex
Mean age at marriage by spouses’ sexNote: Mean ages determined from marriage rates at all ages.
Coverage: Whole of France.
64Between 2017 and 2018, fewer men under age 30 and fewer women under 28 married for the first time, but this decline was offset by an increase at older ages that contributed to a slight increase in the total first marriage rate in 2018, for men especially (Table 14). First marriages after age 40 accounted for 24.3% of the total first marriage rate in 2018, and those after age 50, 8.9%, up by 2.8% and 1.4%, respectively, since 2014. Similar increases are observed for women (+2.7% after age 40 and +1.4% after age 50), even though fewer women than men marry for the first time at a late age (18.7% after age 40 and 6.4% after age 50 in 2018).
A decline in remarriage, especially after divorce
65Remarriages of widows and widowers have been declining rapidly since the 1970s. The annual number in metropolitan France has fallen fourfold (from 10,000 in 1965 to around 2,600 in 2018), with widows and widowers accounting for just 1.2% of newlyweds in 2018 versus more than 3% in 1965. This trend is due mainly to the increase in age at widowhood linked to the rise in life expectancy at birth.
66Most remarriages are between divorcees, and increasingly so since the 1970s, due to the fall in marriages between widows and widowers and the increase in those between divorcees (Bonnet et al., 2019). In 2018, 94% of people who remarried did so after a divorce and just 6% after being widowed. The figures are quite similar for both sexes.
67Since the mid-2000s, fewer divorcees marry each year. From a high of around 50,000 in 2005, just 40,000 divorced women and 41,500 divorced men remarried in 2018. Since 2005, with the decrease in first marriages, the share of divorcees among newlyweds has now stabilized at around 19% for men and 18% for women due to the lower frequency of remarriage after divorce. For all divorce cohorts, the proportions who have remarried 5, 10, and 15 years later is decreasing very sharply across cohorts (Figure 16).
68Finally, while remarriages of divorcees are declining, numbers are still quite high. Among women who divorced in 2000, 1 in 3 was married again 15 years later, and among men the proportion was slightly higher (36.8%).
Total remarriage rates, per 100 divorces, by sex, year of divorce and time since divorce
Total remarriage rates, per 100 divorces, by sex, year of divorce and time since divorceNote: Total rate calculated for each sex with respect to the initial number of divorces.
Interpretation: Out of 100 men who divorced in 1990, 24 had remarried within 5 years.
Coverage: Metropolitan France.
One-third of couples who marry for the first time already have common children
69In 2018, one-third of the different-sex couples who married for the first time reported common children born before the marriage.  This proportion has remained stable in recent years, but it increases with the wife’s age, older women having had more time to conceive before marrying. Slightly above half of the women who marry for the first time at ages 35–44 have already had children with their husband (Figure 17, all years except 2010).
Percentage of couples with common children at the time of marriage* by wife’s age
Percentage of couples with common children at the time of marriage* by wife’s age*Different-sex marriages between never-married partners.
Coverage: Whole of France.
2 – Divorce and PACS dissolutions
Incomplete recent data due to new registration methods
70Little new information is available for divorces and PACS dissolutions in 2018. No figures have been published for PACS dissolutions, and following the new divorce procedures introduced in 2017, divorce data are very incomplete. Divorce by mutual consent can now be registered before a notary without going to court.  The number of divorces pronounced in court fell sharply in 2018 to 62,321, slightly below half the number recorded in 2016 before the reform  (128,043, Table 15). As most divorces are by mutual consent, it is not surprising that many are settled out of court, and statistics on these divorces are not yet available. Data for 2017 and 2018 are still incomplete. 
Divorces and PACS dissolutions in France
Divorces and PACS dissolutions in France(a) Direct divorces and separations converted into divorces.
(b) By virtue of law no. 2016-1547 of 18 November 2016, divorces can also be registered by a notary. These divorces are not included in this table, so the figures for 2017 and 2018 are incomplete.
(c) Provisional data.
71Since 2007, annual numbers of divorces have been trending downward (below 130,000 since 2012), and the total divorce rate has remained relatively stable since 2006, at around 45 divorces per 100 marriages (Table 15).
72Given the steady rise in new PACS unions each year since 2011, the increase in PACS dissolutions is expected to continue. This increase is resulting in a higher number of couples in recent PACS unions, with higher frequency of dissolutions (Breton et al., 2019). For the first time since the PACS union was introduced, the provisional figures for 2018 nonetheless show a slight decrease in dissolutions (82,345, down 2,300 from 2017), but this may be due to a data updating problem.  The data for 2018 and 2019 to be issued shortly should clarify this point.
73Between 50%–55% of PACS unions registered between 1999 and 2007 had been dissolved by the end of 2017 (Breton et al., 2019). Some of these couples dissolved their PACS in order marry (1 in 4 couples who entered a PACS in 2016), while for others the dissolution marked the end of the relationship (between 25% and 36% across cohorts).
Divorce affected 170,000 children in 2016, of whom 116,000 were minors
74According to figures issued by the Ministry of Justice, among the 169,830 children whose parents divorced in 2016,  around two-thirds were minors (115,945). With the relative stability of divorce numbers over the last decade, the number of children concerned fluctuates around 120,000 each year, with no significant change since the drop observed in 2013 (Mazuy et al., 2015b). In 2016, each divorce affected 0.91 minors on average and 1.33 sons and daughters of all ages. These figures have remained stable since 2010. They increased to 0.93 and 1.40, respectively, in 2017, but this change is difficult to interpret pending the publication of statistics on divorces pronounced by notaries.
75In 2016, 52.4% of divorces affected at least one child, most often an only child (29,848; 23% of divorces) or two children (27,638; 21.6% of divorces). There were fewer divorces among couples with at least three children (7,789 with three children and 1,178 with four or more, representing 6.1% and 1.4% of divorces, respectively). Divorce most often concerns children when the woman is aged 35–45 (Figure 18). Three-quarters of all divorces involving children occur in this age group, and the mean number of minors concerned is highest at these ages (1.5 minors per divorce). Women aged over 45 at the time of divorce most often do not have a child under 18 in the household. In 2016, more than 30,000 divorces—almost a quarter of the annual total—concerned women aged over 50 with no children under 18. These are divorces of couples with adult children or of those who never had children. However, the share of divorces among couples with children has increased in recent years for women aged 45–60, reflecting the rise in age at childbearing among the cohorts of women born after 1960. Married women over age 50 now more often live in households with children under age 18.
Divorces in 2016 by woman’s age and number of children under 18
Divorces in 2016 by woman’s age and number of children under 18Coverage: Whole of France.
VI – Mortality
76According to provisional INSEE figures, 612,000 deaths occurred in the whole of France in 2019, up from 610,000 in 2018. As in previous years, this uptick does not reflect a deterioration of population health, but rather an increase in mortality among the large post-war cohorts now reaching advanced ages (Breton et al., 2019).
77In 2020, however, the number of deaths is expected to rise because of the COVID-19 epidemic. On 2 December 2020, Santé Publique France had recorded more than 53,000 COVID-related deaths (Santé Publique France, 2020). By comparison, the annual death toll of flu epidemics in previous years was 14,000 at most. For now, it is impossible to determine or even estimate the final impact of the COVID-19 epidemic on mortality in France. The deferred effects of the epidemic—long-term health consequences for survivors, collateral effects of a healthcare system stretched to breaking point, economic and social impact of lockdowns on the general population—will need to be taken into account. The most optimistic scenario assumes that the epidemic will have no lasting effects and that mortality trends will resume their secular pattern after the crisis, as was the case after the 1918–1919 flu pandemic, the deadliest of the 20th century.
78According to current estimates, excess deaths in the first wave of the COVID-19 pandemic (estimated at 30,000 between March and July 2020) will reduce life expectancy at birth by 0.2 years for men and 0.1 years for women (Guillot and Khlat, 2020). This decrease corresponds exactly to the gain recorded between 2018 and 2019, when life expectancy at birth in metropolitan France rose from 79.6 to 79.8 years for men and from 85.6 to 85.7 years for women (from 79.5 to 79.7 years for men and from 85.5 to 85.6 years for women in the whole of France). While the COVID-19 pandemic raises many new questions, it is still useful to understand the epidemiological context in which it occurred, and which is described in this section.
1 – Limited progress below age 15
79INSEE’s 3-year life tables show that the probability of surviving to age 15 increased very slowly over the 10 years between 2006–2008 and 2016–2018 (Table 16), unlike previous decades. While the probability of dying at ages 0–14 fell by a third for both sexes between 1986–1988 and 1996–1998, and by a quarter between 1996–1998 and 2006–2008, it decreased by less than 10% for males and 5% for females between 2006–2008 and 2016–2018.
80After the first year of life, the probability for a child of dying before their 15th birthday is now negligible (1 per 1,000), so the potential for further progress is very limited, although there is still scope to reduce accidental deaths (mainly avoidable road traffic accidents) which are the main cause of death at ages 1–14. However, with 3.6 infant deaths per 1,000 births in metropolitan France (3.8 per 1,000 in the whole of France), little progress has been made in reducing infant mortality over the last 10 years. It has even increased slightly since 2013–2014, when it stood at 3.4 per 1,000 in metropolitan France and 3.5 per 1,000 in the whole of France (Appendix Table A.11), unlike many European countries where it has continued to fall. It is now below 3.0 per 1,000, and in some cases 2.0 per 1,000 in several northern countries (Estonia, Finland, Iceland, Slovenia, and Sweden, followed closely by Norway; Table 17).
Probabilities of dying by age group in France, by 10-year period between 1986–1988 and 2016–2018*
Probabilities of dying by age group in France, by 10-year period between 1986–1988 and 2016–2018** Provisional data for the 2016–2018 life table.
Coverage: Metropolitan France.
81Not all components of infant mortality are affected in the same way by this stagnation (Figure 19). The post-neonatal mortality rate (deaths between the end of the first month of life up to the first birthday) stabilized at around 3.7–3.8 per 1,000 in the 1980s, before falling sharply in the mid-1990s, thanks notably to the success of measures to reduce sudden infant death syndrome (Barbieri, 1998). It has since halved again, falling from 2.0 per 1,000 in 1995 to 1.0 per 1,000 in 2018. The neonatal mortality rate (deaths in the first month per 1,000 live births) fell steadily until the mid-2000s, then more slowly up to 2011, reaching a low of 2.2 per 1,000 before rising again to 2.7 per 1,000 in 2017 (2.6 per 1,000 according to provisional figures for 2018). The reasons for this upturn remain unclear (Papon, 2018). One possibility is that thanks to advances in neonatal care, severely premature infants who would previously have been stillborn now survive for a few hours or days after birth. This hypothesis is consistent with the rising share of perinatal illnesses (which account for 2.0 per 1,000 deaths), in causes of infant death. Another possible explanation is the steady rise in high-risk births linked to the growing number of multiple births and the increase in age at childbearing.
Infant mortality rate in Europe, 2009–2018, per 1,000 live births
Infant mortality rate in Europe, 2009–2018, per 1,000 live births
82The long-term trend shows that mortality is increasingly concentrated in the first days of life. The share of neonatal mortality in total infant mortality has risen from 55% in the early 1980s to 70% in the late 2010s. A quarter of deaths before age 1 occur during the first day, a quarter during the following 6 days, another quarter during the following 3 weeks, and the last quarter between the second and twelfth month. That said, infant mortality is now so low in absolute terms that its impact on life expectancy at birth is small compared to that of other ages, for which mortality is still decreasing.
Absolute and relative trends in infant mortality and its components, 1980–2018
Absolute and relative trends in infant mortality and its components, 1980–2018Coverage: Metropolitan France.
2 – A steady decline in mortality among adults below age 45
83The downtrend in mortality between 2006–2008 and 2016–2018 was especially pronounced at ages 15–25. The probability of dying fell by almost one-third for males (from 6.3 per 1,000 to 4.5 per 1,000) but much more slowly for females (from 2.2 per 1,000 to 1.8 per 1,000), at a pace comparable to that of previous decades. Male excess mortality is at its maximum in the 15–25 age range, with a probability of dying that is 3 times higher for males than females. Compared with the previous period, the slower decline in female mortality has gradually narrowed the gap, however, as the male-to-female probability ratio stood at almost 3.5 in 1986–1988 (Figure 20). Mortality has also declined substantially at ages 25–45 (−20%), at a pace similar to that of the 2 preceding decades, and the gender mortality gap has remained stable, with ratios of probabilities ranging between 2 and 3 in favour of women.
Excess male mortality by age in 1986–1988, 1996–1998, 2006–2008, and 2016–2018
Excess male mortality by age in 1986–1988, 1996–1998, 2006–2008, and 2016–2018Note: The graph represents the male-to-female ratio of probabilities smoothed over 3 years of age, except at age 0.
3 – A slower decrease in mortality after age 45
84At ages 45–64, progress over the last decade continued at a similar pace as in the preceding decade for males, while for females it slowed considerably. The probabilities of dying fell by almost 20% for males and 15% for females between 1986–1988 and 1996–1998; by 16% and 9%, respectively, between 1996–1998 and 2006–2008; and by just 15% and 7% between 2006–2008 and 2016–2018 (Table 16).
85At ages 65–80, mortality fell by 15% for males in the last decade, a decrease identical to that observed at ages 45–65, while for females the decrease was slightly stronger (11%). Over age 80, the mortality decline between 2006–2008 and 2016–2018 was very limited (−1% for males and −2% for females), and the gender gap narrowed steadily with age, becoming negligible at around age 100.
4 – A growing contribution of old-age mortality to life expectancy at birth
86Over time, mortality below age 65 has become very low. More than 85% of males and 92% of females lived to this age in 2018, compared with 74% and 88%, respectively, 40 years earlier. Today, the increase in life expectancy at birth is increasingly determined by mortality at advanced ages (Figure 21). Over the period 2006–2008 to 2016–2018, 50% of the years of life gained by men and 70% of those gained by women were the result of progress achieved after age 65. This proportion reflects an acceleration of progress in this age group over time. Between 1986–1988 and 1996–1998, its contribution to years of life gained was 40% for men and 63% for women. Examining changes in the age structure by medical cause of death can help to shed light on this trend.
Contribution of age groups to life expectancy gains in years from 1986–1988 to 1996–1998, from 1996–1998 to 2006–2008, and from 2006–2008 to 2016–2018
Contribution of age groups to life expectancy gains in years from 1986–1988 to 1996–1998, from 1996–1998 to 2006–2008, and from 2006–2008 to 2016–2018
5 – Mortality trends by cause of death
87The distribution of deaths by medical cause of death is published by the National Institute for Health and Medical Research (INSERM), the body responsible for coding all the death certificates issued in France and for disseminating the information obtained. In 2020, however, the COVID-19 epidemic disrupted the routine activities of the Centre d’épidémiologie sur les causes médicales de décès (CépiDc), the laboratory that performs these coding operations. At the time of writing, data are available only for deaths occurring up to 2016, as was already the case in 2019.
Remarkable progress in cardiovascular disease management
88The decrease in cardiovascular disease mortality has been the main driver of progress in life expectancy since the 1970s. Its impact has increased over time, accounting for almost half (45%) of the years of life gained for males and almost three-quarters (70%) of the years gained for females between 2006 and 2016. This situation reflects a long-term trend stemming from more effective management of these diseases by comparison with the other main causes of death (Figure 22). Substantial progress has been achieved in ischaemic heart diseases, whose standardized mortality rate fell from 58 deaths per 100,000 for men and 22 per 100,000 for women in 2006 to 39 and 13 per 100,000, respectively, in 2016 (Appendix Table A.13). Deaths from cerebrovascular diseases also declined sharply, with rates of 25 and 18 per 100,000, respectively, in 2016 versus 35 and 25 in 2006. The progress attributable to an overall decline in diseases of the circulatory system slowed, however. Between 1986 and 1996, and again between 1996 and 2006, more than a year of life expectancy was gained for both sexes, but between 2006 and 2016, this gain was just 0.92 years for men and 0.79 years for women.
Standardized mortality rate by sex and group of causes of death, 1980–2016
Standardized mortality rate by sex and group of causes of death, 1980–2016Coverage: Metropolitan France.
Cancer, the leading cause of death in France
89Cancer has been the leading cause of death in France since around 1990 for males and 2000 for females. Male and female cancer mortality trends are very different. Steady progress has been achieved for males, with a 20% decrease in the standardized mortality rate between 2006 and 2016 (from 251 to 208 per 100,000). For females, standardized rates have fallen very little (from 123 to 117 per 100,000). While the decline in cancer mortality increased male life expectancy at birth by 0.76 years over the last decade, the lack of progress for females is reflected in a gain of just 0.16 years. These overall trends conceal highly contrasting tendencies across different tumour sites, however (Figure 23).
90For males, deaths from all the deadliest forms of cancer, excepting cancers of the hematopoietic system (blood cancers), have been declining over the last 40 years (since 1990 for prostate cancer) under the combined effect of lower alcohol and tobacco consumption and progress in screening and treatment. A similar trend is observed for women, except for deaths from uterine cancer, which have stagnated since the early 2000s and, above all, lung cancer, which has been increasing steadily for 4 decades. While lung cancer is still the most deadly cancer for males, the standardized mortality rate fell from 63 per 100,000 in 2006 to 52 in 2016. For females, it is set to become the leading cause of cancer death, ahead of breast cancer. While the standardized mortality rate for breast cancer fell from 25 to 22 per 100,000 between 2006 and 2016, it rose from 14 to 18 per 100,000 for lung cancer over the same period, largely cancelling out the gains achieved for other types of cancer.
Standardized mortality rate by cancer site and by sex, 1980–2016
Standardized mortality rate by cancer site and by sex, 1980–2016Coverage: Metropolitan France.
Favourable trends for most of the other well-defined categories
91The decrease in deaths from external causes also contributed to the increase in life expectancy at birth for both sexes, though much more so for men (+0.31 years, representing 15% of the total) than for women (+0.12 years, 11% of the total). The fall in road deaths was especially beneficial for men regarding years of life gained, lowering the standardized mortality rate from 12 to 7 per 100,000 between 2006 and 2016, compared with a drop from 3 to 2 per 100,000 for women, for whom the rate has become negligible. Suicide rates also fell from 24 to 19 per 100,000 for men and from 8 to 5 per 100,000 for women.
92Given their relatively small contribution to overall mortality, the other groups of diseases (notably infectious, respiratory, and digestive diseases) had a much smaller impact on progress in life expectancy at birth. For some diseases, the standardized rates nonetheless decreased substantially, as was the case for all digestive diseases, whose mortality rates fell by 25% for both sexes, but which contributed just 6% to gains in life expectancy at birth for males and 7% for females. After stabilizing in the 1990s and then declining in the early 2000s, respiratory disease mortality stagnated for all age groups, reaching a ‘plateau’ for men and showing a renewed upturn for women. The COVID-19 epidemic will inevitably exacerbate this trend in 2020—and in 2021 if the disease is not brought rapidly under control.
A convergence of male and female mortality rates due largely to cancer deaths
93Analysing the contribution of causes of death to the gains in male and female life expectancy at birth helps to explain why the gender gap has narrowed. One can do so by calculating the contribution of each age group and each category of causes of death to the gender difference in life expectancy in 2006 and 2016. The difference between contributions for each age group and each group of causes can then be represented on a graph (Figure 24) to reveal the difference between male and female life expectancy gains. The age groups and cause-of-death categories that helped to narrow the gender gap have positive values, while those that widened it have negative values.
Contributions of age groups and cause-of-death categories to the narrowing of the gender gap in life expectancy at birth between 2006 and 2016
Contributions of age groups and cause-of-death categories to the narrowing of the gender gap in life expectancy at birth between 2006 and 2016Coverage: Metropolitan France.
94The results confirm the key role of cancers, which explain 42% of the male life expectancy gain (+1 year) with respect to females. The contribution of cancers to the narrowing of the gender gap mainly concerns adults aged 45–85. The faster decrease in male deaths from circulatory diseases also contributed significantly (explaining around one-third of the convergence). Among young adults aged 20–45, deaths from external causes contribute most to the convergence, with males benefitting more than females from the decrease in road deaths. At advanced ages (beyond 85), the gender gap continues to widen for all causes of death.
95The appendix tables are available on INED’s website and on the Archined open archive:
These rates are different from those published in INSEE’s demographic report (Papon and Beaumel, 2020) which take account of natural increase (2.7 per 1,000) and net migration (2.8 per 1,000) but are not adjusted for migration.
These tables are available on INED’s website and on the Archined open archive: http://hdl.handle.net/20.500.12204/AXWs9WivkgKZhr-blhHr
Increases calculated by the authors using the population estimates on 1 January 2020 (series by region, department, sex, and age from 1975 to 2020), published by INSEE.
The figure of −55,000 corresponds to (population increase between 1 January 2019 and 1 January 2020) − (natural increase 2019).
Under the worst-case scenario, the number of excess deaths is double that observed by INSEE between 2 March and 19 April 2020, and all these additional deaths are among people aged 75 and above (Gascard et al., 2020).
The family structure pyramids of these different subpopulations are not presented here, but they reveal large differences in the proportions of children in lone-parent families or of older adults in residential care, for example.
Under the census definition, a ‘household’ comprises all the people living in one dwelling. It may include one or more families (all people in the household linked by union or filiation) or no families. ‘Children in a family’ are defined as all children in a household with at least one parent living in the same household. According to census data, 97.1% of children with a completed age of 0–17 were in this situation in the 2017 census, with a maximum of 98.0% at age 0 and a minimum of 93.6% at age 17. Those not in families are either living outside private households (in a collective dwelling; Mainaud, 2011) or are already the reference person of a household (possible over age 15).
Stepsiblings are defined here as children with no shared parent but whose respective parents are in a cohabiting union. In most cases, they were born to the two partners in previous unions. This relationship has no legal basis, it is a de facto situation.
Immigrants are defined as persons born outside France to non-French parents, whether or not they subsequently acquire French nationality.
Austria, Belgium, Bulgaria, Croatia, Cyprus, Czech Republic, Denmark, Estonia, Finland, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Liechtenstein, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, and United Kingdom.
The nationalities considered may vary from one demographic report to the next in response to legislative changes in rights of residence. Appendix Table A.3 takes account of changes in scope.
The Ministry of the Interior also publishes a complementary series of migration flow statistics based on a count of all first residence permits issued to adults. Its scope is different; it includes residence permits valid for less than 1 year which will not necessarily be renewed by a longer-term permit. It thus includes cases of temporary migration.
Data drawn from http://ec.europa.eu/eurostat/en/data/database, accessed on 15 September 2020.
In most cases, this permit is called a document de circulation pour étranger mineur. It was instituted by a decree published on 24 December 1991.
Date from which immigration can be analysed using AGDREF data (d’Albis and Boubtane, 2015).
Permit granted to people who are highly qualified or who wish to set up a company or invest in France, and to artists and performers.
Admissions for humanitarian reasons only include people whose asylum application has been processed and approved, so this figure does not include all asylum seekers.
According to the French Ministry of Justice, 17,022 unaccompanied minors were recorded in 2018, half of whom came from Mali and Guinea (Ministry of Justice, 2020). These figures are not comparable because some minors (accompanied or otherwise) do not have a residence permit.
At the time of writing, the provisional number of births is known up to June 2020. The correlation between the number of births reached at the end of June and the final number of births is very strong (R² = 0.983), so if this remains the case in 2020, a total of 748,000 births should be recorded in France. Any effect of lockdown on birth numbers will not be visible until early 2021.
Mean age at childbearing is calculated from the fertility rates of a given year and not from births, so it does not depend on population age structure.
Among OECD countries, South Korea and Japan also have a ‘late’ fertility regime but with much lower fertility rates beyond age 30.
Under French law, becoming a parent is not a valid reason for a judge to emancipate a minor, unlike a marriage which may be authorized before reaching majority under exceptional circumstances, such as a pregnancy, and which automatically leads to emancipation. Paradoxically, minors who become parents automatically acquire ‘parental authority’ over their child while remaining under the authority of their own parents (Bisleau, 2012).
A further 963 births were registered to mothers aged over 18 at the time of birth but under 18 when the child was conceived (births recorded before October of the current year).
These figures take account of the higher share of minors aged 15–17 in completed years, notably in French Guiana and Mayotte.
This is a provisional figure published on the INSEE website. The detailed statistics on births in 2019 will not be available until June 2021.
The weakness of the correlation is due partly to the existence of a ‘confounding variable’, i.e. the proportion of births to couples comprising at least one foreigner. In metropolitan France, the higher the proportion of children with a foreign parent, the lower the proportion of non-marital births (R² = 0.86 in 2018), which explains both the small share of non-marital births and the widespread use of the father’s name only. In the Paris region, 41% of children have at least one foreign parent (the highest share in metropolitan France) versus just 11% in Brittany (the lowest).
Health and abortion data were matched with information from the permanent demographic sample (Vilain et al., 2020) to obtain social information on the women concerned via their tax returns.
This includes people on fixed-term contracts, temporary contracts and work experience contracts, and inactive people who want to work (https://www.inegalites.fr/Le-mal-emploi-toujours-au-plus-haut, September 2020).
As some abortions are anonymous (mainly those concerning adolescent girls), this decrease may be slightly overestimated as age is unknown in these cases. In 2019, almost 1,000 anonymous abortions were performed.
Non-hospital: treatment provided outside healthcare settings.
The legal scope of midwives’ practice may soon be expanded if the bill adopted in parliament on 8 October 2020, based on a report submitted in September 2020 by the delegation for women’s rights (Battistel and Muschotti, 2020), is ratified by the Senate.
PACS unions have outnumbered marriages (in metropolitan France) since 2016, if marriages between couples already in a registered union (i.e. PACS partners who dissolved their PACS in order to marry) are excluded. They represented more than 40,000 marriages in 2016.
Responsibility for registering PACS unions and dissolutions was transferred to municipal registry offices on 1 November 2017.
The total marriage rate corresponds to the number of marriages that would be observed in a cohort of 1,000 people if at each age they experienced the same marriage frequencies as those observed in a given year.
That is, marriages between never-married individuals.
This information is mentioned on the marriage certificate when the marriage is registered. A variable (common children or not) is thus included in the detailed dataset of marriages.
Notaries should register divorces in the civil records.
Law no. 2016-1547 of 18 November 2016.
Statistics on out-of-court divorces registered by notaries were not communicated to the Ministry of Justice before October 2020.
The new PACS registration and dissolution procedures introduced in November 2017 may have delayed the registration of certain dissolutions until the following year, i.e. 2018.
Direct divorces and legal separations converted into divorces.