Housing and Household Size in Local Population Dynamics

Alfred Dittgen; Catriona Dutreuilh

Housing and Household Size in Local Population Dynamics
The example of Paris
Suivre cet auteur Alfred Dittgen, Translated from the French by Suivre cet auteur Catriona Dutreuilh
In Population Volume 60, Issue 3, July 2005, pages 259 to 298

format_quote Cite share Share Switch to French

1Compared with other major European capitals, the city of Paris, enclosed in an area of 100 square kilometres, is confined within very narrow administrative boundaries. In this article, Alfred Dittgen shows how this constraint, which limits new housing construction, has affected the population of Paris over the last fifty years, with an outward migratory trend towards the agglomeration of which Paris forms the centre. This very real population decline is the combined result of a reduction in household size and in the number of primary residences over the period. At the same time, the demographic structure and social profile of the city and its districts has changed substantially.

2As in other cities, the population dynamics of Paris are dependent upon external factors. But this study provides a valuable decision aid for policymakers in charge of a territory whose boundaries are fixed.

3As a general rule, population change in a country depends more on natural change than on migration. In other words, the difference between the numbers of births and deaths is generally higher than net migration [1]. The population of a region, on the other hand, is often shaped more by migration than by births and deaths. This is even more true for the population of a locality such as Paris intra muros, where there are twice as many arrivals as births, and four and a half times more departures than deaths (Table 1).

Table 1

Paris intra muros. Mean annual arrivals and departures over the period 1990–1999 (rounded numbers)

Arrivals Immigration Births +70,200 +30,500 Departures Emigration Deaths – 85,000 – 18,700 Net change Migration Natural –14,800 +11,800 Total net change –3,000 Sources: INSEE, population censuses and population movements.

Paris intra muros. Mean annual arrivals and departures over the period 1990–1999 (rounded numbers)

At this local level, natural change and migration are closely linked —though this is not the case at national level—via the housing factor. A new birth may cause a family to move to a new locality; a dwelling left vacant by the death of its last occupant may be reoccupied by a household from another locality.

4Generally speaking, dwellings—more specifically the number of dwellings and its variation over time—have a major, if not decisive, impact on local population size and population change wherever families and adult generations are no longer willing to cohabit, as is the case in Western countries. For this reason, such a population cannot generally increase unless the number of dwellings increases. Indeed, this increase is often necessary simply to maintain the status quo. Firstly, because the share of dwellings occupied by the residents in relation to the share of second residences and vacant dwellings may decrease. And secondly, because the number of people occupying each dwelling, i.e. the household size, may also decrease, as is the case throughout the Western world.

5In this relationship between population and housing, the characteristics of the dwellings—primarily their size—are naturally of key importance. Dwelling size determines the size of households, but also their age-sex structure. Large dwellings are often occupied by families, i.e. adults and children of all ages, whereas small dwellings are occupied mainly by singles or couples without children, i.e. young or elderly adults. Another important characteristic is the tenure status of the household. Social housing tenants are often families, while private sector tenants, paying higher rents, are often singles.

6These relationships between population and housing are especially evident in Paris, doubtless more so than in other French cities, and more so than in many other European capitals, since Paris has a number of very specific features (see Appendix).

7Paris is almost entirely urbanized, so although the housing stock has increased substantially since the Second World War, there is now little scope for further development, except through the construction of tower blocks, though this alternative has been practically ruled out [2]. Many secondary and occasional residents are attracted to Paris, thereby reducing the number of dwellings occupied by permanent residents, and this reduction is further exacerbated by high vacancy rates.

8So the size of Parisian households is decreasing, following the same trend as French and Western households in general, and for the same key reasons, namely the “couple crisis”, with later couple formation and more frequent dissolutions, and population ageing. But the decrease in city-centre household size is also, and above all, due to the appeal of these districts for small households and singles in particular. This selection of small households is more pronounced in Paris than elsewhere because the city centre houses only a small proportion of the overall population of the agglomeration.

9This centripetal movement of small households is also favoured by the fact that dwellings in Paris are smaller than in the rest of the agglomeration and, due to their scarcity, more expensive, so families are often excluded. In addition, a larger proportion of these dwellings are rented and therefore tend to attract young adults and young couples.

10In this article [3] we will examine the relationships between population and housing in Paris intra muros over the last half century (more precisely since the 1954 census), a period in which the Parisian population has undergone unprecedented change. We will conclude by examining data obtained with a population projection model which, given the importance of the relationship in question, and unlike the models generally used [4], integrates the dwelling and its occupancy.

I – Population, dwellings and primary residences

1 – A shrinking population, an expanding housing stock

11The population of Paris peaked between the 1911 and 1954 censuses, fluctuating between 2.8 and 2.9 million people over the period [5]. It then declined steadily, falling by a quarter to just above 2 million in the 1999 census (Table 2 and Figure 1). This depopulation was rapid in the 1960s and 1970s, with an annual decline of 1.25% on average between 1962 and 1982, and much more moderate from then on, with a reduction of 0.14% annually between 1982 and 1999.

Table 2

Population change (total, household, non-household) between 1954 and 1999 in Paris intra muros and intercensal net change in total population (thousands)

Population Census 1954 1962 1968 1975 1982 1990 1999 Total population 2848.8 2798.8 2586.6 2296.9 2176.7 2152.3 2125.9 Household population 2782.4 2732.3 2525.1 2245.8 2127.7 2101.7 2077.8 Non-household population 66.4 66.5 61.5 51.1 49.0 50.6 48.1 Intercensal net change (total population) Period 1954-1962 1962-1968 1968-1975 1975-1982 1982-1990 1990-1999 1954-1999 Total net change –50.0 –212.2 –289.7 –120.2 –24.4 –26.4 –722.9 Mean annual net change –6.3 –35.4 –41.4 –17.2 –3.1 –2.9 –16.1 Sources: INSEE, population censuses and population movements.

Population change (total, household, non-household) between 1954 and 1999 in Paris intra muros and intercensal net change in total population (thousands)

Figure 1

Population change in Paris intra muros between 1954 and 1999 (thousands)

12Over this same interval, the number of dwellings rose from 1.2 to 1.3 million. This apparently paradoxical increase, observed in practically all the intercensal periods, was especially rapid from 1975 to 1982, with 6,000 additional dwellings an average per year (Table 3). Though this increase is the logical result of new construction, the total housing stock did not increase in line with the number of new dwellings, firstly because many existing dwellings were converted into offices or warehouses—a trend which explains the drastic decrease in housing stock in central districts (infra)—and secondly because old buildings were demolished to make way for new ones [6]. Hence, during the last intercensal period, 1990-1999, during which 5,000 new dwellings were built annually, the housing stock increased by only 2,000 units per year. Since very few apartments are now converted for other uses, the difference between these two numbers is due mainly to the demolition of old buildings, a practice that has become increasingly necessary to free up space for new construction. And in the relationship between population and housing, it is the net increase in housing stock which counts [7].

2 – Fewer dwellings occupied by Parisians

13With regard to population change, the dwellings which count are more specifically those occupied by residents, i.e. primary residences, which, according to French statistical convention, each represent one household. The numbers of these dwellings do not change in the same way as dwellings as a whole, since they are in competition with dwellings occupied as second or occasional residences and with vacant dwellings. In Paris, contrary to the total number of dwellings, the number of primary residences fell between 1954 and 1999, from just under 1.2 million to slightly more than 1.1 million. This diverging trend can be explained by a very rapid parallel increase in second and occasional [8] residences and in vacant dwellings [9] (Table 3 and Figure 2).

Table 3

Change in the number of dwellings by category in Paris intra muros between the 1954 and 1999 censuses and intercensal net change (thousands)

Number of dwellings Census 1954 1962 1968 1975 1982 1990 1999 All dwellings 1,206.0 1,234.7 1,214.8 1,237.6 1,279.3 1,304.3 1,322.6 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Primary residences 1,185.3 1,191.7 1,140.2 1,109.1 1,097.4 1,095.1 1,110.9 98.3% 96.5% 93.9% 89.6% 85.8% 84.0% 84.0% Second and occasional residences 13.3 23.0 35.4 37.8 56.4 90.9 75.1 1.1% 1.9% 2.9% 3.1% 4.4% 7.0% 5.7% Vacant dwellings 7.4 20.0 39.2 90.7 125.5 118.3 136.6 0.6% 1.6% 3.2% 7.3% 9.8% 9.1% 10.3% Intercensal net change Period 1954-1962 1962-1968 1968-1975 1975-1982 1982-1990 1990-1999 1954-1999 All dwellings Total net change 28.7 –19.9 22.8 41.7 25.0 18.3 116.6 Mean annual net change 3.6 –3.3 3.3 6.0 3.1 2.0 2.6 Primary residences Total net change 6.4 –51.5 –31.1 –11.7 –2.3 15.8 –74.4 Mean annual net change 0.8 –8.6 –4.4 –1.7 –0.3 1.8 –1.7 Sources: INSEE population censuses.

Change in the number of dwellings by category in Paris intra muros between the 1954 and 1999 censuses and intercensal net change (thousands)

Figure 2

Change in the number of second and occasional residences and of vacant dwellings in Paris intra muros between 1954 and 1999 (thousands)

This sharp increase in dwellings not occupied as primary residences was made possible by a general increase in housing supply following a period of severe shortage following the war. A growing number of dwellings serve as a pied-à-terre for provincials and foreigners attracted to Paris. Moreover, in a city where household turnover is high due to the much larger proportion of private sector tenants than elsewhere in France (infra), the periods of inoccupancy between two tenants have tended to lengthen, leading to a higher vacancy rate.

14The number of primary residences has varied over time. It increased slightly during the last intercensal period, rising from 1.095 million in 1990 to 1.111 million in 1999. But this 1.4% increase did not raise the household population, which fell from 2.102 million to 2.078 million (a 1.1% decrease) over the same period. This non-parallel variation is observed throughout the period from 1954 to 1999. During this time, the number of primary residences fell by 6% from 1.185 million to 1.111 million, while the household population fell four times faster, from 2.782 million to 2.078 million (a 25% decrease). These diverging trends, just as surprising as those of the population and of dwellings, can be explained, as we shall see, by the decrease in household size.

II – Population and household size

1 – Small households and small dwellings

15The average number of persons per household in Paris in 1999 was 1.87. This is very low compared with France as a whole (2.40) and with the Paris region (2.38). Household size has fallen considerably over the last half-century, as is the case throughout France and in industrialized countries in general (Alders and Manting, 2001), though this decrease began earlier in the capital—from the 1950s—while elsewhere in the country it was not observed until the 1970s. It reflects a rapid increase in the proportion of one-person households at the expense of larger ones (Table 4). These one-person households, forming just under one-third of the total in 1954 (32.1%), represented more than half by 1999 (52.4%).

Table 4

Total number of households in Paris intra muros, Average size and breakdown by size. Change between 1954 and 1999 (in thousands and %)

Census 1954 1962 1968 1975 1982 1990 1999 Total number of households 1,185.3 1,191.7 1,140.2 1,109.1 1,097.4 1,095.1 1,110.9 % 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Average household size 2.35 2.29 2.21 2.02 1.94 1.92 1.87 One-person households 379.9 416.7 432.4 487.6 528.8 545.8 581.7 % 32.1 35.0 37.9 44.0 48.2 49.8 52.4 Two-person households 386.8 376.2 350.3 328.3 304.7 294.7 288.1 % 32.6 31.6 30.7 29.6 27.8 26.9 25.9 Households of 3 persons or more 418.6 398.8 357.5 293.2 263.9 254.6 241.1 % 35.3 33.5 31.4 26.4 24.0 23.2 21.7 Household population 2,782.4 2,732.3 2,525.1 2,245.8 2,127.7 2,101.7 2,077.8 Sources: INSEE population censuses.

Total number of households in Paris intra muros, Average size and breakdown by size. Change between 1954 and 1999 (in thousands and %)

16The current difference in household size between Paris and the Paris region is linked to a difference in dwelling size: dwellings in Paris have 2.53 rooms on average, compared with 3.29 [10] for the Paris region. As Parisians have, on average, as much space per person in their homes as their neighbours in the Paris suburbs—1.35 rooms vs. 1.38 (Table 5)— their households are necessarily smaller [11]. Beyond these somewhat uninformative averages, we note that one-third (32.2%) of dwellings in the Paris region are small one- or two-room dwellings, and that one-third (34.6%) of households comprise only one person. In Paris, these proportions are respectively 56.1% and 52.4%.

Table 5

Household size and dwelling size in France, Paris region and Paris intra muros in 1999

France Paris region Paris Average number of persons per household 2.40 2.38 1.87 Average number of rooms per dwelling 3.86 3.29 2.53 Average number of rooms per person 1.61 1.38 1.35 Proportion of one-person households (%) 31.00 34.60 52.40 Proportion of one- or two-room dwellings (%) 19.10 32.20 56.10 Source: INSEE, 1999 population census.

Household size and dwelling size in France, Paris region and Paris intra muros in 1999

17The reduction in Parisian household size between 1954 and 1999 was accompanied by a paradoxical increase in dwelling size, which rose from 2.23 rooms in 1954, to 2.53 rooms in 1999. This can be explained by the fact that more large dwellings than small ones were built over the period, with the share of dwellings of four rooms or more rising from 13.1% to 20.8%, and that of one-room dwellings—studios and furnished rooms— falling from 33.9% to 24.0%. This did not halt the decline in the household population, though it did succeed in slowing down the trend [12]. If the average number of rooms per dwelling had remained the same as in 1954, taking account of the increase in the average number of rooms per person, the population in 1999 would have been 1.87 million [13], rather than the 2.13 million actually observed.

2 – Paris attracts small households

18The reduction in household size in Paris is the result of factors identical to those affecting the whole of France (Figure 3), and industrialized countries in general, but also of factors specific to the capital. In France, two factors were important in previous decades, but no longer play a major role: residential separation of adult generations and a reduction in the number of children per family. Households of cohabiting adults are now rare. The number of children in households has stopped falling now that fertility has stabilized and that grown-up children tend to remain longer in the parental home. But two other factors still play a role: the growing proportion of people living alone compared with those living in couples, and population ageing. With the later age of first union, the increasing frequency of union dissolution, the declining frequency of subsequent unions and their postponement, the number of adult singles is continuing to increase. And population ageing, which increases the proportion of couples with no children and of persons who have lost their partner, is an ongoing process.

Figure 3

Paris and France. Change in average household size between 1954 and 1999

19Note, however, that population ageing only reduces the number of persons per household because adult cohabitation is rare in today’s society and because this non cohabitation is made possible by a plentiful supply of housing.

20These sociodemographic phenomena underlying the reduction in household size have been observed, and are still being observed, in Paris as elsewhere. Nevertheless, Paris, like other city centres, tends to attract small households, singles in particular, for a number of reasons: these households are ready to accept the constraints of a small dwelling; with no children to raise, they can afford the very high cost of Parisian real estate; and lastly, the facilities and leisure activities offered by the capital are more tailored to their needs than to those of families. This selection of the agglomeration’s small households is all the more striking, given that the population of Paris itself represents only a small proportion of that of the agglomeration as a whole—less than one-fifth [14] compared with one-third to one-half for the other French cities—and that this share is decreasing, since the population of Paris is falling and that of the agglomeration is increasing.

21Paris is especially attractive to young adults, a fact which explains the predominance of the 25-39 age group, in which singles are more numerous: more than one-third of the people in this age group live alone. And it is this high proportion of singles, and not the lower fertility of couples, which explains the low birth rate in Paris, around 1.5 children per women between 1990 and 1999, versus 1.8 in the Paris region and in France.

3 – Smaller household size, the main cause of depopulation

22Between 1954 and 1999, the population of Paris fell by 25%. During this interval, the number of primary residences fell by 6% (Table 6) and the average household size fell from 2.35 to 1.87 persons, a 20% reduction. So this second factor is the key contributing factor to the depopulation of Paris observed over the last half century [15]. The reduction in household size alone would have reduced the population from 2.86 to 2.26 million [16], while the reduction in the housing stock, with size remaining constant, would have brought it down to 2.66 million [17].

23This reduction in household size was the main contributing factor to depopulation in each of the intercensal periods, with the exception of the period 1962-1968, when the sharp decrease in primary residences was the main factor (Table 6).

Table 6

intercensal rate of change in total population, household population, number of primary residences and household size in Paris intra muros (%)

Period 1954-1962 1962-1968 1968-1975 1975-1982 1982-1990 1990-1999 1954-1999 Total population –1.76 –7.58 –11.20 –5.23 –1.12 –1.23 –25.38 Household population –1.80 –7.58 –11.06 –5.26 –1.22 –1.14 –25.32 Number of primary residences +0.54 –4.32 –2.73 –1.05 –0.22 +1.44 –6.28 Household size –2.33 –3.41 –8.57 –4.26 –1.01 –2.54 –20.32 Sources: INSEE population censuses.

intercensal rate of change in total population, household population, number of primary residences and household size in Paris intra muros (%)

We will now examine the age pyramid in Paris which, like the total population, is closely linked to housing.

III – Age-sex structure, migration and housing

1 – Young adults and women are over-represented

24To understand the specificity of the age structure [18] in Paris in 1999, it is useful to examine that of the region and of the country on the same date (Figure 4). The shape of the Paris region age pyramid is similar to that of France. However, the proportion of elderly persons—aged 65 and above—is smaller than that of young adults aged 20-39. These differences are due to migration: the Paris region attracts young adults from the French regions and from abroad, while its pensioners tend to move away.

Figure 4

Age-sex distribution of the population of France, the Paris region and Paris intra muros by five-year age group for a total of 1,000 inhabitants. 1999 census

In the Parisian age pyramid, the proportion of young adults is even higher, between the ages of 25 and 40 especially. Conversely, the proportion of old people is much closer to that of France. Moreover, the proportion of young people aged 0-19 is much lower than in the region and the country. These particularities are also due mainly to migration in and out of Paris.

25The resulting “ace of spades” profile of the age pyramid is not a recent phenomenon. The shape of the pyramid based on the previous census results (Figure 5) is practically the same, and the two figures look almost identical at first sight, except for the position of the smaller cohorts born during the 1914-18 War. The growing proportion of young adults (age 20-39) first becomes noticeable on the 1968 age pyramid, where it represents 31.4% of the total, compared with 29.9% in 1954. And it has increased steadily since then, reaching 36.0% in 1999. The more massive 1954 pyramid differs from that of the country as a whole, mainly by the smaller proportion of children and young people.

Figure 5

Age-sex distribution of the population of Paris intra muros by five-year age group at different census dates for a total of 1,000 inhabitants

26Between 1954 and 1975, the Parisian population aged faster than that of France, with a growing proportion of both “elderly” persons aged 65 and over and of “very elderly” persons aged 80 and above (Table 7). But since then, the proportion of old people in Paris has declined, though it has continued to increase in France as a whole. The proportion of “very elderly” continued to increase until 1990, and has been declining rapidly since then, though it is still higher than in the country as a whole [19].

Table 7

Proportions of persons aged 65+ and persons aged 80+ between 1954 and 1999 in France and in Paris intra muros (%)

Census 1954 1962 1968 1975 1982 1990 1999 65 + France 11.5 11.8 12.6 13.4 13.5 13.9 15.8 Paris 11.0 13.5 16.0 18.0 17.1 16.0 15.4 80 + France 1.7 2.1 2.2 2.4 3.0 3.7 3.7 Paris 1.4 2.2 2.8 3.6 4.6 5.2 4.5 Sources: INSEE population censuses.

Proportions of persons aged 65+ and persons aged 80+ between 1954 and 1999 in France and in Paris intra muros (%)

27In 1999, like practically all populations, the French population comprised more women than men due to excess male mortality (51.5% vs. 49.5%). The percentage of women is practically the same in the Paris region (51.6%), though it is substantially higher in Paris (53.1%). This over-representation of females in Paris is decreasing, however, since in 1954 the capital comprised 55.0% women, compared with 51.8% in France.

2 – An age-sex structure determined by migration

28Figure 6 gives the mean annual coefficients of age-specific population change due to migration between the last two censuses [20]. They were obtained by determining the ratio of the population size of each birth cohort in 1999 to that of 1990, after neutralizing the effect of mortality. The same calculation for the 1982-1990 intercensal period gives relatively similar values. These values differ more substantially for previous periods, though their age-specific variation produces similar curve profiles, which shows that this migration pattern is very old.

Figure 6

Coefficients of age-specific population change in Paris intra muros due to migration. Mean annual values 1990-1999

29Values well above one around age 20 reflect the inflow of students and young people taking up a first job. This inward migration explains the predominance of young adults in the pyramid. These young people come mainly from the provinces, but also from abroad and increasingly from the Paris suburbs [21]. There are more women than men among these new arrivals, leading to a general over-representation of women in the Parisian population. In the 20-24 age group, which counts the highest number of immigrants of this kind, the proportion of women is 54.2%, compared with 49.6% for the country as a whole.

30Coefficients below one around age 30 can be linked to the similarly low coefficients for young children. They reflect the outflow of families towards the suburbs and the provinces. This movement explains the narrowing of the pyramid from age 40, and at the base (newborns and young children). There were only around 20,000 children aged 5 in 1999, though we saw above that the average annual number of births between 1990 and 1999 topped 30,000!

31The values below one at around age 60 correspond to the outward movement of newly retired persons towards the provinces. Up to these ages, the values for both sexes are fairly similar. They start diverging at age 70, with a reduction in mobility among men and, on the contrary, a second wave of emigration among women. This divergence reflects the fact that women more often outlive their husbands than vice versa, and certain widowed women choose to leave Paris to join a household of their family outside the capital.

32The ageing of the Parisian population, like that of France as a whole, is due to the decline in the birth rate since the 1970s and the increase in life expectancy among the elderly. However, the differences with respect to the population of France have little to do with differences in natural change. Once again, they are explained by migration. The migration patterns of the successive Parisian birth cohorts are different from those of the rest of France. Hence, the current decrease in average age, i.e. the declining proportion of old people, can be linked to the fact that a larger share of the birth cohorts currently reaching retirement age now leave the capital than in the past.

3 – Migration patterns closely linked to dwelling characteristics

33Not all types of migration are linked to housing, notably the outflow of pensioners, which can be explained by their desire to retire to the country, to a sunnier climate, or to their region of origin. Migration to and from the suburbs is strongly conditioned by this factor however. After the Second World War, there was a severe housing shortage in France, in Paris especially, as shown by the very small proportion of vacant dwellings and second residences in 1954, and by the small average number of rooms per person: 0.95 (2.23 rooms / 2.35 persons), compared with 1.35 today. Many new homes were built in Paris in the 1960s and 1970s, though the pace of construction was even faster in the suburbs, enabling many young families to settle outside Paris, thanks also to improvements in the public transport system. The capital thus began to “specialize” in small dwellings, for young households especially.

34Moreover, the young provincials arriving in the Paris region choose to live in Paris not only because more jobs are available in the capital [22], but also because, as we have just seen, singles and childless couples prefer to live in Paris rather than the suburbs.

35Living in Paris is also facilitated by a proportion of dwellings in the private rental sector which, though declining, is higher than elsewhere: in 1999, 42% of households were private sector tenants, compared with 25% in the Paris region and 22% in France. This factor, and the dwelling tenure status in general, also has a major influence on the age structure of the population, as we shall see.

IV – Households by tenure status and age of dwelling

36The census questions identify five types of tenure: 1) owner; 2) private sector tenant; 3) social housing tenant; 4) non rent-paying tenant [23]; and 5) tenant of a furnished dwelling or a hotel room.

1 – Owners enjoy more space than tenants

37The number of persons per dwelling varies according to tenure status. The size of owner households (1.90 persons) is close to that of households in general (1.87). The households of private sector tenants (1.79) and non rent-paying tenants (1.80) are slightly smaller, while those of social housing tenants are much larger (2.20) and those of tenants of furnished accommodation or hotel rooms are substantially smaller (1.34).

38These differences in household size are naturally linked to dwelling size (Table 8). Hence, the household size of social housing tenants, larger than that of private sector and non rent-paying tenants, reflects a larger dwelling size and not a higher occupation density. In fact, the members of these three categories all have the same amount of floor space: 1.24 to 1.25 rooms on average per person. On the other hand, the small size of households in furnished accommodation, due mainly to the fact that this type of dwelling has very few rooms—notably hotel rooms which, by definition, comprise only one—is also a consequence of restricted living space, with only 1.05 rooms per person. Likewise, though owner households have more members than private sector tenant households, it is because they have more space (1.63 rooms).

Table 8

Household size and dwelling size by type of tenure in Paris intra muros in 1999

Type of tenure Owner Unfurnished dwelling private sector tenant Unfurnished dwelling social housing tenant Non rent-paying tenant Tenant, furnished dwelling or hotel room Total Proportion of households (%) 29.6 41.8 16.7 7.3 4.6 100.0 Average household size 1.90 1.79 2.20 1.80 1.34 1.87 Average number of rooms of primary residences 3.09 2.24 2.72 2.24 1.41 2.53 Average number of rooms per person 1.63 1.25 1.24 1.24 1.05 1.35 Proportion of one-person households (%) 49.0 55.0 43.0 55.8 78.5 52.4 Proportion of one- or two-room dwellings (%) 38.9 67.2 43.9 67.3 91.3 56.1 Source: INSEE population census.

Household size and dwelling size by type of tenure in Paris intra muros in 1999

39The amount of space available to households is naturally linked to their socioeconomic status. The limited space available to persons living in furnished accommodation is an aspect of the economic insecurity widespread among this sub-population. Conversely, the extensive space available to members of owner households reflects their economic affluence. However, as we shall see, this lower occupation density also results from the fact that owner households are often older, with grown-up children who may already have left home.

2 – A Parisian pyramid whose shape reflects that of private sector tenants

40The pyramids corresponding to the five types of tenure have very contrasting profiles (Figure 7). The pyramid of social housing tenants is the only one which resembles the country pyramid (Figure 4), indicating that it includes the same diverse range of households. Note however that women aged between 25 and 55 are over-represented with respect to the national pyramid, indicating a stronger presence of women living alone and of lone-mother families.

Figure 7

Age-sex distribution of the household population of Paris intra muros by tenure status in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

The pyramid of owner households is older, with fewer children. Buying a home takes time and the adults in these households generally stay in their home until the death of the last occupant.

41For the three other categories, the pyramid characteristics are similar to those of the Parisian pyramid, though even more pronounced: a large proportion of young adults, small proportions of children and old people. These characteristics are especially marked among the population living in furnished accommodation, which essentially comprises adults of working age, and a majority of men. Note also that the pyramid of non rent-paying households, though very similar to that of private sector tenant households, includes far fewer old people since a substantial proportion of these households are housed by their employer and are therefore of working age [24]. Given the similarities between these three populations and the limited size of two of them—non rent-paying households and those living in furnished accommodation—, they will be combined into a single category called “private sector and non rent-paying tenants” to examine the age structure in relation to dwelling age.

3 – Owner households and social housing tenants stay put for life

42The age pyramids of the owner population by dwelling age (Figure 8) shows that for dwellings built since the Second World War, the older the dwelling, the older the pyramid. The average age of occupants rises from 38.6 in dwellings built between 1990 and 1999 to 51.8 in those built between 1948 and 1967 (Table 9). This shows that a substantial proportion of this sub-population has been living in the same dwelling since it was first built. In other words, household age increases with dwelling age.

Figure 8

Age-sex distribution of the owner household population of Paris intra muros by building age in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

Table 9

Average age of household members by type of tenure and building age in Paris intra muros in 1999

Building age Type of tenure Owners Social housing tenants Private sector tenants and non rent-paying tenants Before 1915 45.8 40.3 36.1 1915 - 1948 46.6 42.8 36.8 1949 - 1967 51.8 43.9 38.1 1968 - 1974 50.1 41.4 37.1 1975 - 1981 47.3 38.2 36.8 1982 - 1989 41.4 32.7 35.0 1990 - 1999 38.6 28.6 31.5 Total 46.9 38.6 36.3 Source: INSEE population census.

Average age of household members by type of tenure and building age in Paris intra muros in 1999

On the other hand, households living in dwellings built between the Wars (from 1916 to 1948) have a younger age structure, with an average age of 46.6. A substantial proportion of the people who first occupied these older dwellings have died and been replaced by others. As this renewal is a gradual process, the pyramid in question is much more regular than the previous ones. This renewal is complete in dwellings built before the First World War (before 1915), in which the population is even younger, with an average age of 45.8.

43The population of social housing tenants is younger than the owner population (Table 9), which is logical since people often enter this sector when they start to found a family. Nevertheless, for dwellings built since the Second World War, the average age varies in the same way as for owners: 28.6 in dwellings built between 1990 and 1999, compared with 43.9 in those built between 1948 and 1967. It decreases likewise for the oldest dwellings. In dwellings built before the First World War, the average age is 40.3. Like owners, social housing tenants tend to “stay put” for life. According to the INSEE surveys (Laferrère and Warsée, 2002), this observation is true for France as a whole, but more pronounced in the Paris region. This difference is clearly linked to property prices, which do not encourage households lucky enough to occupy social housing in the capital to move elsewhere.

4 – High mobility among private sector tenants

44The pyramids of private sector and non rent-paying tenants, unlike the previous ones, vary little by dwelling age (Figure 9), with similar average ages of between 35 and 38 (Table 9), except in the case of dwellings completed in the 1990s, which are occupied by younger households. Private sector dwellings are expensive, so their occupants do not stay for long. They are also smaller and hence more suited to people living alone or childless couples. There is a high turnover in this sector, with few tenants remaining in the same dwelling for long periods, and this explains the similar age structures observed for dwellings of all ages. Note however that the two pyramids for the most recent periods are slightly different, with more children, the one for the period 1982-1988 especially. The dwellings concerned are larger, with 2.5 rooms on average compared with 2.1 for all dwellings, so they are more suited to family occupancy.

Figure 9

Age-sex distribution of the population of private sector and non rent-paying tenant households in Paris intra muros by building age in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

V – Populations and dwellings in the arrondissements

45Once again, the population trends and characteristics in the different Parisian arrondissements can be explained by the changing characteristics of the housing stock (see map in Appendix).

1 – Parallel trends

46In practically all the Parisian arrondissements, the population declined between 1954 and 1999. In the central arrondissements, it fell substantially, with a halving of the population in the 1st and 9th, and a decline of between one-quarter and one-third in the 10th and 11th. In the outer arrondissements, the population fell likewise by between one-quarter and one-third in the 14th, 16th, 17th and 18th, while in the 12th, 15th and 20th the decrease was smaller. The only arrondissements with population growth were the 13th, where it was minimal, and the 19th, where it was slightly higher. In line with the population trend for the whole of Paris, the period of maximum decrease was between 1962 and 1975.

47These trends can be linked to those affecting primary residences (Table 10). Figure 10, which compares changes in household population with changes in the number of primary residences, shows that the two are closely linked (linear correlation coefficient r = +0.97). Beyond this predictably strong relation, the comparison once again shows that it is not sufficient to increase the number of dwellings in order to raise the population, nor even to prevent it from falling. Hence, in the 13th arrondissement, where the number of dwellings increased by almost one-third between 1954 and 1999, the population has risen by a mere few percent, and in the 19th, where the housing stock has increased by a quarter, the population has grown by less than 11%. The nine central arrondissements, from the 1st to the 9th, have lost one-third of their housing stock but half their population. As we have already seen, the fact that changes in population size do not tally with changes in the housing stock can be explained by the decrease in the average household size.

Table 10

Changes in population size and number of dwellings by arrondissement of Paris between 1954 and 1999 (%)

Arrondissement Total population Household population Dwellings Primary residences 1st –56.5 –56.5 –24.0 –44.0 2nd –55.1 –55.1 –26.4 –44.9 3rd –47.3 –47.6 –13.5 –33.7 4th –54.0 –54.0 –22.4 –36.8 5th –44.9 –44.7 –14.1 –28.5 6th –49.0 –48.5 –1.4 –31.2 7th –45.5 –45.1 –6.5 –23.5 8th –51.2 –50.9 –23.1 –37.7 9th –45.3 –45.2 –17.6 –32.7 10th –30.6 –30.2 –1.8 –17.7 11th –25.6 –26.2 +6.9 –7.9 12th 13.7 12.4 +25.4 +12.0 13th +3.6 +2.7 +45.6 +29.6 14th –26.8 –27.7 +9.0 –5.2 15th –9.7 –7.7 +37.8 +19.2 16th –24.3 –23.6 +30.9 +8.4 17th –30.7 –30.7 +2.0 –13.0 18th –30.8 –31.1 –0.9 –15.5 19th +11.5 +10.6 +38.6 +24.5 20th –8.4 –9.2 +20.9 +8.6 Paris –25.4 –25.3 +9.9 –6.2 Sources: INSEE population censuses.

Changes in population size and number of dwellings by arrondissement of Paris between 1954 and 1999 (%)

Figure 10

Comparative change in the household population and in the number of primary residences in the arrondissements of Paris between 1954 and 1999

48This being the case, then why has the housing stock, and even more so the stock of primary residences and hence the number of households, decreased in most arrondissements? In the outer arrondissements, in the 1970s especially, the stock was increased by building large numbers of new dwellings, many of them in the form of high-rise blocks, on the plots of land still available [25] in this part of the capital. By contrast, very few new dwellings were built in the central arrondissements due to the lack of available land and the reluctance to replace existing building with high-rise blocks, so their housing stock decreased. Moreover, with the development of commercial activities in the city centre, large numbers of dwellings were converted into offices, shops, warehouses and other business premises. Lastly, in 1999, the proportion of dwellings used as second or occasional residences rather than as primary residences was higher than in the outer districts due to the strong appeal of central Paris for foreigners and wealthy provincials. A full 20% of dwellings in central Paris (1st to 11th arrondissements) were not occupied by residents (second and occasional residences and vacant dwellings) compared with 14% in other arrondissements and a general average of 16%. The proportion tops 25% in the 1st and 2nd arrondissements and is close to one-third (30%) in the 6th.

2 – Three population models, and one dominant model

49The populations of the different arrondissements are changing in different ways, and their age structures are also very dissimilar. They can be grouped into three models (Table 11 and Figure 11):

a model in which the number of young adults, aged 20-39, is proportionally higher than the average [26], with an
“ace of spades” pyramid shape that is even more pronounced than that of the total population. The 2nd arrondissement is the most representative of this model, followed by the 3rd. They are inhabited by a high proportion of young adults living alone (Table 12). It is also known that many are managers or professionals (Dittgen, 2002b), often referred to as “yuppies” [27] in the Anglo-Saxon literature. Many other arrondissements have a similar profile, notably the 1st, 10th and 11th, but also the 4th, 5th, 9th, 14th and 18th.
a model in which the proportion of old people, aged 65 and above, is higher than the average. This is true above all for the 16th, to a lesser extent for the 6th and 8th, and also to a certain degree in the 12th, 15th and 17th arrondissements. This is the “oldies” model.
a model in which the proportion of young people, aged 0-19, is higher than the average. This model, much less common, applies only in the 19th arrondissement, to a lesser extent, in the 20th. The high proportion of young people in these districts reflects the strong presence of families with children, both couples with children and lone-parent families (Table 12). This is what we call here the “family model”.

Table 11

Breakdown by age group of the population in the three “model” arrondissements of Paris in 1999 (%)

Age group 2nd “yuppies” model 16th “oldies” model 19th “families” model Paris 0-19 14.8 18.7 23.9 18.3 20-39 44.1 28.5 32.8 36.0 40-64 29.4 31.9 30.9 30.3 65+ 11.7 20.9 12.4 15.4 Total 100.0 100.0 100.0 100.0 Source: INSEE population census.

Breakdown by age group of the population in the three “model” arrondissements of Paris in 1999 (%)

Figure 11

Age pyramids of the three “model” arrondissements of Paris in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

Table 12

Breakdown of households by type in the three “model” arrondissements of Paris in 1999 (%)

2nd “yuppies” 16th “oldies” 19th “families” Paris Persons living alone 64.3 53.4 48.6 56.2 – under 40 34.6 20.0 18.8 25.1 – age 40 to 64 19.4 15.6 17.8 17.2 – age 65+ 10.3 17.8 12.0 13.9 Childless couples 15.9 19.4 17.0 18.1 Families 19.2 26.6 33.1 25.0 – couples with children 14.0 20.2 22.8 17.6 – lone-parent families 5.2 6.4 10.3 7.4 Other households 0.6 0.6 1.3 0.7 Total 100.0 100.0 100.0 100.0 Source: INSEE population census.

Breakdown of households by type in the three “model” arrondissements of Paris in 1999 (%)

The pyramids of the two remaining arrondissements, the 8th and 13th, have fewer differences with respect to the overall pyramid, though there is a slightly higher proportion of adults aged 40-64.

50As shown by the population breakdown by household type (Table 12), most young adults in the 2nd arrondissement live alone. The old people in the 16th are persons living alone and couples with no children, two categories much more strongly represented in this model than in the other two. The high proportion of young people in the 19th can be linked to the higher proportion of families with children in this model, be it couples with children or lone-parent families.

51The geographical breakdown of these three models is easy to explain. Private sector dwellings in Paris, more expensive than those of the region, are beyond the means of many families with children, who live mainly in social housing (Table 13) often built in the 1970s in districts where space was still available, i.e. certain outer arrondissements, the 19th in particular. The 16th, and the other arrondissements close to this “oldies” model, typically comprise many old dwellings, larger than the average and generally owned by their occupants. Households therefore tend to stay put until the last occupant has died, which explains why the population is older than elsewhere. In the 2nd arrondissement and those with a similar profile, the central districts especially, the dwellings are smaller and most are in the private rental sector. They are better suited to young adults wishing to make the most of life in the capital.

Table 13

Characteristics of households and dwellings of the three “model” arrondissements of Paris. 1999 census

Characteristics 2nd “yuppies” model 16th “oldies” model 19th “families” model Paris Average household size 1.70 1.94 2.15 1.87 Average number of rooms of primary residences 2.17 2.96 2.57 2.53 Average number of rooms per person 1.28 1.53 1.20 1.35 Breakdown of households by tenure status (%) – Owners 28.3 38.7 23.9 29.6 – Social housing tenants 2.2 5.5 34.7 16.7 – Private sector and non rent-paying tenants 69.5 55.8 41.4 53.7 Total 100.0 100.0 100.0 100.0 Breakdown of primary residences by building age (%) – Before 1915 91.4 40.5 20.4 46.5 – 1915-1948 5.1 26.5 19.6 18.4 – 1949 or after 3.5 33.0 60.0 35.1 Total 100.0 100.0 100.0 100.0 Source: INSEE population census.

Characteristics of households and dwellings of the three “model” arrondissements of Paris. 1999 census

VI – Future population and housing trends in Paris

52Given the relationships between population and housing presented above, it is clear that future population trends in Paris will depend upon changes in the number of primary residences and their occupancy. This is why housing is taken into account in our projections (Dittgen, 2002b). These analyses were conducted by 5-year “leaps” over a period of 20 years, an interval compatible with the new local urban planning map.

53The projection method is the standard components method—which consists in ageing the population while taking mortality, fertility and migration into account—but with two important differences: changes in the size of households and in the number of primary residences are also taken into account. Moreover, the adjusted method is applied exclusively to the household population. The non-household population, comprising mostly institutional households [28] whose size depends above all on the vocations of the institutions concerned and the number of places available in them, is analysed separately [29].

54To take account of changes in the number of dwellings, the population of the additional dwellings that become available during each 5-year period are added at the end of each period. To determine this population, the number of dwellings in question is multiplied by an appropriate average household size and the result is then broken down using an appropriate age-sex structure [30].

55To take account of changes in household size, the total population resulting from the projection was compared at the end of each 5-year period with the population equal to the average household size multiplied by the number of dwellings. As household size is assumed to be decreasing, the first population is larger than the second. We then corrected the projected population numbers by age and sex so that the total number corresponds to that obtained with the projected household size. This correction was made exclusively on the sub-population of households comprising at least two people, since the population of one-person households—for a constant number of dwellings—cannot, by definition, decrease.

1 – Hypotheses that suggest a continuation of past trends

The decisive role of migration

56As Parisian mortality is little different from that of the country as a whole, we used French tables for the 21st century produced by Jacques Vallin and France Meslé (2001). According to these tables, male life expectancy should rise from 75.4 years in 2000 to 80.1 years in 2020, and that of women from 82.9 to 86.6 years. We limited ourselves to this one hypothesis, as is generally the case when making projections. This is justified by the considerable inertia of the phenomenon and, all the more so in this case, by the fact that natural change has much less impact in Paris than migration.

57For national projections, it is common to make several fertility assumptions—high, medium and low—since fertility, though contained within certain limits, is very unpredictable. In Paris, this is not necessary, since a variation in the birth rate would have little impact on population increase. Indeed, households often leave the capital when they have a child, generally to move into a larger dwelling in the suburbs. We saw above that one-third of babies born in a particular year had left Paris five years later. Hence, if fertility increased slightly, the number of departures would increase likewise, and if it fell slightly, so would the departures. In other words, whatever the assumed fertility level, within reasonable limits, the number of children present in Paris would vary little. This is why we used only the fertility level observed during the last intercensal period, i.e. 1.5 children per woman (see above).

58We could have taken several different migration assumptions however, since migration plays a more important role in Parisian population dynamics than natural change. But we have seen that migration trends are also subject to strong inertia. Moreover, this would further multiply the number of scenarios, already potentially large due to the variety of hypotheses concerning changes in housing stock and occupancy. So here again a single assumption was chosen, i.e., a prolongation of the migration trends observed in the last intercensal period [31].

On ongoing decrease in household size

59For the change in household size, we used three assumptions: a central assumption, with a continuation of recent trends (1982-1999) and hence a reduction in household size from 1.87 persons in 2000 to 1.79 in 2020. This corresponds to an increase in the proportion of one-person households from 52% to 58%. The high assumption, with a 50% smaller decrease, gives a value of 1.83 persons in 2020, while the low assumption, with a 50% larger decrease, results in 1.75 persons.

60How can the general assumption of a continuous reduction in the number of persons per household be justified? There are two arguments, one based on the analysis of figures by arrondissement, and the second based on analysis of the factors affecting changes in Parisian household size.

61In 1999, household size in the different arrondissements ranged from 1.66 persons in the 4th to 2.15 in the 19th, with a general average of 1.87. One might think that the low value in the 4th indicates a smaller dwelling size than elsewhere, though in fact, the average number of rooms per dwelling in this arrondissement (2.43) is not much different from that of the 19th (2.57). And more importantly, the value of 1.66 persons is not yet considered to be a floor value, since a further decrease was observed recently; in 1990, the average household size in the 4th was still 1.78 persons. So the average value used in the central assumption (1.79 persons) and even that of the low assumption (1.75) is still well above the lowest observed value.

62The substantial decrease in Parisian household size since the Second World War can be explained by the development of the rest of the agglomeration, which is better suited to large households, while Paris has “specialized” in smaller ones. As the Parisian agglomeration is still growing, the Parisian selection of small households is bound to increase, resulting in a smaller household size [32].

63For the number of additional dwellings, we chose a figure of 2,000 per year, slightly above the figures observed in recent years (see above), to take account of the current ambition of policy makers to stimulate new construction and convert offices into dwellings [33]. High and low assumptions are added: one with 3,000 more dwellings per year, and a second with just 1,000.

64The three assumptions for the number of additional dwellings combined with the three assumptions for household size give rise to nine scenarios. We chose just three: a central scenario based on the central assumptions for additional dwelling numbers and household size, and two extreme scenarios resulting from these hypotheses: a high scenario with 3,000 more additional primary residences per year and a reduction in average household size to 1.83 persons in 2020; a low scenario, with 1,000 additional primary residences per year and a reduction in average household size to 1.75 persons in 2020.

2 – Predictable results

A total population which varies little in the central scenario

65In the central scenario, the Parisian population continues to fall slowly, reaching 2.1 million by 2020 (Figure 12). The decrease in household size is partly offset by the larger number of primary residences. This naturally raises the question of how many primary residences would be needed to bring an immediate halt to population decline. A simulation based on the same assumptions of change in household size gives a number of 2,500 dwellings per year, not much higher than that of the scenario in question, which stands at 2,000.

Figure 12

Observed change in the population of Paris (in thousands) between 1954 and 1999 and projected change between 2000 and 2020

66Under the much more favourable assumption of the high scenario, the population would increase steadily to almost 2.2 million by 2020. In the low scenario, on the other hand, it would decline steadily to just above 2 million by 2020.

67Is the central scenario the most probable and are the extreme scenarios also probable? Close observation of the situation in Paris tends to suggest a negative answer to these two questions, for two reasons. It is increasingly difficult to increase the housing stock in Paris. Once the “gaps” in the urban landscape have been filled and the former industrial sites have been occupied, and if the high-rise option is ruled out, then further new housing construction will become impossible [34]. So, even if the 2,000 additional dwellings per year of the central scenario are feasible in the short term, they are more improbable over the long term. Consequently, the low scenario is just as likely as the central scenario. The high scenario, for its part, produced by simple symmetry, is practically ruled out.

A slowly evolving age structure

68Over the 20 years of the projection, the population age distribution does not change fundamentally, either in the central scenario or in the two others. This is because the migration pattern by age, a key factor for deter mining this structure, is assumed to remain constant. However, the proportion of elderly people, in decline since 1975, continues to fall until 2010 (Table 14 and Figure 13) before rising again, though to levels below those of the country. The proportion of very elderly, in decline since 1990, varies little over the two decades in question, by contrast with foreseeable trends for the country as a whole (Table 14 and Figure 14). We saw earlier that these decreases were due to more frequent departures at the moment of retirement. In other words, the change in the population of old people in Paris, like that of other age groups, is largely dependent upon migration. This said, the growing proportion of seniors from 2010 is due to the progressive entry of the baby-boom cohorts into this group. They do not enter the “very old” age group until 2026 [35], and the proportion of this age group will not increase until after this date.

Table 14

Projected change in the proportion (in %) of old people aged 65+ and 80+. France and Paris intra muros, 2000-2020 (central scenario)

Year 2000 2005 2010 2015 2020 65+ “elderly” France 16.0 16.6 17.0 19.0 21.0 Paris 15.4 14.6 14.6 15.5 16.2 80+ “very elderly” France 3.6 4.6 5.4 6.0 6.3 Paris 4.5 4.6 4.8 4.8 4.7 Sources: France, Brutel and Omalek (2003), central scenario; for Paris, author’s projections, central scenario.

Projected change in the proportion (in %) of old people aged 65+ and 80+. France and Paris intra muros, 2000-2020 (central scenario)

Figure 13

Observed change in the proportion of people aged 65+ in France and Paris intra muros from 1954 to 1999 and projected change from 2000 to 2020

Figure 14

Observed change in the proportion of people aged 80+ in France and Paris intra muros from 1954 to 1999 and projected change from 2000 to 2020

Conclusion

69The population of Paris does not evolve independently, since from geographical and demographic viewpoints, the capital of France is simply the centre of the Parisian agglomeration. Our focus on the centre is justified by the fact that Paris intra-muros is a distinct administrative entity and its councillors need to know how the population of their municipality is evolving. Moreover, the analysis of this population enables us to study the effects of housing and spatial constraints on populations living in enclosed territories.

70This dissociation between administrative reality, the city population and demographic and geographical realities does not exist in other major European capitals. In some cases, this is because the cities cover a territory which encompasses the entire agglomeration. Berlin, for example, covers more than 900 km² and Rome 1,500 km², while Paris covers only around 100 km². In other cases, the cities form part of larger entities covering the entire agglomeration and which constitute an equally important reference, such as London and Greater London, Brussels and Brussels-Capital for example. In such cases, unlike Paris, the inevitable population decrease in the city centre is not perceived negatively, since the population of the agglomeration as a whole is not limited in its expansion.

71The decline in the Parisian population is not the only consequence of its restricted territory. There are many others, though we will mention just three.

72Paris is unable to house all its inhabitants, and the poorest ones are the first to suffer. To house the twelve thousand extra Parisians each year resulting from natural increase (Table 1), more than seven thousand additional dwellings [36] would be needed, a figure that is impossible to achieve. So a large proportion of young households move out of the capital, to the immense regret of the Paris City Hall. And with an available housing stock of around 5,000 dwellings, the municipality is far from satisfying the demand for social housing in Paris (around 100,000 applications each year).

73Paris does not give a say to all “Parisians”. Jobs in the Parisian agglomeration are concentrated in the centre, which is not a bad thing given the star-shaped transport network. But as a consequence, more than half the people who work in Paris do not live there [37], and have no say in the running of the city. Yet certain issues, such as transport for example, may be of equal or even greater concern to suburban dwellers than to the Paris Inhabitants.

74Paris is increasingly a city for the rich. The capital attracts the agglomeration’s most affluent households and is becoming “gentrified” [38], to use a term applied widely in sociology. More specifically, the Parisian population is increasingly dichotomous, with a majority of rich people, at the expense of the middle classes, and a minority of poor people living in low-quality rented accommodation (garret rooms, furnished accommodation, hotel rooms, etc.). Clearly, this trend is linked to the appeal of the capital and the fact that housing prices are bound to rise because supply cannot be increased.

75If Paris were to spread to its agglomeration, the departure of young couples and of social housing applicants towards the suburbs would not be viewed as an exile for the persons concerned and as a policy failure for the City Hall. Most of the people who work in the centre would have a voice in the management of the city. True, this extension would not put an end to city-centre gentrification, but a wealthy district is probably more acceptable than a city for the wealthy.

76Contrary to the trends observed in other European countries, the municipal boundaries in France are unlikely to change for a long time to come. Perhaps the many forms of municipal community now being set up across the country will eventually carry more weight than the municipalities that they encompass, making it possible to bring administrative reality more into line with demographics and geography.

APPENDIX

77

The Paris region (Île de France), which corresponds almost exactly to the Paris agglomeration, comprises a central zone and two concentric rings:

Paris intra muros, the capital city, covering the central area of 87 km². It is divided into 20 arrondissements that spiral out from the centre (Municipality of Paris);
a densely urbanized inner suburban ring outside the administrative boundaries of Paris (inner suburbs);
an outer ring of distant suburbs with a more varied urban density (outer suburbs).

Notes

[*]
Institut de démographie, University of Paris I.
Translated from the French by Catriona Dutreuilh.
[1]
With net migration figures well above those of natural change, many European countries are an exception to this rule. This situation is relatively recent however and only concerns this part of the world.
[2]
In the autumn of 2004, the mayor of Paris launched the idea of building new residential tower blocks to address housing needs and halt depopulation, but his proposals met strong opposition.
[3]
The results presented here were obtained as part of a population projection study for the city of Paris commissioned by APUR (Atelier Parisien d’Urbanisme) prior to the adoption by the Paris city council of a new local urban planning map (Plan Local d’Urbanisme, PLU) to replace the former development area map (Plan d’Occupation des Sols, POS).
[4]
These models, INSEE’s Omphale model especially (Descours, 1992), are based on migration patterns observed in the past. Yet in a city where the number of dwellings is expected to change substantially, future migration patterns will be totally different from those of the past.
[5]
The Paris city limits were extended between the two World Wars up to the old fortifications (and to the Bois de Boulogne and Bois de Vincennes woodland areas), thereby increasing its population by around 40,000 people.
[6]
Another more secondary reason for the decrease is the fact that more conversions are performed to merge two existing dwellings into one than to divide one dwelling into two.
[7]
For example, the creation of social housing in the central districts by the incumbent city council has not modified the total number of dwellings in the capital since the dwellings already existed. Only their status has been changed.
[8]
The “occasional residence” category dates from the 1990 census. Previously, these dwellings, used for business purposes, were grouped with second residences. However, it is not easy for the enumerator in the field to distinguish between these two categories, so we have grouped them together.
[9]
The irregular variations in each of these two categories since 1982 is not necessarily a true reflection of reality, since here again, it is difficult to distinguish between the two in the field.
[10]
Dwelling size can be measured by floor area or number of rooms. We use the number of rooms for our study, since questions about floor area were not asked prior to the 1999 census, and the answers were imprecise (categories). The following are counted as rooms in the censuses: the living or dining room, the bedrooms and the kitchen if larger than 12m².
[11]
But people in the Paris suburbs have less space than French people in general: 1.38 rooms vs. 1.60.
[12]
The disappearance of live-in servants, which brought garret maids’ rooms onto the rental market, contributed to the increase in the number of dwellings and their decrease in size. However, this factor did not cancel out the general increase in dwelling size, except in the 16th arrondissement, where maid’s rooms were particularly common, and where the average number of rooms per dwelling fell from 3.16 in 1962 to 2.96 in 1999.
[13]
This population would be equal to the average number of rooms per dwelling in 1954 divided by the average number of rooms per person in 1999, multiplied by the number of households in 1999, the total of which is divided by the proportion of the total population represented by the household population: [(2.23 / 1.35) × 1,110,900] / 0.98.
[14]
The Paris agglomeration, which corresponds almost exactly to the Paris region (Île de France) totalled 11.2 million inhabitants in the 1999 census.
[15]
The sum of decreases in the two factors is not exactly equal to the population decrease due to their combined effect.
[16]
Average number of persons per household in 1999 multiplied by the number of households in 1954, the total of which is divided by the proportion of the total population represented by the household population: (1.87 × 1,185,300)/0.98.
[17]
Average number of persons per household in 1954 multiplied by the number of households in 1999, the total of which is divided by the proportion of the total population represented by the household population: (2.35 × 1,110,900)/0.98.
[18]
The numbers, taken from censuses conducted in March, are broken down by age reached during the year, i.e. by birth cohort. For this reason, the numbers given for age 0 and for the 0-4 age group are not complete.
[19]
The proportion of persons aged 65 and over in France increased very little between 1975 and 1982 and the proportion of persons aged 80 and over remained stable between 1990 and 1999 because these age groups correspond to the smaller birth cohorts of the First World War years.
[20]
A coefficient of 1.05, for example, means that the population grows by 5% per year due to migration, with arrivals outnumbering departures, while a coefficient of 0.98 means that it shrinks by 2% because departures outnumber arrivals.
[21]
In the 25-29 age group, the proportion of persons who were living in the suburbs at the time of the previous census is 15.2% in 1982, 16.8% in 1990 and 19.2% in 1999.
[22]
In 1999, there were around 1.6 million jobs in Paris for a working age population of one million (Davezies, 2001).
[23]
Non rent-paying tenants may be housed free of charge by their parents or friends or by their employer, in which case the rent-free accommodation may be counted as a fringe benefit.
[24]
For the proportions of households housed by their employer and by family or friends —64% and 36% respectively—the most recent information dates back to the 1982 census. As most of the people housed by their employer are building caretakers (concierges), whose numbers are decreasing, the proportions have probably become much more equal in recent years.
[25]
Many of these plots were former industrial sites.
[26]
For all models, the 20-39 age group is the largest in all arrondissements if the population is broken down into age groups of equal span.
[27]
Young urban professional.
[28]
Institutional households, classified in decreasing order by their number of residents (in 1999) are: workers’ hostels, student hostels, retirement homes, religious communities, long-stay hospitals and accommodation centres. The non-household population also includes: prison inmates, boarding school students and soldiers living in barracks who have no other residence, bargemen, travelling populations and persons with no fixed address.
[29]
Since 1982, the populations of certain insitutional households have decreased, while others have increased. Their overall population has remained constant to within slightly less than 50,000. For the projection, we prolonged the trends observed in the various institutional households while maintaining the overall population constant. We also maintained a constant age and sex structure for these populations, except in the case of retirement homes, where we aged the population, since observation shows that people tend to enter and “leave” retirement homes at increasingly advanced ages.
[30]
Household size and structure were based on the structures of households living in dwellings built during the most recent intercensal period.
[31]
The migration coefficients presented in Figure 6 concern the entire population. Those used in the projection concern the household population only. They take account of exchanges between Paris and the suburbs or provinces, but also of exchanges with the population of institutional households.
[32]
At national and regional level, the future change in household size depends on the factors mentioned above, mainly population ageing and an increase in the number of one-person households at the expense of cohabiting couples. At these levels, household size can be projected on the basis of assumed trends relating to these factors (Deckneudt et al., 2003; Sagot, 1997). For the Parisian population, this projection would need to take account firstly of the appeal of Paris for small households and its impact on their size, a factor which is very difficult to quantify. Hence the decision to opt for the continuation of past trends.
[33]
These same policy makers are also seeking to reduce the vacancy rate, so we have maintained the number of dwellings not occupied by Parisians at a stable level, despite the foreseeable increase in the number of second and occasional residences.
[34]
Moreover, the few plots still available will not all be used for housing, since space is also needed for new amenities and business premises.
[35]
The first large post-war cohort was born in 1946.
[36]
Taking account of the average number of persons per household and the proportion of primary residences in the total housing stock.
[37]
Out of a total of 1.6 million jobs in 1999, only 700,000 were occupied by Parisians.
[38]
On this question, see Pinçon and Pinçon-Charlot, 2004.

English

Abstract

Territorial population change is the result of natural change and migration, with the first primarily affecting national populations, and the second local populations. At local level, migration patterns depend largely upon housing or, more specifically, on changes in the number of primary residences and in their occupancy.
Hence, the depopulation of Paris “intra muros” since the Second World War can be explained by a decline in the number of primary residences—despite an increase in the number of dwellings—and, above all, a decrease in household size which, like elsewhere in France and the Western world, reflects the “couple crisis” and demographic ageing, but equally the appeal of Paris for singles and small households and the “banishment” of families to the suburbs.
This pattern of urban residence can be explained by several factors, above all by the size of dwellings – small in Paris and larger in the suburbs – and their higher cost inside Paris than beyond the city limits. These factors are raising the proportion of young adults and of professionals and higher-level occupations in the Parisian population. This trend is especially pronounced in the central districts, where the small dwellings and animated social life are particularly attractive to the singles population.

Français

Résumé

L’évolution de toute population territoriale résulte du mouvement naturel et du mouvement migratoire. Mais alors que le premier est prépondérant pour les populations nationales, le second l’est pour les populations locales. À ce niveau, les migrations sont fortement dépendantes du logement, soit plus précisément de l’évolution du nombre de résidences principales et de celle de l’occupation de ces résidences.
Ainsi, le dépeuplement de Paris intra-muros depuis la guerre s’explique par la baisse du nombre de ces résidences – en dépit de l’augmentation du nombre de logements – et surtout par la diminution de la taille des ménages. Celle-ci résulte, comme ailleurs en France, voire dans le monde occidental, de la « crise » du couple et du vieillissement démographique, mais tout autant de l’attirance exercée par Paris sur les personnes seules et les petits ménages de l’agglomération et du « refoulement » des familles en banlieue.
Ce partage de l’espace de l’agglomération s’explique par plusieurs facteurs, au premier rang desquels la taille des logements, petite à Paris, plus grande en banlieue, et leur coût, élevé à Paris, meilleur marché en banlieue. Il en résulte que les jeunes adultes et les cadres sont de plus en plus nombreux dans la population parisienne. Cette tendance est surtout marquée dans les quartiers centraux, dont les petits logements et l’animation sont adaptés aux personnes seules.

Español

Resumen

La evolución de la población es el resultado del crecimiento natural y del crecimiento migratorio. Así como la importancia del primero es preponderante a nivel nacional, el segundo predomina a nivel local. A tal nivel, las migraciones dependen fuertemente de la vivienda y, más concretamente, de la evolución del número de residencias principales y de la ocupación de tales residencias.
La disminución de la población de París intramuros después de la guerra, por ejemplo, se debe a la disminución de la oferta de tales residencias –a pesar del aumento en el número de viviendas- y especialmente a la disminución del tamaño de los hogares. Tal disminución es el resultado, tanto en París como en el resto de Francia y del mundo occidental, de la “crisis” de la pareja y del envejecimiento demográfico, pero también de la atracción que ejerce París sobre las personas solas y de los hogares pequeños de la aglomeración y la consiguiente “expulsión” de familias hacia los suburbios.
Tal división del espacio de la aglomeración se debe a varios factores: el factor principal es el tamaño de la vivienda, pequeño en París y mayor en los suburbios, pero también hay razones de coste, más elevado en París que en los suburbios. Estos dos factores explican que en la población predominen los adultos jóvenes y que los cuadros sean cada vez más numerosos. Esta tendencia es más marcada en los barrios centrales, ya que las viviendas pequeñas y la animación atraen a las personas solas.

I - Population, dwellings and primary residences
1. 1 - A shrinking population, an expanding housing stock
2. 2 - Fewer dwellings occupied by Parisians
II - Population and household size
III - Age-sex structure, migration and housing
IV - Households by tenure status and age of dwelling
V - Populations and dwellings in the arrondissements
1. 1 - Parallel trends
2. 2 - Three population models, and one dominant model
VI - Future population and housing trends in Paris
Conclusion

REFERENCES

Alders M., Manting D., 2001, “Household scenarios for the European Union, 1995-2025”, Genus, LVII, n° 2, pp. 17-47.
Atelier parisien d’urbanisme (APUR), 1994, Paris 1954-1990 : données statistiques. Population, logement, emploi. Paris et arrondissements, 90 p.
Atelier parisien d’urbanisme (APUR), 2000, Le recensement de 1999 à Paris : les résultats du dénombrement. Population et logement, 38 p.
Beckouche P., 2004, “Les relations Paris-banlieue à l’ère des régions urbaines”, in Y. Michaud (dir.), Université de tous les savoirs, Odile Jacob.
OnlineBonvalet C., Lefèbvre M., 1983, “Le dépeuplement de Paris 1968-1975. Quelques éléments d’explication”, Population, 38(6), pp. 941-958.
OnlineBonvalet C., Lelièvre É., 1991, “Mobilité en France et à Paris depuis 1945. Le filtre parisien”, Population, 46(5), pp. 1161-1183.
OnlineBonvalet C., Tugault Y., 1984, “Les racines du dépeuplement de Paris”, Population, 39(3), pp. 463-481.
Brutel C., Omalek L., 2003, “Projections démographiques pour la France, ses régions et ses départements (horizon 2030/2050)” (with CD-ROM), Insee Résultats (série Société).
Davezies L., 2001, L’économie et l’emploi à Paris, Éléments de diagnostic, APUR/L’ŒIL.
Deckneudt J., Jacquot A., Macrakis B., 2003, “Projection de ménages pour la France métropolitaine, ses régions et ses départements (Horizon 2030)”, Insee Résultats (série Société).
Descours L., Poinat F., 1992, “Modèle de projection démographique OMPHALE,”, Insee Méthodes.
Dittgen A., 2001, Analyse de la dynamique de la population parisienne et éléments de perspectives, CRIDUP/APUR.
Dittgen A., 2002a, Perspectives de la population de Paris, 2000 -2020, APUR/CRIDUP, 53 p.
Dittgen A., 2002b, La population des ménages parisiens selon le statut d’occupation et l’ancienneté du logement, APUR/CRIDUP, 63 p.
Dubois-Taine G. (ed.), 2004, European Cities Insights on Outskirts from Helsinki to Nicosia. Eleven Case Studies and Synthesis, Ministry of Infrastructure, France.
Esponda Marc, 2002, “Les jeunes adultes toujours aussi nombreux à Paris”, in Atlas des Franciliens, volume 3, Population et modes de vie, INSEE-IAURIF.
Farcy J.-C., Faure A., 2003, La mobilité d’une génération de Français. Recherche sur les migrations et les déménagements vers et dans Paris à la fin du xix^e siècle, Paris, INED (coll. Les cahiers de l’Ined, n° 151) 616 p. + CD-ROM.
OnlineGiblin D., Papin D., Subra P., 2001, “Paris/Londres: enjeux géopolitiques de villes capitales”, Hérodote, n° 101, 2^nd quarter.
Gollain V., 2003, Berlin capitale de la Mitteleuropa?, Île-de-France, Agence régionale de développement.
Hassan M.-E., 2001, “99% des Franciliens vivent dans l’aire urbaine de Paris”, Île-de-France à la page, n° 196, INSEE.
Insee, Recensement de la population de 1999, Évolutions démographiques 1982-1990-1999, Données définitives, Paris, s.d.
Insee, Recensement de la population de 1999, Tableaux références et analyses, Exploitation principale, Paris, s.d.
Insee-Iaurif, Atlas des Franciliens, volume 1, 2000, Territoire et population; volume 2, 2001, Logement; volume 3, 2002, Population et modes de vie; volume 4, 2003, Activité et emploi.
Jean-Marie A., 2004, “Paris ville de riches? Les évolutions de la sociologie parisienne”, in Y. Michaud (dir.), Université de tous les savoirs, Odile Jacob.
Jean-Marie A., 1987, Les densités résidentielles à Paris et dans quelques villes étrangères, APUR.
Laferrère A., Warsée C., 2002, “Secteur libre et HLM, deux dynamiques différentes”, Insee Première, n° 839.
Laroche N., 2001, “L’Île-de-France pourrait dépasser 12 millions d’habitants en 2030”, Île-de-France à la page, n° 201, INSEE.
Louchart P., 1997, L’influence du parc de logements sur le structure et la dynamique par âge des populations à l’échelon local : éléments d’analyse, IAURIF.
Mairie de Paris, 2001, Bilan PLU. Bilan des données la construction à Paris, DAUC/Bureau de l’information et de la construction.
Mairie de Paris/Iaurif, 2004, L’accès au logement social à Paris.
Merlin P., 2003, “L’Île-de-France hier, aujourd’hui et demain”, Notes et études documentaires, n° 5181.
Navarre D., Blumenfeld H., Jacob C., Manz K., 2002, Paris-Berlin. Comparaison des systèmes de transport, IAURIF.
Omalek L., 2001, “Projections régionales de population pour 2030 : l’impact des migrations”, Insee Première, n° 805.
OnlinePinçon M., Pinçon-Charlot M., 2004, Sociologie de Paris, La Découverte (coll. Repères, n° 400).
Sagot M., 1997, Perspectives de ménages et de logements en Île-de-France à l’horizon 2015, IAURIF.
Simon P., 2000, “Paris. La division sociale et ethnique de l’espace parisien”, in P. Simon, Métropoles en mouvement : une comparaison internationale, Anthropos-IRD, pp. 299-309.
Vallin J., Meslé F., 2001, “Tables de mortalité françaises pour les xix^e et xx^e siècles et projections pour le xxi^e siècle”, INED (coll. Données statistiques, n° 4).

Alfred Dittgen [*]

Alfred Dittgen, Institut de Demographie de l’Université Paris I, email: alfred.dittgen@univparis1.fr

[*]
Institut de démographie, University of Paris I.
Translated from the French by Catriona Dutreuilh.

Translated from the French by

Catriona Dutreuilh

Latest publication on cairn or another partner portal

Religions and Demography in France
Contrasting Developments
In Population & Avenir Volume 684, Issue 4, September 2007

Latest publication on cairn or another partner portal

Recent Demographic Trends in France
Ongoing Impact of the COVID-19 Pandemic at Both Regional and National Levels
Avec Didier Breton, Nicolas Belliot, Magali Barbieri, Hippolyte d’Albis, Magali Mazuy, Catriona Dutreuilh, Paul Reeve
In Population Volume 77, Issue 4, October 2022

Uploaded on Cairn-int.info on 03/03/2014

Cite article French English

ISO 690		Copy DITTGEN Alfred, « Logements et taille des ménages dans la dynamique des populations locales. L'exemple de Paris », Population, 2005/3 (Vol. 60), p. 307-347. DOI : 10.3917/popu.503.0307. URL : https://preprod.cairn.info/revue-population-2005-3-page-307.htm
ISO 690		Copy DITTGEN Alfred, "Housing and Household Size in Local Population Dynamics. The example of Paris", Population, 2005/3 (Vol. 60), p. 259-298. URL: https://preprod.cairn-int.info/journal-population-2005-3-page-259.htm
MLA		Copy Dittgen, Alfred. « Logements et taille des ménages dans la dynamique des populations locales. L'exemple de Paris », Population, vol. 60, no. 3, 2005, pp. 307-347.
MLA		Copy Dittgen, Alfred. "Housing and Household Size in Local Population Dynamics. The example of Paris", Population, vol. 60, no. 3, 2005, pp. 259-298.
APA		Copy Dittgen, A. (2005). Logements et taille des ménages dans la dynamique des populations locales: L'exemple de Paris. Population, 60, 307-347. https://doi.org/10.3917/popu.503.0307
APA		Copy Dittgen, A. (2005). Housing and Household Size in Local Population Dynamics: The example of Paris. Population, 60, 259-298. https://preprod.cairn-int.info/journal--2005-3-page-259.htm.
DOI		Copy https://doi.org/10.3917/popu.503.0307

Download citation

file_downloadZotero (.ris) file_downloadEndNote (.enw) open_in_new RefWorks

Cite

Distribution électronique Cairn.info pour I.N.E.D © I.N.E.D. Tous droits réservés pour tous pays. Il est interdit, sauf accord préalable et écrit de l’éditeur, de reproduire (notamment par photocopie) partiellement ou totalement le présent article, de le stocker dans une banque de données ou de le communiquer au public sous quelque forme et de quelque manière que ce soit.

Full-text

Paris intra muros. Mean annual arrivals and departures over the period 1990–1999 (rounded numbers)

Paris intra muros. Mean annual arrivals and departures over the period 1990–1999 (rounded numbers)

I – Population, dwellings and primary residences

1 – A shrinking population, an expanding housing stock

Population change (total, household, non-household) between 1954 and 1999 in Paris intra muros and intercensal net change in total population (thousands)

Population change (total, household, non-household) between 1954 and 1999 in Paris intra muros and intercensal net change in total population (thousands)

Population change in Paris intra muros between 1954 and 1999 (thousands)

Population change in Paris intra muros between 1954 and 1999 (thousands)

2 – Fewer dwellings occupied by Parisians

Change in the number of dwellings by category in Paris intra muros between the 1954 and 1999 censuses and intercensal net change (thousands)

Change in the number of dwellings by category in Paris intra muros between the 1954 and 1999 censuses and intercensal net change (thousands)

Change in the number of second and occasional residences and of vacant dwellings in Paris intra muros between 1954 and 1999 (thousands)

Change in the number of second and occasional residences and of vacant dwellings in Paris intra muros between 1954 and 1999 (thousands)

II – Population and household size

1 – Small households and small dwellings

Total number of households in Paris intra muros, Average size and breakdown by size. Change between 1954 and 1999 (in thousands and %)

Total number of households in Paris intra muros, Average size and breakdown by size. Change between 1954 and 1999 (in thousands and %)

Household size and dwelling size in France, Paris region and Paris intra muros in 1999

Household size and dwelling size in France, Paris region and Paris intra muros in 1999

2 – Paris attracts small households

Paris and France. Change in average household size between 1954 and 1999

Paris and France. Change in average household size between 1954 and 1999

3 – Smaller household size, the main cause of depopulation

intercensal rate of change in total population, household population, number of primary residences and household size in Paris intra muros (%)

intercensal rate of change in total population, household population, number of primary residences and household size in Paris intra muros (%)

III – Age-sex structure, migration and housing

1 – Young adults and women are over-represented

Age-sex distribution of the population of France, the Paris region and Paris intra muros by five-year age group for a total of 1,000 inhabitants. 1999 census

Age-sex distribution of the population of France, the Paris region and Paris intra muros by five-year age group for a total of 1,000 inhabitants. 1999 census

Age-sex distribution of the population of Paris intra muros by five-year age group at different census dates for a total of 1,000 inhabitants

Age-sex distribution of the population of Paris intra muros by five-year age group at different census dates for a total of 1,000 inhabitants

Proportions of persons aged 65+ and persons aged 80+ between 1954 and 1999 in France and in Paris intra muros (%)

Proportions of persons aged 65+ and persons aged 80+ between 1954 and 1999 in France and in Paris intra muros (%)

2 – An age-sex structure determined by migration

Coefficients of age-specific population change in Paris intra muros due to migration. Mean annual values 1990-1999

Coefficients of age-specific population change in Paris intra muros due to migration. Mean annual values 1990-1999

3 – Migration patterns closely linked to dwelling characteristics

IV – Households by tenure status and age of dwelling

1 – Owners enjoy more space than tenants

Household size and dwelling size by type of tenure in Paris intra muros in 1999

Household size and dwelling size by type of tenure in Paris intra muros in 1999

2 – A Parisian pyramid whose shape reflects that of private sector tenants

Age-sex distribution of the household population of Paris intra muros by tenure status in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

Age-sex distribution of the household population of Paris intra muros by tenure status in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

3 – Owner households and social housing tenants stay put for life

Age-sex distribution of the owner household population of Paris intra muros by building age in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

Age-sex distribution of the owner household population of Paris intra muros by building age in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

Average age of household members by type of tenure and building age in Paris intra muros in 1999

Average age of household members by type of tenure and building age in Paris intra muros in 1999

4 – High mobility among private sector tenants

Age-sex distribution of the population of private sector and non rent-paying tenant households in Paris intra muros by building age in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

Age-sex distribution of the population of private sector and non rent-paying tenant households in Paris intra muros by building age in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

V – Populations and dwellings in the arrondissements

1 – Parallel trends

Changes in population size and number of dwellings by arrondissement of Paris between 1954 and 1999 (%)

Changes in population size and number of dwellings by arrondissement of Paris between 1954 and 1999 (%)

Comparative change in the household population and in the number of primary residences in the arrondissements of Paris between 1954 and 1999

Comparative change in the household population and in the number of primary residences in the arrondissements of Paris between 1954 and 1999

2 – Three population models, and one dominant model

Breakdown by age group of the population in the three “model” arrondissements of Paris in 1999 (%)

Breakdown by age group of the population in the three “model” arrondissements of Paris in 1999 (%)

Age pyramids of the three “model” arrondissements of Paris in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

Age pyramids of the three “model” arrondissements of Paris in 1999 (distribution by five-year age group for a total of 1,000 inhabitants)

Breakdown of households by type in the three “model” arrondissements of Paris in 1999 (%)

Breakdown of households by type in the three “model” arrondissements of Paris in 1999 (%)

Characteristics of households and dwellings of the three “model” arrondissements of Paris. 1999 census

Characteristics of households and dwellings of the three “model” arrondissements of Paris. 1999 census

VI – Future population and housing trends in Paris

1 – Hypotheses that suggest a continuation of past trends

The decisive role of migration

On ongoing decrease in household size

2 – Predictable results

A total population which varies little in the central scenario

Observed change in the population of Paris (in thousands) between 1954 and 1999 and projected change between 2000 and 2020

Observed change in the population of Paris (in thousands) between 1954 and 1999 and projected change between 2000 and 2020

A slowly evolving age structure

Projected change in the proportion (in %) of old people aged 65+ and 80+. France and Paris intra muros, 2000-2020 (central scenario)

Projected change in the proportion (in %) of old people aged 65+ and 80+. France and Paris intra muros, 2000-2020 (central scenario)

Observed change in the proportion of people aged 65+ in France and Paris intra muros from 1954 to 1999 and projected change from 2000 to 2020