Slightly slower growth

1The total population of France on 1 January 2010 (mainland France, Corsica and overseas départements) is estimated at 64.7 million, of whom 62.8 million in metropolitan France (mainland France and Corsica) (Pla and Baumel, 2010). In 2009, the total population rose by 346,000, of whom 325,000 in metropolitan France (Appendix Table A.1). [1] Total growth was therefore slightly weaker than in 2008 (when the population of metropolitan France had risen by 339,000) owing to the conjunction of three somewhat less positive factors: births and estimated net migration registered mild declines of 3,000 and 5,000 respectively, while the number of deaths edged up by 6,000. [2] The total growth rate is therefore estimated at 5.2 per 1,000, down from 5.4 per 1,000 in 2007 and 2008 (Appendix Table A.1).

2France’s population growth remains relatively strong in European terms. In the European Union (EU), according to Eurostat, the population is still decreasing in seven countries – Germany and six new Member States (Bulgaria, Estonia, Latvia, Lithuania, Hungary, and Romania) – plus Malta in 2009, owing to a migration balance that is now negative. In these eight countries, the population decline is due to negative natural growth as well as negative or zero net migration, except in Hungary, where positive net migration failed to offset the excess of deaths over births. Austria, Italy, and Portugal also recorded more deaths than births in 2009, but their populations rose slightly thanks to immigration.
In France, population growth was mainly due to natural increase. With a natural growth rate of 4.3 per 1,000 in 2009 for the whole country, France was outpaced only by Ireland (10.2 per 1,000) and Cyprus (5 per 1,000). It ranked ahead of Luxembourg (4 per 1,000), the United Kingdom (3.7 per 1,000), and the Netherlands (3.1 per 1,000). No other country registered natural growth in excess of 3 per 1,000.

A slowly ageing population structure

3This relatively favourable natural growth can be attributed to three causes. First, French fertility is among the highest in Europe. Second, life expectancy at birth is high – notably for women – and still rising. Third, the population structure is unfavourable to mortality; the “depleted cohorts” born during the First World War are currently generating a relative deficit of deaths at certain ages (approximately 90), in particular among women, whose deaths tend to be concentrated around that age. [3] However, these depleted cohorts are being gradually replaced by larger cohorts, resulting in an increase in deaths while mortality remains unaffected (Niel and Beaumel, 2010).

4The uptrend in births since 2004 has slightly broadened the base of the population pyramid (Figure 1), but the overall population structure continues to age (Appendix Table A.2). The percentage of under-20s is still declining by 0.1 points a year, reaching 24.4% on 1 January 2010 in metropolitan France. The proportion of over-59s has been rising sharply since 2006, as the first baby-boom cohorts enter their sixties. However, this large increase still mainly concerns the 60-64 age group, as the population aged 65 and over is rising by only 0.1 points a year. The 20-59s are therefore the group whose percentage has been declining rapidly since 2006.

Figure 1

Population pyramid of metropolitan France on 1 January 2010

Population pyramid of metropolitan France on 1 January 2010

5This ageing of the population structure by broad age group will gain momentum in the years ahead. In its new population projections, the French National Institute for Statistics and Economic Studies (Institut national de la statistique et des études économiques, INSEE) forecasts a further rapid increase in the population aged 60 and over until around 2035, when all the large cohorts born between 1945 and 1975 will be 60 and older (Blanpain and Chardon, 2010). In INSEE’s baseline scenario, nearly one person in four (24.8%) will be aged 60 or over by 2015, and more than three in ten (30.6%) by 2035. Even under the high fertility or low life expectancy scenarios, the 30% threshold will be crossed in 2035. From the 2020s, ageing will be most pronounced above age 75, when the first baby boom cohorts have reached that age. Among centenarians, likewise, numbers will surge from 2045, reaching between 120,000 and 400,000 by 2060 (versus 15,000 in 2010) aunder the declining mortality hypotheses (Blanpain, 2010). The ageing of the French population structure will thus continue, for it is built into the population pyramid. The only unknown is the speed of the process, which will depend on the future pattern of mortality but also, to a lesser extent, on fertility and migration.

6Today, however, French population ageing is relatively moderate by comparison with Germany and Italy, where more than one in five inhabitants was aged 65 or over on 1 January 2009 (20.4% and 20.1%, respectively, according to Eurostat), and where the under-20s, at 19% of the total population, are outnumbered by the over-65s. Ireland remains the “youngest” country in the European Union, with 11% of its population aged 65 and over, and 27.2% aged under 20.

Small increase in residence permits issued in 2008

7As in previous years, the analysis of immigration flows to France is mainly based on initial residence permits valid for one year or more issued to foreigners still subject to a residence permit requirement for settlement in France. Since 2004, this category essentially consists of citizens of countries outside the European Union. [5] However, for foreigners from the EU, recent flows can now be estimated from the annual census surveys. Respondents who report that they were residing outside France five years earlier are asked to give the year of arrival in France. Based on the answers to this question, it is estimated that an average of 55,000 EU citizens arrive in France every year. [6] The estimate of total flows, based on a lower provisional value of 40,000, has therefore been revised (Appendix Table A.3). The number of foreigners who “establish residence” [7] in France annually is thus believed to have exceeded 210,000 since 2003, except in 2007, when they totalled just under 200,000.
These recent fluctuations in total flows are due solely to the change in the number of residence permits issued to nationals from “third countries” (i.e. outside the European Economic Area [EEA]). After peaking in 2003, the number of new permits declined gradually until 2006, then more rapidly in 2007 (–8.3% on a like-for-like basis, i.e. EU membership in 2007), before rising again in 2008 (+7.9%), with 156,056 first permits issued, versus 144,658 in 2007. The breakdown by reason for admission shows that the increase is confined to specific types of permits (Table 1).

New permits issued to “workers”

8The reason for admission that recorded the largest growth both in absolute numbers (+10,000) and in relative terms (+134%) was admission for employment purposes. The increase is a consequence of the Act of 24 July 2006, which introduced a new set of residence permits designed to promote immigration of workers. These include a “skills and talents” permit, and, especially, “employee on assignment” and “seasonal worker” permits (Rapport au Parlement, 2010). However, the increase is partly artificial, since the statistics now include seasonal workers (5,400 in 2008), for whom the permit is optional. [8]

9Two other reasons for admission, which had been trending down in recent years, also increased in 2008: “student” (+11%) and “humanitarian protection”(+14%), the latter a consequence of the rise in permits issued to refugees and stateless persons following the surge in asylum applications since the fourth quarter of 2007.

Table 1

Residence permits issued to non-EU nationals (EU* membership in 2007) by reasons for admission

Reason for admission Numbers Change 2007/2008 (%) 2005* 2006* 2007 2008 Family member 88,274 90,270 80,098 77,044 –4 Minor child 13,177 9,897 9,799 9,506 –3 Spouse of foreign national 13,378 11,097 11,531 11,938 +4 Spouse of French national 41,635 41,569 36,365 35,225 –3 Parent or child of French national 9,713 9,824 10,197 9,799 –4 “Personal and family life” permit 10,371 17,883 12,206 10,576 –13 Worker 6,843 7,365 7,496 17,561 +134 Student 37,629 36,417 36,916 40,979 +11 Humanitarian protection 17,827 12,807 11,050 12,603 +14 Refugee and stateless person 11,905 7,120 6,078 7,533 +24 “Ill foreigner” permit 5,922 5,687 4,972 5,070 +2 Legalization 2,448 2,350 1,300 1,463 +13 Economically independent 8,201 8,445 7,759 6,326 –18 “Visitor” permit 6,139 6,596 6,425 5,188 –19 “Retired person” permit 2,062 1,849 1,334 1,138 –15 Reason unknown 122 66 39 80 +105 Total 161,344 157,720 144,658 156,056 +8 *Excluding Bulgaria and Romania in 2005 and 2006 (admitted to EU on 01/01/2007). Population: Metropolitan France. Source: Collated by INED (X. Thierry) using information from the central residence permit register (AGDREF) supplied by the Ministry of Immigration, Integration, National Identity, and Co-development.

Residence permits issued to non-EU nationals (EU* membership in 2007) by reasons for admission

10Admissions for other reasons are declining. The decrease in admissions as “visitors” and “retired persons”, fell quite sharply (a total of –18%), as did family immigration, but by just 4% versus 11% in 2007. The share of family immigration in total admissions to residence therefore declined slightly and, for the first time since 2003, accounted for just under one-half of new permits in 2008 (49.4%). In absolute terms, the “personal and family life” and “spouse of French national” reasons for admission registered the largest decreases in 2008. The former often consists of legalization of foreigners with close “personal and family ties” in France. After the exceptional legalization of parents of children enrolled in school in France in 2006, admissions for this reason have been falling. Permits issued to spouses of French nationals recorded a much smaller decline in 2008 than in 2007. The sharp drop in such permits in 2007 reflected the first year of implementation of the Act of 14 November 2006 on the control of validity of marriages. This tightened the formalities to be accomplished before celebrating a marriage between a French national and a foreign spouse in a French consulate and before its official registration in France (Prioux and Mazuy, 2009).

Almost as many Moroccans as Algerians among new permit holders

11Moroccans have been the main beneficiaries of the new three-year permits issued to seasonal workers, accounting for 80% of the total. Almost as many Moroccans as Algerians obtained an initial permit in 2008 (23,382 and 23,605 respectively, for all reasons of admission combined). The two nationalities received 30% of all permits issued to third-country nationals in 2008. [9] China remained in third place with 11,893 admissions (+18%). While China ranks well ahead of Morocco and Algeria for student permits, these two countries remain in the lead for family immigration, and Morocco, thanks to the new “seasonal worker” permits, is the number-one sending country for immigrant workers (Rapport au Parlement, 2010).

12Tunisians received 9,103 residence permits in 2008 (+3%) and Turks 7,607 (+10%), mainly for family immigration, keeping their fourth and fifth places, respectively. They were followed by Malians with 4,535 permits, a 71% increase on the 2,657 issued in 2007. Mali ranks third among non-EU sending countries for worker immigration, behind Morocco and Algeria.

Persons of Algerian origin are now the largest immigrant group

13On 1 January 2006, there were 5,137,000 immigrants in metropolitan France, representing 8.1% of the total population (TEF, 2010, p. 41). As a proportion of these immigrants had acquired French nationality (most often through naturalization or marriage), the number of foreigners living in France on that date was 3,648,000, or 5.8% of the population. While the two groups partly overlap, the nationalities most represented among immigrants and among foreigners are not necessarily the same. Acquisitions of French nationality do not involve all foreign nationalities to the same extent, owing to differences in the history of migration flows and in behaviour (e.g. degree of motivation to obtain French citizenship and frequency of marriages with French nationals).

14Among immigrants, persons of Portuguese origin were equal in number to those of Algerian origin in the 1999 census. This was no longer the case in 2006, when Algeria (691,000 immigrants) and Morocco (634,000), whose number of immigrants had risen, largely outranked Portugal, whose immigrant population remained stable (569,000) (Figure 2). Numbers of immigrants from Italy (330,000) and Spain (269,000) have fallen, but the two countries remain in fourth and fifth place respectively. They precede Turkey (229,000) and Tunisia (227,000), whose immigrant populations have risen, in particular for Turkey. They are followed by the United Kingdom (134,000), whose number of immigrants has surged, almost matching the numbers from Germany (128,000), followed by Belgium (102,000). All other countries have fewer than 100,000 immigrants in France. China, whose immigrants more than doubled in number to 69,000 between 1999 and 2006, ranks only fifteenth, behind Senegal and Vietnam.

Figure 2

Immigrants to metropolitan France by country of birth

Immigrants to metropolitan France by country of birth

15The rankings differ slightly if we look at persons who report themselves as foreign nationals in the population censuses (whether they are immigrants or born in France). In 2006, [10] the Portuguese formed the largest group (490,000), closely followed by Algerians (481,000) and Moroccans (460,000); Turks (224,000) outnumbered Italians (177,000) and Tunisians (146,000) slightly outnumbered Spanish (134,000). These nationalities exhibit very different population structures: most Italians (58%) and Spanish (54%), who have a long history of immigration to France, are aged 55 and over, compared with around one-third among Algerians (36%) and Portuguese (33%), and one-fifth among Moroccans (21%) and Tunisians (20%); the share is only 12% among Turks, and 7% for all immigrants from sub-Saharan Africa. In addition to differences in the historical timing of migration flows and to the diversity of attitudes regarding the acquisition of French nationality, the age structure of foreign population groups is influenced by the frequency of returns to the home country, particularly after retirement.

A slight decline in births

16In 2009, the number of births in France fell slightly by 3,763, or 0.5%, to 824,641 (Beaumel et al., 2010). The metropolitan départements registered 793,420 births, 2,624 fewer than in 2008 (–0.3%), due primarily to the fact that 2009 was a non-leap year which followed a leap year in 2008 [11] (Pla and Beaumel, 2010). According to initial INSEE estimates of monthly numbers of births in 2010, the world economic crisis and the upsurge in unemployment, which began to affect France in the second half of 2008, do not seem to have undermined the birth rate in late 2009 or in early 2010 (Figure 3).

Figure 3

Monthly numbers of births in metropolitan France since 2000

Monthly numbers of births in metropolitan France since 2000

17The total fertility rate (TFR) declined in 2009, but to a lesser degree: after a record high in 2008 of slightly over 2 children per woman in the whole of France (including overseas départements), and 1.99 in metropolitan France, it dipped by 0.01 to 1.99 and 1.98 children per woman, respectively.

18This mild decrease is due to the lower fertility of women aged under 30. Fertility over age 30 continued to increase, but more moderately than in 2008 (Table 2). For the past five years, fertility has been rising annually by an average 22 children per 1,000 women after age 30, and decreasing by 7 children per 1,000 women below that age. In 2009, as in 2007, the rise was below this average after age 30 (+11), and the decrease was above this average before age 30 (–23). It is mainly between ages 20 and 35 that the comparison with the five-year average is less favourable; at ages 35-40, the increase in 2009 was almost as strong as in 2008.

Table 2

Fertility by age group since 2004 (per 1,000 women), metropolitan France

Age reached in the year Sum of age-specific fertility rates Absolute change(a) Average 2004 2005 2006 2007 2008 2009* 2004-2005 2005-2006 2006-2007 2007-2008 2008-2009 2004-2009 Under 20 39 38 37 36 37 34 –1 –1 –1 0 –2 –1 20-24 276 274 279 270 273 262 –3 +5 –8 +3 –11 –3 25-29 645 641 655 641 641 631 –4 +14 –14 +1 –10 –3 30-34 604 619 642 638 652 654 +15 +23 –4 +13 +3 +10 35-39 270 281 298 300 310 316 +11 +17 +3 +9 +7 +9 40+ 64 67 70 73 76 77 +3 +3 +3 +3 +1 +3 Total 1,898 1,920 1,980 1,959 1,988 1,975 +21 +61 –22 +30 –13 +15 *Provisional. (a) Because of roundings in the sums of rates shown in the left-hand side of the table, the changes calculated here may not correspond to the apparent differences. Source: INSEE.

Fertility by age group since 2004 (per 1,000 women), metropolitan France

19The breakdown of fertility by broad age group is therefore still shifting. Since 2005, women aged 35-39 have been making a greater contribution to fertility than women aged 20-24. Since 2008, the largest contribution has come from women aged 30-34, exceeding that of women aged 25-29, whose fertility has been trending down in the past few years. While, in the aggregate, two-thirds of total fertility is due to women aged 25-34 (Prioux and Mazuy, 2009), the concentration of births between ages 25 and 35 depends heavily on social origin, and particularly on female educational level (Davie and Mazuy, 2010). Among the most highly educated women (with a degree in higher education), three-quarters of total fertility is due to women aged 25-34, and one-fifth to women aged 35 and over, while births before age 25 are rare. This pattern is not reproduced among women with no educational qualifications: only a little over one-half of their fertility is concentrated between ages 25 and 35. In sum, the mean age at childbearing increases with female educational level, particularly because women with a degree in higher education have their first child towards age 30 on average, almost 5 years later than women with no qualifications (Davie and Mazuy, 2010).

20The increasing proportion of women with a degree in higher education is one of the factors that explains the change in distribution of age-specific fertility and the increase in the mean age at childbearing, which reached 30 for the first time in 2009 (Pison, 2010; Appendix Table A.4). The mean age of 30 could be reached as early as the 1973 cohort, and even exceeded in the 1975 cohort if fertility rates over age 34 (age of women born in 1975 at the end of 2009) continue to rise at the same pace (Appendix Table A.5). Under this scenario, completed fertility will rise to 2.04 in the 1975 cohort after falling to 2 children per woman in the 1969-1972 birth cohorts. If fertility after 35 were to stop rising (i.e. if rates levelled off at their 2009 values), completed fertility in the 1971-1974 cohorts would be slightly under 2 children per woman – an all-time low probably never previously reached in France, even in the cohorts born in the late nineteenth century (Daguet, 2002).

An upward trend in European fertility

21Average fertility in Europe is not as low as it was about a decade ago. The total fertility rate (TFR) fell steeply in the 1990s, particularly in the countries of former Eastern Europe, but also in Germany, Italy, and Spain. Today, by contrast, the TFR is on the rise in almost every country, even those where the rate was not especially low (Figure 4 and Appendix Table A.6). The exceptions are rare. Only in Portugal, and possibly Luxembourg, is fertility still falling. In Cyprus and Malta, the TFR appears to have stopped its decline. In Germany, Austria, and Hungary, the sometimes very modest upturn in fertility has levelled off at close to 1.3-1.4 children per woman. Alongside these three countries, it is in Southern Europe that fertility is now lowest (Figure 4C), with levels ranging between 1.32 (Portugal) and 1.45 (Greece) in 2009. Among the new EU members in eastern Europe (Figure 4D), levels are also fairly similar and still rather low (ranging between 1.33 in Hungary and 1.57 in Bulgaria). The one exception is Estonia, which has taken a clear lead since the 2000s, but it may soon be caught up by other countries where the rate has been rising quite vigorously (Bulgaria and Lithuania). In northern Europe, fertility rates are also relatively homogeneous, and distinctly higher (Figure 4B), ranging from 1.84 in Denmark to 1.98 in Norway and possibly more than 2.1 children per woman in Ireland and Iceland. [12] The fertility of the seven countries of continental western Europe is more diverse (Figure 4A), with values ranging from 1.35 in Germany to 1.98 in metropolitan France.

Figure 4

Total fertility rate in Europe since 1990

Total fertility rate in Europe since 1990

22Irrespective of the trends in TFR, the mean age at childbearing has been rising across Europe owing to the increase in fertility among women aged 30 and over. France is not the only country where the mean age has reached 30 (Pison, 2010). In 2008, twelve EU countries crossed this threshold, of which seven are approaching age 31, and in four countries – the Netherlands, Ireland, Italy, and Luxembourg – the mean age is above 31. It is generally lower among new EU entrants in central and eastern Europe, particularly Bulgaria (age 26.8 in 2008) and Romania (age 27.1), which combine two distinctive features: the highest fertility in Europe before age 20 and the lowest after age 30.
This increase in the age at childbearing explains why completed fertility generally exceeds the TFR, notably in the former Eastern European countries (Appendix Tables A.6 and A.7). These are not the countries with the lowest completed fertility, but rather Italy and Spain, where it stands at around 1.4 children per woman for the 1974-1975 birth cohorts. In the EU, Ireland still posts the highest fertility, followed by France. Outside the EU, Iceland ranks above France, while in Norway, where fertility overtook that of France among the 1960s cohorts, the completed fertility of the 1970s cohorts has now dipped below 2 children per woman.

No decrease in abortions

23French law requires doctors who perform induced abortions to file notification forms. Despite this obligation, the statistics compiled from the forms are not yet exhaustive, although there has been considerable improvement in recent years. Since 2002, the total number of induced abortions performed in metropolitan France has been tracked annually using statistics from the Direction de la recherche, de l’évaluation et des statistiques (Directorate for Research, Assessment, and Statistics, DREES, Ministry of Labour and Social Affairs) (Rossier and Pirus, 2007). Until now, DREES compiled the series from the annual statistics on healthcare facilities (Statistique Annuelle des Établissements de Santé, SAE). But as this source under-reports the number of medical terminations performed in doctors’ surgeries (authorized since 2004), it is now supplemented for this category of procedures by data from the national health insurance fund (Caisse nationale d’assurance maladie, CNAM) (Vilain, 2009). The total number of induced abortions performed in 2006 has been revised upwards accordingly, from 206,999 to 215,390. This gives a much higher annual increase between 2005 and 2006 than initially estimated (+4.4% versus +1.6%) (Appendix Table A.8). The 2007 figure of 213,382 induced abortions is down slightly (–0.9%), but remains high by comparison with earlier years and with INED annual estimates of 206,000 for the early 2000s (Rossier and Pirus, 2007).

24The indicators measuring the frequency of induced abortions have therefore been rising in recent years (Appendix Table A.8, last two columns), returning to estimated levels that had not been seen for 15-20 years. The total induced abortion rate reached 14.9 per 1,000 women aged 15-49 in 2006 and 14.7 in 2007, close to the 14.8 per 1,000 estimated for 1990. The total abortion rate, or mean number of induced abortions per woman, came to 0.53 per woman, the same value as in 1986. This rate is a better indicator of the change in frequency, as it does not depend on the age structure of the population at risk. By comparison with 1991 and 1992, when the estimated rate bottomed out at 0.48, the frequency of terminations appears to have risen by approximately 10%. But as fertility also increased by around 12% over the same period, the ratio of induced abortions to live births is lower today than in the mid-1990s (Appendix Table A.8, column 5).

Slightly more repeat abortions

25The breakdown of induced abortions by order, made possible thanks to the notification forms, allows a more detailed analysis of the change in frequency of terminations. The slight increase is due to a rise in the frequency of repeat abortions; the probability of a first termination is stable and has even fallen slightly since the 2000s (Figure 5). It would appear that the rather sharp decline in the frequency of first induced abortions from 0.38 to 0.35 per woman between 2002 and 2005-2007 is due largely to the change in the procedure for recording information on the order of terminations following the introduction of a new, simplified notification form in 2005. [13] By contrast, the increase in the component of second and higher order abortions in the total rate per 1,000 is perceptible as early as the 1990s, although the change in the notification form heightened the trend between 2002 and 2005. If the 2007 figures remained stable over a long period, it would mean that 35% of women would undergo at least one induced abortion in their lives, with 22% undergoing only one and 13% at least two. These numbers differ slightly from the estimate for 2002 based on the old forms, which suggested that 38% of women had at least one induced abortion, with 28% undergoing only one, and 10% at least two (Rossier et al., 2009, p. 517). The fact that a smaller proportion of women are undergoing induced abortion than hitherto estimated – 35% instead of 38% – indicates that repeat abortions are becoming more common, and their share among total induced abortions is gradually rising. These new figures do not challenge the findings of last year’s analysis of repeat abortions, namely, the disappearance of the learning effect after a first abortion (Rossier et al., 2009).

Figure 5

Total abortion rate and breakdown by abortion order since 1976

Total abortion rate and breakdown by abortion order since 1976

Population: Metropolitan France.

26The comparison of age-specific induced abortion rates in 1997 and 2007 shows a rise in abortion frequency under age 29, stability over that age, and even a mild decline around age 35 (Figure 6). The rise is most visible between ages 20 and 27, but is steepest at ages 16-17 in absolute value, and at ages 14-17 in relative terms: at these ages, of course, the statistic concerns a first induced abortion, whose rates increase in the same proportions. But starting at ages 18-20, the main drivers of the rise in induced abortion rates are second order abortions, followed by third or higher order abortions. The decrease in age at first induced abortion is certainly one of the factors behind the increase in repeat abortions.

Figure 6

Abortion rate by age and abortion order, 1997 and 2007

Abortion rate by age and abortion order, 1997 and 2007

Population: Metropolitan France.

A further rise in PACS civil partnerships

27In 2009, the number of new civil partnerships (pacte civil de solidarité, PACS) rose by nearly 20% to 174,504, of which 173,045 in metropolitan France (Table 3). This represents a sharp slowdown from the previous year’s 43% rise. Although the latest figure is the lowest observed since 2002, the ever greater success of the PACS attests to its “relative democratization […], understood as the social dissemination of the contract among social groups that were least attracted to it at the time of its introduction” (Rault and Letrait, 2010). Adopted promptly by same-sex couples in the very first months of its introduction in November 1999 (Carrasco, 2007), the PACS later became increasingly popular among different-sex couples, with each legislative change apparently triggering a new rise in demand: in 2005, the tax regime of new PACS partners was aligned with that of newly married couples; in 2007, separation of property replaced common property as the default matrimonial regime for managing partners’ assets, and gift tax and inheritance tax between partners were reduced; in 2008, inheritance tax was abolished and the provisions relating to gift tax were aligned with those applying to married couples.

The number of registered unions has also increased

28The gradual decline in the number of marriages in recent years has been largely offset by the large increase in new heterosexual PACS unions. This suggests that the PACS has encouraged more couples to officialize their unions in a legal framework. However, the annual number of couples legalizing their unions cannot be determined simply by summing heterosexual PACS and marriages, for some marriages are preceded by a PACS. It is possible to offer an estimate of registered unions for 2007-2009, as PACS unions have been recorded in the margin of birth certificates since 2007. As a result, PACS dissolutions are registered more comprehensively, in particular if they are dissolved by marriage between the partners. [14] This analysis shows marriage to be one of the main causes of PACS dissolutions (Table 3), accounting for almost half of dissolutions in 2007 (47%), and progressively fewer in 2008 (41%) and 2009 (34%). These dissolutions must therefore be subtracted from marriages to estimate the total number of new registered unions. Adding these “net marriages” to new PACS unions, we obtain a total of 357,596 unions in 2007, 393,614 in 2008, and 413,340 in 2009, an increase of 15% in two years. The latter figure is very close to the 416,521 marriages celebrated in 1972, the year with the highest number of marriages since 1945. In fact, it was the year that marked the end of the “golden age” of marriage, for it was followed by a movement away from marriage and the rise of non-marital cohabitation (Toulemon, 1997). The number of unions registered in 2009 can be compared directly with the number of marriages in 1972 since the cohorts old enough to form unions are of comparable size today. It is still much smaller than the number of new unions formed each year (estimated at 550,000 in the 1990s by Beaumel et al., 1999) and well below the number of unmarried couples, even after subtracting an estimated number of couples who have already signed a PACS. Thus, despite its growing success, the PACS is far from having exhausted the entire stock of unmarried couples.

Table 3

PACS and PACS dissolutions since 2007, metropolitan France and overseas départements

2007 2008 2009 PACS registered (total) 102,023 146,030 174,504 PACS concluded by sex of partners: male-male 3,708 4,780 4,894 female-female 2,509 3,423 3,549 male-female 95,708 137,820 166,056 not known 98 7 5 Mean age of partners (years) 32.0 33.4 33.3 Dissolutions (total) 22,783 23,448 26,573 Reason for dissolution: Mutual consent 10,850 12,763 16,232 Unilateral request by one partner 746 709 912 Marriage 10,781 9,610 9,120 Death 371 341 281 Other reasons and not known 35 25 27 Mean duration of dissolved PACS (months) 28.0 29.2 28.0 Source: Ministry of Justice, SDSED.

PACS and PACS dissolutions since 2007, metropolitan France and overseas départements

A further decline in the proportion of same-sex couples

29The number of PACS unions signed between two women or two men has also risen, increasing by 36% between 2007 and 2009, from an overall total of 6,217 to 8,243. However, the growing success of the PACS among heterosexual couples continues to reduce the proportion of same-sex civil unions. The percentage has dropped from 42% of all PACS couples in 1999 (Carrasco, 2007) to just 6.1% of new unions in 2007 and 4.8% in 2009. While male same-sex unions still outnumber female ones, the imbalance is growing smaller: in 2009, 58% of homosexual PACSs involved two men, down from nearly 80% of new same-sex unions formed in 1999.
Since 2007, the statistics published by the Ministry of Justice (Table 3) have included the sex of the partners and the mean duration of dissolved PACS unions. Despite this slightly greater level of detail, it is not yet possible to deepen the demographic analysis of these partnerships. The upcoming release of a database containing selected demographic variables (age of contracting parties, year of PACS dissolution) should allow comparisons with marriages and divorces. But general population surveys will still be needed for in-depth sociological analysis of the reasons for choosing the PACS (Rault and Letrait, 2010).

Fewer marriages among singles, widow(er)s and divorcees

30After rising in the 1990s and levelling off in 2000, the number of marriages began a downward trend (except in 2005) that accelerated in 2008 and 2009, with just 251,400 marriages registered in 2009. Slightly more than 245,000 marriages were celebrated in metropolitan France (Appendix Table A.9), 14,000 fewer than in 2008, and 6,300 were registered in the overseas départements, a decrease of 300. This represents a 5% [15] drop in both cases. This downtrend is observed in all regions of metropolitan France, and for all types of marital status: the number of marriages fell to a similar extent for singles, widow(er)s and divorcees. The majority of marriages were between singles (Beaumel and Pla, 2010b). They represented 80% of newlyweds in 2009 (79.2% of grooms and 80.6% of brides). Divorcees represent slightly below 20% of newlyweds (19.4% of grooms and 18.1% of brides), and widow(ers) less than 1.5% (1.4% for men and 1.3% for women).

31The total first marriage rate (the sum of age-specific marriage rates for single persons) fell between 2008 and 2009, dropping to below 50%. Based on the overall probability, it was just above 50% (Appendix Table A.9). This means that if the conditions of first marriage observed in 2009 remain unchanged, barely more than one man and one woman in two (56% and 53%, respectively) will marry in future cohorts.

32This drop in the first marriage rate, which has reached an all-time record low, is observed at all ages (Figure 7). Even for the ages where marriage rates were relatively stable in recent years, i.e. after age 30, a decline was observed in 2009. The probability of marrying before age 25 is decreasing year on year, especially for men, due to a male model of late marriage, which now also applies to highly educated women (Davie and Mazuy, 2010; Galland, 1999). Few men form a union before completing their education. Marriage timing is more varied for women, who enter conjugal life at an earlier age.

Figure 7

Breakdown of the first marriage rate, based on age-specific probabilities

Breakdown of the first marriage rate, based on age-specific probabilities

33Mean age at first marriage has increased considerably in recent decades. The mean age at marriage in 2009 was 29.7 years [16] for women and 31.6 years for men (a difference of 1.9 years). Thirty years earlier (when the couples marrying in 2009 were born), men married at age 25 and women at age 23. Women still marry two years earlier than men (Bozon, 1990), and age at second marriage is also younger: widows and female divorcees remarried at age 39.0 on average in 2009, and widowers and male divorcees at age 40.7, a difference of 1.7 years.

34The decrease in all age-specific marriage probabilities observed in 2009 implies an increase in the estimated proportion of never-married persons in each cohort. Over one-third of men born in 1973 will be single at age 49 (37% [17]), compared with just 12% of men of the same age 30 years previously. For women, the proportions are very similar: 35% of women born in 1975 will be never-married at age 49 if the conditions of first marriage observed in 2009 remain unchanged, versus 8% thirty years previously (Appendix Table A.10).

35Adults can now live together outside the institution of marriage. The status of husband, and even more so that of wife, are much less universal for the younger generations. The increase in the number of registered unions (see above) is thus explained exclusively by the success of the PACS civil partnership. While couples still want a legal framework for their union, traditional legislation is losing ground in favour of new, alternative forms of union. Conjugal situations are becoming more diverse: living without a partner after a separation (with or without children, see above), marriage, civil partnership, or non-cohabiting unions (living apart together). Nonetheless, men and women still believe in the ideal of conjugal life, and their requirements in this respect are increasing, “[the couple] is still a strong value in France and the apparent deregulation of the family is due primarily to the abundance of co-existing norms rather than to their disintegration” (Déchaux, 2009).

Irretrievable marriage breakdown accounts for one in ten divorces

36The decline in divorces continued in 2009, in both metropolitan France and the overseas départements. A total of 129,504 divorces were pronounced in 2009, versus 132,594 in 2008. This is a drop of 2%, [18] slightly larger than that of 2008 (Prioux and Mazuy, 2009). The number of divorces nonetheless remains above the level recorded in the early 2000s.

37In metropolitan France, 127,578 divorces were pronounced in 2009, versus 132,594 in 2008, down by 1.4%. The total divorce rate also continued the steady decline observed since the peak of 52.3 divorces per 100 marriages in 2005 following the simplification of divorce proceedings (Prioux, 2008). It stood at 44.7 divorces per 100 marriages in 2009 (Appendix Table A.9). If the conditions observed in 2009 remain unchanged, fewer than one marriage in two will end in divorce. The steep rise in the divorce rate is now over (the law of 2005 is having less effect) and its level is tending to stabilize.

38A minority of persons who divorced in 2009 – 10% of divorced women and 5% of divorced men – were aged below 30. The vast majority are in their thirties, forties or fifties. Only 5% of female divorcees and 10% of male divorcees are over 60, while the over-60s represented 31% of married women and 37% of married men on 1 January 2009.

39The risk of divorce is highest after nine years of marriage. Given that many couples are already separated when they file for divorce, the actual duration of marital life before separation is shorter.

40A majority of divorces (53.5%) are by mutual consent. However, petitions for divorce following irretrievable marriage breakdown (which rose from 1.3% to 9.9% between 2005 and 2009) and individual petitions accepted by the spouse (9.6% in 2005 and 24.5% in 2009) are the two types of divorce with the fastest relative growth. Fault-based divorces are becoming less frequent (11.4% of procedures in 2009) and will probably soon be overtaken by irretrievable marriage breakdown. In 2000, the latter represented just 1.4% of proceedings, versus 41% for fault-based divorces, but today the gap between the two is narrowing (Table 4).

Table 4

Number and distribution of direct divorce proceedings in metropolitan France, 2000-2009

2000 2005 2006 2007 2008 2009 Type of proceeding* N % N % N % N % N % N % Joint petition / mutual consent 48,458 43.3 90,843 60.4 75,868 56.5 71,831 55.2 69,358 54.1 67,662 53.5 Uncontested petitions / uncontested divorce 15,983 14.3 14,403 9.6 21,798 16.2 27,438 21.1 29,792 23.3 30,956 24.5 Breakdown of conjugal life / irretrievable marriage breakdown 1,596 1.4 1,942 1.3 7,022 5.2 9,879 7.6 11,573 9.0 12,497 9.9 Mental impairment 63 0.1 27 0.0 27 0.0 71 0.1 61 0.0 54 0.0 Fault 45,780 40.9 42,208 28.1 28,741 21.4 19,900 15.3 16,453 12.8 14,393 11.4 Not specified 993 0.7 936 0.7 941 0.7 885 0.7 960 0.8 Total 111,880 100 150,416 100 134,392 100 130,060 100 128,122 100 126,522 100 * Name of proceeding before 2005 / after 2005. Source: Ministry of Justice.

Number and distribution of direct divorce proceedings in metropolitan France, 2000-2009

Increase in the share of divorces not involving a minor child

41The majority of couples who divorced in 2009 (57%) had one or more minor children, and around 130,000 children were concerned, [19] although the number of such divorces has fallen back to its level of the early 2000s. The increase in divorces over the last decade has thus concerned more couples without minor children, who more frequently divorce by mutual consent (Chaussebourg et al., 2009). This increase is probably linked to the growing number of divorces among childless couples after a short marriage duration. The risk of divorce after many years of marriage has also increased, however, probably reflecting an increase in divorce among couples whose children have reached adulthood (the risks are nonetheless low).

Figure 8

Direct divorces with and without minor children, and estimated number of children concerned

Direct divorces with and without minor children, and estimated number of children concerned

Adults less often live with a partner

43As shown in Figure 9, the proportion of under-25s living in cohabiting unions (marital or otherwise) is no longer falling (Daguet and Niel, 2010). The proportion had declined between the two earlier censuses (Prioux, 2002), mainly because of the increase in age at first union (Prioux, 2003). This age has therefore probably stopped rising, as confirmed by the halt in the fertility decline among the under-25s over the past ten years or so (Breton, 2010). By contrast, between ages 25-65, the proportion of men and women in a union is still falling, owing to the increase in separations. The greater frequency of repartnering [20] does not fully make up for this increase, hence the lower frequency of cohabiting unions in that age bracket. Women over 65 were more often living with a partner in 2006 than women of the same age in 1999. These cohorts (born before 1940) were less affected by the rise in divorce, whose consequences are still largely offset by the decline in mortality, which delays their widowhood. At these ages, men far more frequently have a partner than women: they are fewer in number (owing to their excess mortality), and they are more frequently older than their spouse than the reverse. Among the over-80s, the proportion of men living in a union rose between 1999 and 2006: the decline in female mortality and the improvement in male health are delaying their widowhood and/or their institutionalization.

Figure 9

Proportions of men and women at each age living with a partner in their own home in 1999 and 2006

Proportions of men and women at each age living with a partner in their own home in 1999 and 2006

Population: Metropolitan France (total population).

Education favours living with a partner for men

44The frequency of living with a partner at each age differs with male and female educational attainment (Daguet and Niel, 2010). The most highly educated individuals, who spend more years in education, are less frequently in cohabiting unions before age 25 than the lowest educated (Robert-Bobée and Mazuy, 2005). Over age 25, educational attainment may be linked to a higher or a lower frequency of living with a partner. The “educational endowment”, it has been argued, has a “pernicious effect” for women, as “women with the strongest social and cultural capital are more often never-married than other women” (de Singly, 1987, p. 167). [21] By contrast, the proportion of singles tends to be higher among less educated men (de Singly, 1982) – especially those with no qualifications – whereas the most highly educated men are advantaged on the “marriage market”. In addition to these differences in the frequency of “never-married” status, behavioural differences regarding union dissolutions and repartnering explain the frequencies of living with a partner at each age observed at any given time.

Figure 10

Proportions of men and women at each age living alone in their own home in 1999 and 2006

Proportions of men and women at each age living alone in their own home in 1999 and 2006

Population: Metropolitan France (total population).

45The partnership status of men by educational level in 2006 is fairly consistent with this pattern. Between ages 30 and 75, unqualified men far frequently live with a partner than all other categories of men: the difference ranges from 7 to 11 percentage points. Between ages 30 and 50, it is the most highly educated men who are most often living with a partner or wife (Daguet and Niel, 2010, Figure 5). For example, among men aged 40-44, nearly 78% of higher education graduates live in partnerships versus 74% of holders of the baccalauréat (upper secondary exit examination) or of a lower secondary qualification, [22] and only 66% of men with no qualifications. However, beyond age 55, the most highly educated men no longer differ from men who have completed lower or upper secondary education. Only men with no qualifications less frequently live with a partner than the others.

46For women, the pattern is somewhat different, and does not strictly match the situation described above. Higher education has become far more widespread in younger cohorts and thus no longer reduces the likelihood of being in a union, at least in the 30-45 age group, where women with no qualifications are the category who least often live with a partner. Beyond age 45, the status of the highest educated women has converged towards that of the less educated and the unqualified, among whom the proportion with a partner has declined sharply since 1999 (Daguet and Niel, 2010). At all ages over 45, the women who most frequently have a partner or husband are those who have completed lower secondary education, by far the largest group in those cohorts.
These differences by educational level in the frequency of living with a partner are accompanied by small disparities regarding the legal recognition of unions, which is linked differently to education for men and women. On average, slightly over 81% of women aged 30-59 living in unions report being married, but the proportion of married women decreases with the rise in educational attainment – from 83% for women with no qualifications or with a lower secondary qualification to 77% for women who have completed more than two years of higher education. For men in the same age group, of whom 78% report being married, the marriage frequency is, on the contrary, highest for the best-educated, at just over 80%, and lowest for men who have completed two years of higher education (of whom only 76% report being married). Higher educational attainment is probably associated with greater personal wealth, which acts as an incentive to legal recognition of unions.

Men and women more often live alone

47The decline in the frequency of living with a partner at adult ages is accompanied by an increase in the proportion of men and women living alone (Figure 10), a situation that is not confined to the 25-65 age group, however. The proportion of 20-25s living alone is still rising, because of the higher frequency of enrolment in higher education. Around age 22, roughly one-fifth of baccalauréat holders live alone, as do one-quarter of higher education graduates, versus only one-fifth of persons with a lower secondary qualification in the same age group, and one-twelfth of persons with no qualifications (Figure 11).

Figure 11

Proportions of women (A) and men (B) living alone, by age and educational level, 2006

Proportions of women (A) and men (B) living alone, by age and educational level, 2006

Population: Metropolitan France (total population)

48Beyond age 25-30, separations are the main reason for the rising proportion of persons living alone (Figure 10). Below age 55, the proportion rises more slowly for women than for men, because when a couple with children breaks up, it is more often the man who ends up living alone, while the woman forms a lone-parent family. Over age 65, the proportion living alone decreases slightly because more people live with a partner at older ages (see above). From approximately age 54, women live alone more frequently than men, a pattern that intensifies with age, at least up to ages 85-87, when the proportion peaks at around 56%. This illustrates the classic effect of greater female longevity and age difference between spouses.

49For women, the frequency of living alone is closely correlated with educational level (Figure 11A). Under age 75 or so, the proportion of women living alone rises with education. Although a high educational level makes it easier to achieve residential autonomy, the disparities before age 50 are largely due to differences in the timing of family formation. Women with no qualifications form unions sooner, have children earlier (Robert-Bobée and Mazuy, 2005), and live far more frequently in lone-parent families (Figure 12B), while the highest educated women have children later (Davie and Mazuy, 2010) and more often remain childless (Robert-Bobée, 2006). Over 75, the pattern is reversed: the highest educated women (relatively scarce in these cohorts) less often live alone and more often live in institutions than unqualified women and women with lower secondary qualifications.
For men as well, the frequency of living alone increases with education among young adults (Figure 11B), but the differences are smaller. Beyond age 50, unqualified men are the category who most often live alone. The difference with respect to more educated men is small (whatever their level), since for unqualified men, the lower frequency of living with a partner, mentioned above, mainly reflects other modes of cohabitation. At young ages, they live more often with their parents; at adult ages, they live more often in non-family situations (households composed of several unrelated persons, or “non-household” census categories [23]).

Women without qualifications more often live in lone-parent families

50The increase in union dissolutions is also accompanied by a rise in the proportion of adults – particularly women – living in lone-parent families (Figure 12A). For women, the steepest rise occurs between ages 40 and 55, and between 45 and 50 especially (more than two percentage points). For men as well, the increase is sharpest around age 50, but does not exceed half a percentage point.
For women, being the head of a lone-parent family is also closely correlated with educational level (Figure 12B): under age 45, the proportion of lone parents falls as educational level increases. In addition to early family formation among the least educated women, the data suggest a greater frequency of union dissolution among couples with children, which would explain the sizeable decline in the frequency of living with a partner at these ages between 1999 and 2006 among women with lower secondary qualifications or less (Daguet and Niel, 2010). From age 60, unqualified women much more frequently live in lone-parent families [24] than other women. This may reflect two factors: the integration problems of adult children who continue (or return) to live with their mother, and the rising number of older women who are cared for by their children.

Figure 12

Proportions of men and women living in lone-parent families in 1999 and 2006 by age (A), and proportions of women in lone-parent families in 2006 by age and educational level (B)

Proportions of men and women living in lone-parent families in 1999 and 2006 by age (A), and proportions of women in lone-parent families in 2006 by age and educational level (B)

Population: Metropolitan France (total population).

51Education is a less differentiating factor for men living in lone-parent families, but they share two points in common with women: the highest educated men slightly less frequently live in lone-parent families when they are young because they become parents at a later age; and unqualified men aged 60 and over live more often in lone-parent families than other men in the same age group.

52The diversity of situations among adult men and women at each age by educational level is therefore largely due to education-specific differences in the timing of family formation (Robert-Bobée and Mazuy, 2005).

53The transition from a model of early and universal marriage to one where marriage is later and less frequent, and where other lasting forms of union are possible, results in a multiplication of norms. This plurality gives rise to socially differentiated behaviours (Déchaux, 2009) that affect each sex in a specific manner. For women, early union formation is often associated with a larger family size, long-term withdrawal from the labour force and unequal division of domestic tasks (Régnier-Loilier, 2009). By contrast, women who form a union later have fewer breaks in their working career. For men, later union formation is accompanied by more frequent exclusion from the marriage market of men in the most disadvantaged social categories (Toulemon and Lapierre-Adamcyk, 2000).

A renewed increase in female life expectancy

54The 546,000 deaths in 2009 correspond to a crude death rate of 8.5 per 1,000 inhabitants and a life expectancy at birth of 77.8 years for men and 84.5 years for women (provisional estimates). These levels of life expectancy represent a two-month gain for both sexes with respect to the previous year and indicate that the stagnation observed in 2007 and 2008 for women was not a lasting break in the downward female mortality trend. These figures also show that the gender gap in mean length of life (6.7 years) remained unchanged with respect to 2008.

55If these provisional estimates are confirmed, male life expectancy at birth will have increased by almost three years over a decade (2.95 years between 1996-1998 and 2006-2008), representing an acceleration of progress with respect to the two previous decades (gain of 2.4 years from 1976-1978 to 1986-1988, and 2.5 years from 1986-1988 to 1996-1998). For females, the gains were smaller, totalling 2.5 years, 2.1 years, and again 2.1 years over the last three decades. This slower increase in female life expectancy is a recent trend, first observed in the 1990s. Throughout the second half of the twentieth century, female mortality decreased much more quickly than that of men, and it was not until the 1980s that the speed of mortality decrease among males caught up with that of females, finally overtaking it in recent years.

France is well placed with respect to its European neighbours

56With the exception of eastern Europe, all countries of Europe have reached a life expectancy of 80 years for women, even 84 years in the most advanced countries, led by Switzerland followed by France, Spain and Italy. Compared with Bulgaria, Latvia, Lithuania and Romania, where female life expectancy is below 78 years (Appendix Table A.12), the difference is almost 8 years.

57Dispersion is even wider for male life expectancy, with a difference of almost 13 years between Latvia, where it stands at 67.0 years, and Switzerland, with 79.8 years and Sweden with 79.2 years. France, which ranked twelfth in 2007-2008, is in the middle, alongside Austria, Germany and the United Kingdom.

58France has an unusually large gender gap in life expectancy, only equalled or exceeded by that of the eastern European countries. It is above 10 years in Estonia, Latvia and Lithuania (where it reaches a record 11.3 years) compared with less than 5 years in most European countries.
The countries of eastern Europe also have the highest infant mortality, with a rate of above 10 per 1,000 in Romania. In the other European countries, the probability of dying before age one is below 4 per 1,000 practically everywhere. Scandinavia has a particular lead in this area, with an infant mortality rate of below 3 per 1,000 in Sweden and in Finland (Appendix Table A.13).

Progress especially strong before age 45

59As shown in Table 5, the strongest decrease in the probability of dying in France over the last decade is observed before age 45. To limit random variations, the calculations are based on multi-year tables published by INSEE, the most recent of which covers the years 2006-2008. Between 1996-1998 and 2006-2008, the probability of dying between ages 15 and 25 fell by around one-third for both men and women alike. A decrease of almost equal proportions was observed for men between ages 25 and 45 (–31%). The decrease was –27% for boys below 15 and –25% for girls below 15 and for women aged 25-45. Although less marked, a decrease of 20% was also observed for both sexes at ages 65-80. The smallest decrease is observed at ages 45-65, particularly among women, for whom it was below 10% (15% for men).

Table 5

Change in probabilities of dying at certain ages in metropolitan France between 1996-1998 and 2006-2008*

Probability of dying between Males Females 1996-1998 2006-2008 Change (%) 1996-1998 2006-2008 Change (%) Probability (‰) Probability (‰) Ages 0-15 8.3 6.0 –27.5 6.4 4.8 –25.3 Ages 15-25 9.4 6.3 –32.6 3.4 2.2 –34.9 Ages 25-45 40.0 27.6 –31.1 17.0 12.8 –24.9 Ages 45-65 169.7 143.4 –15.5 72.1 65.5 –9.3 Ages 65-80 434.7 350.5 –19.4 233.1 185.6 –20.4 * Provisional data for the 2006-2008 life table. Source: Calculations based on INSEE life tables (Division of Demographic Surveys and Studies).

Change in probabilities of dying at certain ages in metropolitan France between 1996-1998 and 2006-2008*

60The decline in infant mortality is just slightly below that recorded for the under-15s in general, with a 23% decrease in the probability of dying in the first year of life between 1996-1998 and 2006-2008 (Appendix Table A.11). Over this period, the infant mortality rate fell from 4.7 to 3.6 per 1,000. The improvement was slightly slower for mortality in the first weeks of life, with a 20% drop in the neonatal mortality rate over the same period. It fell from 3.0 per 1,000 in 1996-1998 to 2.4 per 1,000 in 2006-2008 (Appendix Table A.11). In any case, infant mortality has now reached such a low level that its contribution to overall mortality (or life expectancy at birth) has become negligible. The levels observed in other European countries (Appendix Table A.13) nonetheless show that further progress in reducing mortality at these early ages is still possible.

The growing role of oldest-old mortality

61Table 6 gives an overview, by ten-year period, of the contribution of mortality at different ages to progress in life expectancy at birth over the last 30 years and for both sexes. The increasing concentration of mortality at advanced ages is reflected in their growing contribution to the increase in life expectancy at birth. Over the last ten years, 74% of the years of life gained by men and 85% of those gained by women are the result of progress achieved after age 65. And this progress has accelerated over time; between 1976-1978 and 1986-1988, the contribution of this age group was 65% for men and 77% for women, and between 1986-1988 and 1996-1998 it was 68% for men and 81% for women. An examination of changes in the causes of death provide an explanation for this trend.

Table 6

Contribution of age groups to life expectancy gains (years)

Age group Period 1976-1978 to 1986-1988 1986-1988 to 1996-1998 1996-1998 to 2006-2008 1976-1978 to 1986-1988 1986-1988 to 1996-1998 1996-1998 to 2006-2008 Males Females Age 0-14 0.08 0.07 0.02 0.06 0.05 0.02 Age 15-24 0.08 0.06 0.04 0.05 0.04 0.02 Age 25-44 0.18 0.20 0.21 0.14 0.11 0.09 Age 45-64 0.50 0.48 0.50 0.33 0.21 0.18 Age 65-79 0.91 0.88 1.02 0.75 0.50 0.44 Age 80+ 0.65 0.82 1.16 1.15 1.22 1.33 Total 2.40 2.51 2.95 2.47 2.13 2.07 Source: INSEE life tables (Division of Demographic Surveys and Studies).

Contribution of age groups to life expectancy gains (years)

Cause-specific mortality

62Since the 1950s, the two main causes of death in France have been cancers and cardiovascular diseases (Appendix Table A.14). Since the early 1980s, they account for almost 60% of the standardized mortality rate from all causes, for men and women alike. The respective shares of these two major groups of diseases have reversed over time, however.

Substantial progress in combating cardiovascular diseases

63While cardiovascular diseases represented more than one-third of overall mortality (34% for men and 37% for women) in 1980, the proportion has gradually shrunk to around one quarter (24% and 25%, respectively) in 2007. This represents a decrease of around 60% in the standardized mortality rate from this cause for both sexes (Appendix Table A.14).

64Progress has been especially marked in the fight against cerebrovascular diseases, for which the standard mortality rate is in steady decline. It has fallen by almost 75% for both sexes since 1980. Deaths from ischaemic heart diseases have also fallen considerably, with a decrease of 50% in the male standardized rate and of 60% in the female rate over the same period. While the decline has been quite linear for cerebrovascular diseases, the decline in deaths from ischaemic heart diseases accelerated between 1980 and 2007 at a rate which almost doubled for men and tripled for women between 1980-1990 and 2000-2007 (Appendix Table A.14). It is this very favourable trend in the diseases most common at advanced ages which explains the substantial progress achieved in mortality above age 65 (Meslé, 2006).

An accelerating decline in cancer mortality

65Over the same period, the contribution of cancer has followed a trend that is inversely proportional to that of cardiovascular diseases (Appendix Table A.14). Cancers are the leading cause of death in France today. Despite a significant decrease in the standardized rate (–22% for men, –18% for women between 1980 and 2007), the share of cancers in overall mortality has increased from 26% to 35% among men and from 22% to 32% among women. For women, mortality from all types of cancer is nonetheless only half that observed for men (121 versus 247 per 100,000 in 2007).

66The contribution of cancer mortality is especially high at ages 45-65 and only slightly lower after age 65 (Table 7). At these ages, it represents around half the standardized rate for all reported causes: 48% for men and 57% for women at ages 45-65; 47% and 45% at ages 65-80. The acceleration of progress over the last 20-25 years is nonetheless encouraging. This progress concerns the main cancers affecting each sex, i.e. lung cancer among men (with a decline in the standardized rate from 70 to 62 per 100,000 between 1990-1996 and 2007) and cancers of the breast and uterus among women (whose combined rate fell from 38 to 30 per 100,000 between 1980 and 2007). But an increase in lung cancer has been observed among women, with a rate that rose from 6 to 15 per 100,000 over the same period. These contrasting trends between the sexes reflect the much more recent decline in smoking among women than among men. The number of male smokers is still one-third higher than that of female smokers (Arwidson et al., 2004).

Other causes of death

67“Other diseases” constitute the third major cause of death in 2007, as was already the case in 1980 (Appendix Table A.14). In most cases, these other diseases are linked to particular causes of death either in childhood or more especially in old age. The causes of death of young children and adolescents (0-15 years) are very specific: congenital abnormalities and childhood diseases for the most part, but also accidental deaths. All other diseases represent almost three-quarters of the standardized rate for all causes before age 15 (Table 7). After the first year of life, mortality is very low, however. It is lowest at ages 9-10, when the risk of dying is below 1 in 10,000. Among the oldest-old (age 80 and above), cancers and cardiovascular diseases predominate, although the other diseases still represent 17% of standardized mortality, all causes, for men, and 23% for women (Table 7).

Table 7

Standardized mortality rates by age groups in 2007* (per 100,000) and distribution by cause of death (%)

Cause of death Age group Age 0-14 Age 15-24 Age 25-44 Age 45-64 Age 65-79 Age 80+ All ages Males Standardized rates, all causes (per 100,000) 4 6 13 70 245 1150 682 Infectious diseases 2.7 0.7 2.8 1.8 1.6 2.0 1.9 Cancers 7.4 9.7 17.1 48.0 46.5 24.9 37.0 Cardiovascular diseases 2.4 3.6 11.0 17.6 24.9 36.4 25.6 Respiratory diseases 1.9 1.2 1.8 3.0 5.9 10.2 6.3 Cancers of the digestive organs 1.3 0.5 5.4 8.1 5.0 3.7 5.3 Other diseases 71.7 9.8 13.1 9.9 11.0 17.4 13.8 Deaths from external causes 12.6 74.4 48.8 11.7 5.0 5.3 10.2 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Females Standardized rates, all causes (per 100,000) 3 2 6 31 112 765 358 Infectious diseases 2.2 2.5 3.0 1.6 1.8 2.0 1.9 Cancers 8.0 17.4 38.8 56.6 44.9 17.6 34.5 Cardiovascular diseases 3.1 4.3 10.2 11.8 23.2 40.2 26.9 Respiratory diseases 2.3 2.1 1.9 2.5 4.7 7.7 5.3 Cancers of the digestive organs 1.1 0.8 4.8 6.9 5.2 4.2 5.0 Other diseases 72.0 18.7 13.8 10.8 15.2 23.1 18.7 Deaths from external causes 11.3 54.2 27.5 9.7 4.9 5.3 7.7 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 * These rates are slightly different from those of Appendix Table A.14 because of the calculation method used. Ill-defined causes have been distributed across other causes. For a definition of the major groups of causes and of the method used to distribute ill-defined causes, see Meslé (2006). Source: Calculations by Meslé (2006) updated using INSERM statistics (CépiDc).

Standardized mortality rates by age groups in 2007* (per 100,000) and distribution by cause of death (%)

68Death from external causes is by far the leading cause of death for young people aged 15-25 (69% of the standardized mortality rate for all causes) and remains so at ages 25-45 (42% of the rate for all causes). While road traffic accidents account for 53% of the standardized rate for all external causes among men and 49% among women at the youngest ages, suicide is the main killer at ages 25-45. It is the leading cause of death among men aged 23-34 and the second for women, after cancers (Aouba et al., 2009).
Last, the standardized rate of deaths from respiratory diseases is still decreasing in stages, although its contribution to the total number of deaths remains high as the population grows older.