| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Age Concern Institute of Gerontology (ACIOG), King's College London
Correspondence: Karen Glaser, ACIOG, King's College London, Waterloo Bridge Wing, Franklin-Wilkins Building, Waterloo Road, London SE1 8WA, UK. E-mail: karen.glaser{at}kcl.ac.uk.
Vernon L. Greene, PhD
 |
Abstract |
|---|
Key Words: Coresidence • Distance • Elderly women • Intergenerational relations
There has been a substantial increase in the proportions of elderly people living alone in the post-World War II period. This change is apparent throughout the European Union, even in Southern European countries, which have historically been characterized by high rates of coresidence (
Coleman 1997
;
Pampel 1992). This trend has led to increasing concerns that family ties are growing weaker and that older people are less supported by their families. Declines in coresidence with kin among older people may, however, reflect preferences for living near, but not with children, an attitude described as "...intimacy at a distance" (
Rosenmayr and Kockeis 1963). Studies have shown that norms of mutual responsibility between parents and children can often be met through proximity and do not necessarily require coresidence (
Mancini and Bleiszner 1989). Thus, it has become increasingly important to analyze the spatial proximity of elderly people and their kin, in addition to intergenerational coresidence, in order to provide a more comprehensive view of the availability of family support to older people.
Research on the spatial proximity of elderly people and their kin is largely American (
Clark and Wolf 1992;
Clark, Wolf, and Schulte 1994;
Lin and Rogerson 1995), although there have been some European studies (
Warnes 1986). A few descriptive studies have examined intercultural variations in spatial closeness between older people and their kin (
Finch 1989; Hölinger & Haller, 1990). These studies have shown that individuals from Northern European countries (e.g., Britain) are much less likely to live close to their mothers than those from Southern European countries such as Italy (
Finch 1989;
Hollinger and Haller 1990). Differences in proximity were thought to reflect the strength of cultural attitudes in maintaining close kin interactions, rather than socioeconomic factors such as the availability of housing and jobs (
Finch 1989).
This study examined similarities and differences in the effects of demographic and socioeconomic determinants on the proximity of older women (aged 60–74) to their children in Britain and Italy using data from the British Retirement Survey follow-up in 1994, and the 1995 Italian Indagine Multiscopo sulle Famiglie. Demographic and socioeconomic factors, which may be responsible for the higher levels of parent–child proximity observed in Italy, are discussed in this article. Differences in residential proximity are not expected to reflect variations in the availability of children, as women in the cohorts under comparison (those born between 1920 and 1935) experienced similar fertility trajectories. It is suggested that a combination of factors, involving differences in housing and education policies, job opportunities, and especially cultural attitudes with respect to the strength of family ties, provide a better explanation of variations in residential propinquity between the two countries. As previous research demonstrated important differences in cultural values between Britain and Italy in terms of the desirability of intergenerational interactions and contact (
Glaser, Hancock, and Stuchbury 1998;
Walker and Maltby 1997), it is hypothesized that parent–child proximity in Britain is more likely to arise from the needs of the older generation, whereas in Italy it is more likely to reflect a cultural preference regardless of parental need. In addition, research has shown that children's characteristics (in terms of the number of their own young children, their health, or economic circumstances) are also important for determining intergenerational proximity (
Clark and Wolf 1992;
Lin and Rogerson 1995). In Italy, delays in the transition to adulthood (e.g., higher ages at school leaving, labor force entry, and marriage) have been shown to affect intergenerational coresidence (
Billari and Ongaro 1998;
Iacovou 1998;
Sabbadini 1999). Thus, this study also investigates the association between adult children's (aged 25–39) demographic and socioeconomic characteristics and proximity to the mother (for those with a living mother) using data from the 1995 British Social Attitudes Survey and the 1995 Italian Indagine Multiscopo sulle Famiglie.
This research used multinomial logit models to examine differences and similarities in the correlates of proximity to children among these older women. Our findings showed that certain parental characteristics such as health have a significant effect on proximity in Britain but not in Italy, providing support for the hypothesis that parental needs may have a greater impact in determining parent–child proximity in societies with relatively low levels of desired kin interaction. Our analysis of the relationship between adult children's demographic and socioeconomic characteristics with proximity to mothers showed less of an association in Britain than in Italy. The importance of adult children's needs for intergenerational proximity in Italy may explain, in part, why certain parental characteristics were not as strong predictors of propinquity. Consequently, differences between Britain and Italy in the spatial proximity of older women and their children may be a result of several factors such as the effects of parental need, cultural factors emphasizing close family ties, and the socioeconomic circumstances of young adults.
|
Trends in Living Arrangements and Their Implications |
|---|
), although recent evidence for Italy suggests a decline in the proportion living alone and an increase in the proportion living with children (
Sabbadini 1999;
Tomassini and Wolf in press). However, the proportion of older women living alone has continued to remain higher in Britain than in Italy (Table 1 ). British General Household Survey (GHS) data from 1996 show that 48% of women aged 65 and older lived alone compared to 35% in Italy. The decrease in intergenerational coresidence among older people has important implications for the exchange of help and support (
Sundstrom 1994). Although most older people remain healthy and are able to look after themselves, the risk of disability rises with age, increasing the vulnerability of those living alone. If the needs of frail older people cannot be met from within the household, it is likely that these needs will be met by kin living in close proximity. For this reason the issue of residential proximity of elderly people to their kin has become increasingly important (
Wolf 1994). Furthermore, as nearby kin may be able to provide similar support to coresident kin (
Mancini and Bleiszner 1989), declines in coresidence may, in part, reflect preferences among older people for living close to, but not with, their children, implying little or no loss of potential support.
|
|
Proximity |
|---|
). There are two possible avenues to intergenerational proximity: (a) adult children's economic needs are an important influence on geographic distance earlier in the parental life course, and (b) parental economic and health needs are likely to influence geographic proximity to adult children at older ages (
Greenwell and Bengtson 1997). In particular, declining health among older people is believed to be an important trigger for coresidence or proximity to an adult child, reflecting the need for a greater level of support than that available within the elderly person's own household (
Al-Hamad, Flowerdew, and Hayes 1997;
Ford 1993;
Silverstein 1995;
Silverstein and Angelelli 1998). Previous research has also found that the demographic and socioeconomic characteristics of older people, such as age, marital status, number of children, education, and housing tenure (whether an individual is an owner–occupier, that is, someone who owns his/her home, rents from a private landlord, or is in social sector housing as either a local authority or housing association tenant) are important predictors of spatial proximity (
Clark and Wolf 1992;
Lin and Rogerson 1995). Age showed a curvilinear effect on distance between older parents and their children, with the odds that a respondent lived near a child declining among the "young elderly" and increasing thereafter (
Clark and Wolf 1992). Widows were more likely than divorced or separated women to live close to a child (
Clark and Wolf 1992;
Lin and Rogerson 1995). The number of children had a significant positive effect on the likelihood that the older respondent lived near a child (
Clark and Wolf 1992;
Lin and Rogerson 1995). The more education an older parent had, the less likely they were to live near a child (
Clark and Wolf 1992;
Lin and Rogerson 1995). In Britain, housing tenure was found to be an important determinant of proximity between older parents and their adult children. Those in social sector housing were found to be living closer to their children than owner–occupiers (
Warnes and Ford 1995b). |
Fertility Patterns and Their Implications |
|---|
|
), differences in proximity may reflect differences in mean age at childbirth. However, the mean ages at childbirth in 1920 / 21 were similar in England and Wales and Italy (28.6 and 29.1, respectively;
Santini 1974). |
Cultural Values and Their Implications |
|---|
). In addition, numerous studies using both the 1992 Eurobarometer data (
Glaser et al. 1998;
Walker 1993;
Walker and Maltby 1997), and data from the 1986 International Social Survey Programme (ISSP) on social networks (
Finch 1989;
Hollinger and Haller 1990) have shown that the actual frequency of family contact among older people is less in Britain than in Italy, even when distance to mother is controlled for. The 1986 ISSP data showed that 32% of Italians who lived close to their mother saw her every day compared with 11% of British respondents (
Finch 1989). British individuals were also less likely to be caring for a dependent than their Italian counterparts. For example, 10% of women aged 50–64 were providing coresident care for an ill, handicapped, or elderly household member in Britain compared with 17% in Italy (
Glaser et al. 1998). Given empirical evidence of lower levels of both desired and actual kin interaction in Britain compared with Italy, it is likely that differences in cultural values play an important role in determining the lower levels of parent–child proximity in Britain. As cultural norms reflecting the desirability of intergenerational contact are weaker in Britain than in Italy, proximity in Britain may be more likely to reflect other factors, such as the need for assistance among those with an illness or disability. |
Educational Levels and Their Implications |
|---|
). In Italy, these cohorts of older women would also have experienced a minimum school-leaving age of 14 (introduced by law in 1923), although this was largely disregarded until the mid 1970s (
D'Alessandro 1996). In general, the educational attainment of the adult population is higher in Northern Europe than in Southern Europe. For example, while in the United Kingdom 52% of women aged 55–64 had attained at or above an upper secondary education (derived from typical graduation ages in each country, 16–18 in the United Kingdom and 17–19 in Italy), the figure for Italy was 13% (Organization for Economic Co-Operation and Development, 1998). The higher educational levels of older adults in Britain compared with Italy may account, in part, for the lower levels of intergenerational proximity in Britain. |
Housing Tenure and Its Implications |
|---|
). A higher proportion of the older birth cohorts are in local authority (social sector) housing, whereas the younger cohorts are more likely to be owner–occupiers (
Evandrou and Falkingham 2000;
Warnes and Ford 1995a). Few individuals are in privately rented accommodations, reflecting declines in the availability of this type of housing. In contrast, there is little social sector housing in Italy, and the majority of individuals are owner–occupiers or private renters.
Studies have shown that in Britain, social sector housing tenants are more likely to live close to their children than owner–occupiers (
Warnes and Ford 1995b). This may be because they have less choice about the timing and destination of moves and are more likely to move locally rather than long distances (
Warnes and Ford 1995a). In addition, tenure in Britain is strongly associated with social class and income, and is often used as a measure of social status (
Fox and Goldblatt 1982). Studies have found pronounced social class differences in mobility, so that parents at the top of the class structure live further from their children than those at the bottom of the structure (
Warnes 1986). With regards to coresidence between older parents and their adult children, socioeconomic disadvantage, as measured by housing tenure, appears to be associated with a decreased likelihood of coresidence (
Grundy 1993).
In Italy, the inheritance of property appears to play an important role in the acquisition of housing among adult children (
Barbagli 1997). In addition, parents appear to provide considerable financial contributions to adult children toward the purchase of their own homes (
Barbagli 1997). Recent surveys in Italy have shown that 15% of all couples were given a house by their parents or parents-in-law, and 12% received substantial help in purchasing their homes (
Sabbadini 1999). This kind of assistance may provide Italian parents with a greater say in where adult children live, and may be one reason why a high proportion of adult children live close to or in the same building as their parents (
Barbagli 1997;
Cioni 1997).
|
Methods |
|---|
). Data were collected on the current financial circumstances of respondents, pension and job histories, marriage and fertility histories, health, and distance and frequency of contact with both parents and children.
This research was based on female key respondents at the second wave of the survey (aged 60–74 in 1994), resulting in a weighted sample size of 992 ever-married women with at least one child.
The British Social Attitude Surveys (BSAS) are annual cross-sectional surveys of the population aged 18 and older that have been carried out every year since 1983 with the exception of 1988 and 1992 (
Lilley, Brook, Park, and Thomson 1997). They were designed to monitor trends in attitudes and to understand attitudinal change. The 1995 survey contained a module on family networks, which asked questions on proximity between adult children and their parents as well as questions concerning frequency of contact with kin. The 1995 BSAS had a total sample size of 3,633 individuals aged 18 and older. Only a portion of the sample (2,067) were given the module on family networks as a self-completion questionnaire (
Jowell, Curtice, Park, Brook, and Thomson 1996), resulting in a weighted sample size of 547 adults aged 25–39 with a living mother.
Italy
The analysis used data from the 1995 Indagine Multiscopo sulle Famiglie, a survey carried out every year on the private household population of Italy. The unit of the sample is the "de facto" household selected from the Register of Population. The survey covered a wide variety of topics including questions on household structure, demographic background, housing, area of residence, health status, and leisure time. The survey has a total sample size of 60,890 respondents from which two subsamples were selected: ever-married women aged 60–74 with at least one child (a weighted sample size of 3,940 respondents), and adults aged 25–39 with a living mother (a weighted sample of 13,015 respondents).
Method
Dependent Variable
The dependent variable, the older woman's proximity to her nearest child, was composed of three categories: coresidence, lived close (up to 10 miles), and lived far (more than 10 miles). This variable was based on a question concerning distance to children available in both surveys. In Britain, respondents who had children were asked, "How far away does (the nearest/he/she) live?" In Italy respondents were asked how far away each of their children lived (this question was asked for up to three children; respondents with more than three children were asked about the three closest ones). For the Italian data, an algorithm was created to identify the child living closest to his/her mother. The responses were grouped into the following categories: (1) women coresiding with their nearest child (living in the same household); (2) those living close to their nearest child (up to 10 miles in Britain, up to 16 km in Italy); and (3) those living far from their nearest child (more than 10 miles in Britain, more than 16 km in Italy).
Independent Variables
Two dummy variables for age were used in the model (65–69 and 70–74 with the youngest age group, 60–64, as the reference category). Other demographic variables such as marital status and number of children were also included, given the positive association between widowhood and number of children to spatial proximity of an adult child (
Clark and Wolf 1992;
Lin and Rogerson 1995).
Health was included in the models, as "kinship migration" among older people is thought to be triggered by deteriorating health. Both the British and Italian surveys contained questions on self-perceived health. The British health question referred to the respondent's health in the past 12 months (Over the past 12 months would you say that, on the whole, your health has been ...?), whereas the Italian survey asked about the respondent's health in general (Come va in generale la sua salute?). The response categories for the British health question were as follows: (1) good, (2) fairly good, and (3) not good. Respondents in Italy were asked to rate their health on a scale from 1 to 5, with 5 representing those in the best health and 1 those in the worst health. In order to make the British and Italian health variable comparable, the Italian 5-point scale was collapsed as follows: good (4,5); fair (3); and not good (1,2).
Previous studies had shown that education had a negative effect on proximity, so that those individuals who were more educated lived further from their children (
Clark and Wolf 1992;
Lin and Rogerson 1995). In order to retain comparability, given differences in the educational systems of both countries and the more limited information on education available in the British data, individuals with high educational levels were distinguished from those with lower levels. The British Retirement Survey asked respondents for the highest educational qualification attained. For the purposes of this analysis, individuals with a university degree, qualifications obtained from public examinations taken at secondary school usually at ages 16 and 18, or with a clerical, commercial, or trade qualification were grouped into the high-education category, and those with no educational qualifications were grouped into the medium/low category (this included both individuals who may have attended some school but did not achieve an educational qualification, and those with no education). In Italy, the educational category was more detailed, so that those in the high-education group were classified as having a university degree or a high school diploma (usually obtained about age 18), and those in the medium/low-education group had either less than a high school diploma or had received no education.
Housing tenure, given its strong association with proximity in Britain, was also included in the model (
Warnes and Ford 1995b). In order to have similar categories between the two countries, a dummy variable was created that distinguished owner–occupiers from those in other tenures (largely social sector tenants in Britain and private renters in Italy).
Model
The aim of this analysis was to study the correlates of older women's proximity to children and, additionally, to examine demographic and socioeconomic factors that may affect the proximity of adult children (aged 25–39) to their mothers. We used a multinomial logit model for the probability distribution across three possible categories of proximity between older women and their nearest adult child (0 = coresidence; 1 = up to 10 miles; 2 = more than 10 miles). For clarity and ease of interpretation, the coefficients from all three contrasts are presented in Table 4 , although this is redundant as the coefficients are additive.
|
|
Results |
|---|
|
Multivariate Analysis
The coefficients, their corresponding standard errors, and level of significance are presented in Table 4 . In interpreting the Betas, recall that they represent the effects of the explanatory variable on the log-odds ratios of the following probabilities: (1) living close (up to 10 miles) versus coresidence with nearest child; (2) living far (more than 10 miles) versus coresidence with nearest child; and (3) living close versus living far from nearest child.
As indicated in Table 4 , individuals aged 70–74 in both countries were significantly less likely to coreside when compared to those in the younger age group. In Italy, even among those aged 65–69, there was a negative effect of age on coresidence.
Marital status had no effect in Britain, whereas in Italy being widowed had a significant positive effect on coresidence when compared with living close to, or far from, the nearest child.
The number of children had the same effect for both countries, that is, the more children a woman had the more likely she was to coreside with her nearest child (vs living close or far), and the more likely she was to live close to her nearest child rather than to live far. Higher parity women were more likely to have had their last child at a later age than lower parity women, increasing the likelihood of having a child at home in later life.
Health had no effect on coresidence in either country, although in Britain poor health was positively associated with living close to the nearest child versus living far.
Low education had a positive effect for Italy on living close versus coresidence: the fact that more highly educated women were more likely to coreside could be explained by the greater likelihood that their own children were in higher education. For both countries low education had the same effect (those with lower levels of education were more likely to live close than to live far). In Britain, there was also a significant negative relationship between lower education and living far from the nearest child compared with coresidence.
Another way of presenting the results from the multinomial logit model is to calculate the probabilities of each category of the dependent variable for selected characteristics of women. Two profiles were created, the "less favored group" ( Fig. 1, representing women with greater potential need), and the "more advantaged group" ( Fig. 1, reflecting those with more optimal characteristics). As shown in Fig. 1, in Britain the probability of living with a child or in close proximity is greater among women in the less favored profile group. In Italy the pattern is reversed, with women in the more advantaged group having higher probabilities of coresidence with children.
|
;
Greenwell and Bengtson 1997;
Wolf 1994). As women aged 60–74 are representative of the younger elderly, their children's characteristics, rather than their own, may have a greater impact on proximity. Consequently, differences in the socioeconomic circumstances of children, rather than differences in parental characteristics, may provide a further explanation for variations in intergenerational proximity between the two countries. For this reason we also examined similarities and differences in the association of adult children's characteristics and proximity to their mother (for those with a living mother) in Britain and Italy using the 1995 BSAS and the 1995 Italian Indagine Multiscopo sulle Famiglie.
Tabular Analysis
Table 5 shows the demographic and socioeconomic characteristics of British and Italian adult children (aged 25–39) with a living mother, as well as the associations between these characteristics and proximity to the mother (see Appendix, Note 1). In the BSAS, proximity was created from a combination of two variables, one identifying coresidence and the other journey time to mother's home. Thus, those who were living under one hour's journey from their mother's home were classified as living close, and those living more than one hour away were categorized as living far. For Italy, the proximity variable remains the same as in the previous analyses.
|
A lower proportion in Britain were single than in Italy (19% vs 35%), reflecting both higher rates of cohabitation in Britain and the slightly higher mean age at marriage in Italy (see Appendix, Note 2). For example, 23% of childless women aged 25–29 were cohabiting in the United Kingdom versus only 0.3% in Italy (
Iacovou 1998). Unemployment levels among this age group were similar in the two countries, although for those under age 25 unemployment levels in Italy are substantially higher (
Eurostat 1996).
Although there were no sex differentials in proximity in Britain, gender differences were important in Italy, with Italian men being more likely to coreside with parents. Whereas only 9% of British men aged 25–39 lived with their mothers, 39% of Italian men did so. Although 31% of single people in this age group were living with their mothers in Britain, 82% of single people in Italy were living with their mothers. Adults in this age group in Britain with a high level of education were more likely than their less educated counterparts to live far from their mothers. On the other hand, Italian 25–39-year-olds with higher levels of education were more likely to coreside than those with lower levels of education. Finally, only 8% of unemployed adults in this age group coresided with their mothers in Britain, but 64% of their Italian counterparts were living with their mothers.
The postponement of the transition to adulthood in Italy appears to affect proximity, and in particular coresidence, with parents. Consequently, proximity in Italy may be more likely to be a response to the needs of children than of parents in the young elderly age group.
|
Discussion |
|---|
In addition, children's socioeconomic background in Britain had less of an effect on proximity and coresidence than in Italy; thus, intergenerational proximity in Italy may be more likely to be a response to the needs of children than parents. Delays in the transition to adulthood in Italy have meant that adult children are increasingly relying on continued parental support. The strong norms in Italy concerning mutual aid between parents and children are likely to have influenced the continuing assistance provided by elderly parents to their adult children.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
The British Social Attitudes Survey 1995 was made available by Social and Community Planning Research, London via the Data Archive at the University of Essex. Responsibility for the analysis and interpretation of data reported here rests with the authors.
Received for publication October 12, 1999. Accepted for publication May 23, 2000.
|
Appendix ENDIX |
|---|
2. In 1994 the mean age at marriage was 29.4 and 28.2 for Italian and British men, respectively, and 26.4 and 26.1 for Italian and British women, respectively.
|
References |
|---|
This article has been cited by other articles:
 |
|
 |
|
  F. C. Billari Europe and its Fertility: From Low to Lowest Low National Institute Economic Review, October 1, 2005; 194(1): 56 - 73. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
TOP
Abstract
Trends in Living Arrangements...