“The Apple doesn’t Live Far from the Tree”: Living Distances between Parents and their Adult Children in Europe Bettina Isengard Abstract: Geographic proximity or distance between parents and their adult chil- dren is a fundamental, if not the decisive prerequisite for intergenerational solidari- ty. But why do some parents and their children live closer together than others? And why are there national differences in Europe? The objective of this article is to better understand the causes of geographical proximity or distance using the SHARE data of 14 European countries. In addition to personal characteristics of the parents and children, familial structures and cultural contextual differences between the coun- tries are also in the focus of interest. The fi ndings suggest that especially age and family-cycle infl uences have an impact on the living distance between the genera- tions, but that socio-economic and origin-specifi c correlations are also important. A comparison reveals that geographical proximity or distance varies across countries. In the south of Europe parents and adult children live far closer together, which is not merely due to co-residence. The differences can primarily be ascribed to cul- tural as well as institutional infl uences and the associated social consequences. Keywords: Geographical distance between family members · Intergenerational relationships · Europe · SHARE 1 Introductory remarks Even in modern societies familial intergenerational relationships continue to be the most stable interpersonal bonds and have even become more signifi cant in recent decades. Since (measured in numbers) less and less children will – due to an in- creased life expectancy – spend a longer lifetime together with their parents and even their grandparents (cf. Bengtson 2001; Szydlik 2000: 11). This is mainly a con- sequence of ageing processes and the demographic change, which is manifested in (1) the increasingly declining birth rate, (2) a drop in marriages and the associated rise in cohabitation, (3) the constantly rising number of divorces, (4) the decline in Comparative Population Studies – Zeitschrift für Bevölkerungswissenschaft Vol. 38, 2 (2013): 237-262 (Date of release: 06.06.2013) © Federal Institute for Population Research 2013 URL: www.comparativepopulationstudies.de DOI: 10.4232/10.CPoS-2013-09en URN: urn:nbn:de:bib-cpos-2013-09en4 • Bettina Isengard238 multi-person households as well as the increase in one-person households and (5) the rise in the average life expectancy (cf. Hoff 2006). This does not only alter the quantitative, but also the qualitative relationship structures between the generations. While in earlier times low life expectancy meant that the family mainly fulfi lled a socialising and reproductive function, today fam- ily members are increasingly dependent on mutual solidary assistance. Intergen- erational solidarity is a chief element of familial support networks (Bertram 1997). Besides the subjective signifi cance of intergenerational relationships, marked by emotional closeness and affection (affective solidarity), the bonds are also refl ected in the contact frequency and mutual activities (associative solidarity). Yet the most striking aspect in today’s societies is the extent of functional solidarity, i.e. the giv- ing and taking of money, time and space, as numerous empirical studies show (cf. e.g. on fi nancial transfers Deindl 2010; on assistance Brandt 2009; on care Haberk- ern 2009 as well as on co-residence Isengard/Szydlik 2010, 2012). Beside existing family structures and the norm of solidarity (cf. Szydlik 2000: 85), the geographical proximity or distance between generations is a central prerequi- site for many types of intergenerational solidarity. For instance, the living distance is not only decisive for the assistance and care within family networks, but also for the care of grandchildren (Igel 2012). It is only possible for parents and their adult children to render regular acts of assistance, for which direct contact is essential, when they live close enough together. Exceptions to this are fi nancial assistance, which can easily be transferred across greater distances, or types of assistance that can basically also be given over the phone such as emotional support or assistance with administrative matters. Nevertheless, many types of (intergenerational) sup- port are dependent on geographical proximity or distance. Although it is known from earlier fi ndings for Germany and other European countries that parents and their adult children usually do not live far from one an- other (Kohli et al. 1997; Lauterbach 1998; Hank 2007), there are individual and fam- ily related differences in geographical proximity or distance. In addition, cultural contextual circumstances also have a major infl uence on living distances and ex- plain differences within, but primarily also between countries. These convergences and divergences have not yet been given suffi cient attention, however. Although there are some studies comparing geographical proximity or distance, these are in most cases studies that only examine two nations (cf. e.g. Bordone 2009; Glaser/ Tomassini 2000; Höllinger/Haller 1990) and that do not analyse the culturally and institutionally related causes more closely. The study by Hank (2007) examining ten European countries and ascribing the differences to such aspects as divergent familial bonds also does not supply a systematic country comparison that examines the correlation between living distances and social circumstances outside the con- text of the welfare state more closely. Especially, cultural contextual structures may be able to supply a signifi cant contribution to explain the differences within Europe. The following study aims to close this research gap. The research issue is therefore two-fold: (1) What determines the geographical proximity or distance between parents and their adult children in Europe? and (2) What cultural and institutional factors can explain national differences? To answer Living Distances between Parents and their Adult Children in Europe • 239 these two questions, cluster-robust ordered logit models for 14 countries based on the Survey of Health, Ageing and Retirement in Europe (SHARE) are estimated. In addition to personal and familial indicators, the research interest is mainly focused on macro-structural causes in order to illuminate the previously quite neglected sphere of culturally and institutionally related convergences and divergences more precisely. 2 Theoretical background and empirical evidence Intergenerational solidarity depends on three central basic prerequisites (Szydlik 2000: 85): (1) the relevant family structures should exist (in this case: familial gen- erations), (2) the living distance should not be too great for many forms of solidarity and (3) a certain feeling of obligation to assist family members (solidarity norm) should dominate. Against this background, the living distance between the gen- erations is an important prerequisite for a number of types of support. Unlike co- residence, i.e. giving and taking living space, geographical proximity is not a direct form of solidarity, but it makes many forms of intergenerational solidarity possible at all. Nevertheless, the theoretical assumptions and supposed effects of the forma- tion of intergenerational solidarity can be transferred to the living distance between the generations. This is possible since the geographical proximity or distance on the one hand (in the form of the special case of co-residence) is a direct form of func- tional intergenerational solidarity and is on the other hand indirectly closely linked to it (as a potential). In order to explain why some parents and their adult children live more close- ly together than others and why there are specifi c national differences in Europe, the theoretical model designed by Szydlik (2000: 39) to explain intergenerational solidarity is used (cf. also Bengtson/Roberts 1991). According to this model, it is as- sumed that three central structures have an infl uence on intergenerational support. These are (1) individual possibilities and needs, (2) familial factors and (3) cultural contextual structures. All three of these groups are also relevant for explaining the living distance between parents and their adult children. Theoretically, they assume that the personal needs and the possibilities of the adult children and their parents infl uence the formation of solidarity. Transferred to the geographical proximity or distance this would mean that a lesser or a greater living distance depends on the personal needs or opportunities. Yet in this context, being embedded in familial structures is also relevant, since “competing” family members can decrease the necessity for proximity. According to the model, differences within but primarily between countries can be ascribed to divergent cultural contextual circumstances. But how should be adequately dealt with co-residence, in this case cohabitation of parents and their adult children? Previous fi ndings have shown that in addition to age and its associated status within the life and family cycle, mainly the children’s economic needs either promote or impede co-residence (cf. e.g. Aassve et al. 2002; Choi 2003; Deindl/Isengard 2011; Isengard/Szydlik 2012). Therefore, taking co-resi- dence into account in countries with high rates leads to overestimations of the infl u- • Bettina Isengard240 ence of economical resources in a joint approach. At the same time, a theoretical and empirical elimination of co-residence would only lead to an incomplete picture of geographical proximity or distance, since this type of living arrangement is very frequent in some European countries and thus would exclude a fundamental part of the population per se. Against this background, it appears wise to look at both manifestations in the empirical section – including and excluding co-residence – in order to reveal the causes and reasons for living distances. Previous research (with the exception of the study by Hank 2007) mainly focuses on individual countries or on two-country comparisons and identifi es three indica- tor groups at the individual and familial level. First, age and life cyclical patterns are important that also result in gender-specifi c differences. In general, it becomes apparent that the geographical distance or proximity between the generations is considerably determined by age (cf. Clark/Wolf 1992; Lin/Rogerson 1995). Moving out of the parental home is usually the fi rst step into independence, which often does not lead too far away. Using data from the German Socio-economic Panel (SOEP), Leopold et al. (2011) can show that on their fi rst move children on average only move about ten kilometres away. For parents, age usually plays a role in com- bination with health (Silverstein 1995). In addition, increasing isolation caused by divorce or the death of a spouse or partner and of friends with advanced age can lead to parents and children moving (again) closer together (De Jong Gierveld/van Tilburg 1999; Dykstra et al. 2005). Age-specifi c effects have not only a direct, but also an indirect infl uence since they are closely associated with family cycle processes (White 1994) – and hence with familial structures according to Szydlik (2000). Entering or being in a relation- ship, the birth of a child or grandchild, separation or dissolution of relationships through divorce or death: all of these events can have direct effects on the living distance between the generations. Two opposite trends are apparent for geographi- cal proximity or distance. Although children frequently leave their parents’ home when they enter into a relationship (for Europe Iacovou 2001: 10), at the same time, they usually do not move as far away as singles (Glick/Lin 1986; Aquilino 1990; Lau- terbach/Pillemer 2001 for Germany and the USA), although the opposite is the case in the Netherlands (Michielin/Mulder 2007). The distances are lesser for the parent generation if mothers are single or both parents live together. This is not the case for widowed fathers; here the distances between the generations are also greater (Lawton et al. 1994). Smits et al. (2010) fi nd for the Netherlands that children live more rarely with their parents when the parents are not single, but live with them far more often if they themselves are not married but have children. When parents have a number of children this also has a direct effect on the dis- tance to each individual child (Shelton/Grundy 2000). While – from the child’s point of view – siblings enlarge their geographical distance to the parents, for the parents it is more probable that they have at least one child living close to them (Hank 2007: 159). Grandchildren usually also lead to the generations living closer together (Crim- mins/Ingegneri 1990; Rogerson et al. 1997; Pettersson/Malmberg 2009). However, Madigan and Hogan (1991) discover that single mothers do not live closer to their parents than mothers who cohabitate with the child’s father. Living Distances between Parents and their Adult Children in Europe • 241 The gender of the children also has an infl uence on geographical proximity or distance, whereas scientifi cally the correlations are not quite uniform. Although it is unquestioned that adult sons live together with their parents far more frequently (cf. e.g. Billari et al. 2001), but the picture is not as clear for living distances outside the household. Empirically it has been shown that although daughters live less often under one roof with their parents, when they do move out, they do not move very far away (Clark/Wolf 1992; Lawton et al. 1994). For Sweden Pettersson and Malm- berg (2009), however, show that daughters live further away from their parents than sons. Yet even in this context, the familial situation is relevant. Leopold et al. (2011) fi nd out that when daughters leave their parents’ home, they usually move further away than sons, but this is only true when they leave a couple household. A current study for the Netherlands shows that couples live closer to the husband’s parents than those of the wife. However as soon as young children live in the household, the family structures of women gain infl uence (Blaauboer et al. 2011). All in all, based on the theoretical assumptions and previous empirical fi ndings it can be assumed that with increasing age and position in the family cycle (rela- tionship, parenthood, etc.) the probability is greater that parents and their adult children live further apart (hypothesis 1a). Since women usually marry earlier and have children, the geographical distance should be greater for daughters than for sons (hypothesis 1b). The personal needs and opportunity structures used by Szydlik (2000) to explain intergenerational solidarity are closely associated with the second indicator group: the socio-economic variables. Earlier fi ndings suggest that the geographical dis- tance between the generations’ increases with rising status (Bengtson/Harootyan 1994; Clark/Wolf 1992; Silverstein 1995), since this considerably infl uences personal opportunities and needs. Key factors are education and the employment status as well as the associated level of prosperity. A higher level of education is related to a greater geographical distance, since higher educated children and parents usually have better career opportunities, which are often associated with greater demands for mobility and are frequently already relevant during studies (Kalmijn 2006). But the fi nancial opportunities, primarily of the parents, can also have an infl uence. In It- aly for instance, due to the poor labour market situation and high rents, parents can frequently control where their children move either by passing residential property on to them or by fi nancially supporting them and by traditionally being involved in the choice of residence (Glaser/Tomassini 2000; Giannelli/Monfardini 2003). Analo- gous to the assumptions of the heuristic model by Szydlik (2000) and on the basis of the state of research for individual countries, it is expected that better socio- economic opportunities and lesser need structures are accompanied by a greater geographical distance between the generations (hypothesis 2). Thirdly, the effects of origins can infl uence the living distances between the gen- erations. On the one hand, this can be the social origin, which is closely linked with socio-economic structures. Children with a more highly educated background are themselves frequently more highly educated and thus they should be more accus- tomed to mobility than children from less educated social classes (on inherited edu- cation see Schimpl-Neimanns 2000). But regional origins are also important (Hank • Bettina Isengard242 2007; Lauterbach 1998; Lauterbach/Pillemer 2001), because they considerably de- termine the opportunities and needs of individuals. In larger cities and conurbations the labour market opportunities as well as the supply of educational institutions are usually far better than in rural or isolated regions. This results in parents and chil- dren from urban regions living closer together more often since the “compulsion” to move is lesser due to better opportunity structures than in the countryside. The housing market can also have an infl uence. In rural regions there are often only few rented fl ats available and therefore the necessity of moving further away from the parents can be greater than in the city (Georg et al. 1994). Finally, ethnic origins can also explain the geographical proximity or distance between the generations. Two contrary trends are possible. On the one hand generations might live further apart because they live in different countries; on the other hand a migration background can lead generations not to move far apart since they often take advantage of a sup- port network of relatives and friends from their country of origin, which is settled close together (Aslund 2005). Mulder (2007) can empirically prove that parents and children with migration backgrounds live closer together (see also Angel/Tienda 1982 for the USA, who confi rm more frequent co-residence). In the tradition of pre- vious fi ndings for individual countries, it is assumed that also children in European countries with parents of high social status (hypothesis 3a), without a migration background (hypothesis 3b) and from rural regions (hypothesis 3c) live further away from their parents than children from less-educated classes, with migration back- grounds and from cosmopolitan or urban regions. Finally and fourthly, cultural contextual structures, which vary socially and help explain national patterns, are relevant. Generally it can be assumed that not only co- residence of parents with their adult children is a far more common living arrange- ment in southern European countries than in continental Europe and in particular in comparison with northern Europe (Kiernan 1999), but also that living distances in general are closer (Hank 2007). Some of the possible convergences and diver- gences are related to welfare state arrangements. On the one hand, the need or necessity for intergenerational solidarity is structured by the welfare state and on the other hand the attitudes towards and signifi cance of the family are closely tied to the historic development of the welfare states. In countries where the church had a strong role in the development of the so- cial state, the social and, in particular, the family-related benefi ts are still rather weakly today and the family is more important than in countries with extensive benefi ts and well-developed social systems (for a comprehensive portrayal, see Esping-Andersen 1998 as well as modifi cations by Leibfried 1992 and Keune 2009). Correspondingly, in social-democratic systems such as those in Denmark and Swe- den, the role of the family is rather weak, since social benefi ts are very extensive. By contrast, the signifi cance of the family in the continental European, conservative countries, among them Germany, France and Belgium, is rather strong due to the infl uence of the church. In the southern European, familialistic welfare systems such as Italy, Spain and Greece, the church had a signifi cant infl uence on the design of the social benefi ts. In an OECD comparison, this means that the family-related, pub- lic support payments are quite low (cf. for example Ferrera 1997; Sciortino 2004), Living Distances between Parents and their Adult Children in Europe • 243 but at the same time strong, traditional family structures predominate. In the post- Socialist systems such as those in Poland and the Czech Republic even today the state is “universally responsible” due to the experience of earlier times. The effects of the transformation processes can only gradually become effective. Presently we can expect that the demand for intergenerational solidarity is still relatively pro- nounced in these countries due to the lack of state infrastructures. In the liberal wel- fare states, which are mainly spread in the Anglo-Saxon area, there is means-tested public welfare with few universal benefi ts. In this context, Giuliano (2007) ascribes the differences between the countries to cultural divergences, which are expressed in a differing self-concept of fam- ily. According to Reher (1998), family relationships can typically be classifi ed as “strong ties,” which traditionally predominate in southern European countries, and the “weak ties” traditionally existing in northern and western European countries. Thus it is “not diffi cult to identify areas where families and family ties are relatively ‘strong’ and others where they are relatively ‘weak’ [and] these divergent practices appear to have deep historical roots” (Reher 1998: 203 and 204). The role of the family is one of the driving forces behind the cultural divergences, but the actual social benefi ts and their direct and indirect consequences for the objective living conditions also play a role. For example, welfare benefi ts in general have direct ef- fects on poverty and inequality, but also on (regional) unemployment and thus on the necessity of moving or generating the family as a support network. But actual family-related benefi ts such as maternity leave or child benefi ts can have an infl u- ence (Neyer 2003). In general, these benefi ts are far more comprehensive in the north than in the south. Against this (theoretical) background, it is expected that country-specifi c differ- ences can be explained through institutional and cultural circumstances. It can be assumed that (1) the less comprehensive the welfare benefi ts in a country, the small- er will be the geographical distance between the generations (hypothesis 4a); (2) the greater the social inequality, the shorter are the living distances (hypothesis 4b) as well as (3) the more important the familial obligation norms and the role of the family, the closer parents and their adult children will live together (hypothesis 4c). The cultural contextual explanatory factors are closely linked here. In countries with weak social welfare systems – also due to the infl uence of the church– the family has a greater signifi cance and correspondingly social inequality is more pronounced due to a lesser redistribution. Living distance between generations is not only related to institutional and cul- tural circumstances, but is also infl uenced by geographical circumstances. National differences are also ascribed to whether there is a chance at all to move far away. If we disregard emigration, in smaller, densely populated states it is more diffi cult to live far apart from one another than in countries with larger areas. It is therefore assumed that the larger (in area) a country is, the further apart the parents and their children will live (hypothesis 5a), but the denser the settlement structures, the lesser will be the living distances between the generations (hypothesis 5b). • Bettina Isengard244 3 Data and methods 3.1 Data The second wave of the SHARE data (Survey of Health, Ageing and Retirement) co- ordinated by the Munich Center for the Economics of Aging (MEA) were used for the analyses. This is a data set supplying relatively consistent information about older people for fourteen European countries. They are Belgium (BE), Denmark (DK), Ger- many (DE), France (FR), Greece (GR), Ireland (IE), Italy (IT), the Netherlands (NL), Austria (AT), Poland (PL), Sweden (SE), Switzerland (CH), Spain (ES) and the Czech Republic (CZ). The advantage of these data is that the procedure is standardised in the countries and that a broad spectrum of subjects is measured. The survey covers people over 50 years of age as respondents and their (younger) partners and asks about various topics such as demographics, income, health, residence, education and employment, behaviour, social support, activities and expectations. In addition, the partners living in the same household, who may also be younger, are interviewed. A total of approximately 33,000 people were interviewed, each of whom provided information about their parents and their children, of whom the age, gender and living distance are asked. Additionally more specifi c information is included about four children, chosen by the parents, such as marital status, em- ployment status, the number and the age of their own children and highest level of educational attainment. Since not many parents in Europe have more than four children, the number of cases not accounted for is limited (almost 4 percent). The living distances between the parents and all of their adult children, for whom this additional information is available, are analysed in this article. 3.2 Operationalisation The living distance between parents and their adult children is captured by using the following variable: “Please look at card 5. Where does {child name} live?: (1) In the same household, (2) in the same building, (3) less than 1 kilometre away, (4) be- tween 1 and 5 kilometres away, (5) between 5 and 25 kilometres away, (6) between 25 and 100 kilometres away, (7) between 100 and 500 kilometres away, (8) more than 500 kilometres away, (9) more than 500 kilometres away in another country.” Based on this question, the dependent variable is formed. Essentially, the ordinal categories are kept and are used in the fi rst model estimation. Only the categories 8 and 9, which measure living distances of more than 500 kilometres (once within a country and once outside a country), are combined. Since co-residence is a direct form of intergenerational solidarity, in a second model specifi cation children who live in the same household or the same building (near co-residence, see Isengard/ Szydlik 2012) with their parents are excluded to see whether the results differ sub- stantially. This is because co-residence can be explained by the economic needs of the children to a large part (cf. Deindl/Isengard 2011). As explanatory variables factors are used that refl ect the age, gender and life- cycle determinants as well as socio-economic variables and origin effects. In order Living Distances between Parents and their Adult Children in Europe • 245 to explain the country-specifi c particularities, cultural contextual variables are taken into account. Since the age of the parents and their adult children are strongly cor- related a combined age variable is included into the models. Younger children (18+ to 29 years of age) and older children (from 30 years of age) are each combined with younger parents (up to 64 years of age) and older parents (from 65 years of age); the age of the oldest parent in the household is used here. The familial structures taken into account, which are closely associated with the age and the position in the life cycle, are the relationship status of the parents and children, the number of children of the respondents (corresponds to the number of siblings of the children) as well as the number of grandchildren (or number of children’s children). The relationship status of the children is included as a dummy variable, whereby children living in a relationship (marital or non-marital) are compared to children living alone. The status of the parents is linked with the gender of the children, since, as previous fi ndings show, the gender-specifi c living distance varies with the parents’ relation- ship status. For data technical reasons, the gender of the parents is not included because the information about the living distance is available at the household level and is thus identical for both parents. Four dummy variables are taken into account: daughter-relationship parents, daughter-parent alone, son-relationship parents, son-parent alone. Moreover, the additional number of children is included in four dummy variables and the variables measure whether someone has next to the each recorded child no additional child, has one additional child, has two more children or has three and more children. The same differentiation is used for the number of grandchildren (none, 1, 2; 3 and more). The health of the parents is also an indicator that is closely linked with age and refl ects a possible need for solidarity, in particu- lar if the parent has no (or no longer a) partner. For this reason, an interaction term between health and the existence of a relationship is formed. Health is recorded as the number of restrictions according to ADL and IADL and added. These are instru- ments that measure restrictions in activities of daily living (ADL) as well as instru- mental activities of daily living (IADL). ADL includes, for example, eating, dressing, personal hygiene and walking, IADL includes shopping, preparing meals, house- work and managing money. The value for the parent with the most restrictions is used and interacted with the relationship status, which is a dichotomous variable. As socio-economic variable the education of the parents is used as well as a proxy for the economic situation of the household, the “living on earnings”, an indi- cator that measures how well the household makes ends meet. The personal level of education is measured according to the ISCED classifi cation (International Stand- ard Classifi cation of Education), developed by the UNESCO to make different school systems and types comparable (cf. OECD 1999). Three categories are established that portray a low (ISCED 1), intermediate (ISCED 2) and a high level of education (ISCED 3). The so far highest attained school education or professional achievement is recorded as the qualifi cation. People who are still attending school and have no certifi cate yet are subsumed under ISCED 1, as are people who are still attending vocational school but only completed compulsory schooling so far. People with an “other” achievement are also included in the lowest category since this means there is no information about their level of qualifi cation available in this case. Since this • Bettina Isengard246 only applies to less than 1.5 percent of the cases, however, this type of categorisa- tion does not affect the results. The assessment of the individual income situation is stored in four dummy variables and implies the following categories: with great diffi culty, with some diffi culty, with reasonable ease and with ease. This variable is preferred over the actual income since it highly correlates with the educational level thus making the estimation instable. The economic opportunities of the children are reproduced through the education, which was measured analogously with the parental information, as well as through the employment status. This is done us- ing six dummy variables that differentiate between employment, unemployment, in training, economic inactivity (housewife/househusband), (early) retirement and a residual category (other). The third group of variables includes the origin effects. The social origin over- laps with the economic variables. Moreover, the regional origin (city vs. country) and the ethnic origin are measured. The community size, which is developed using variables generated by SHARE, only differentiates between city and country and thus combines the original fi ve categories, which were large city, suburbs of a large city, large towns, small towns as well as rural regions and villages. People living in a (large) city or a suburb are given the specifi cation of city, while people living in smaller towns and communities fall under the category of villages and municipali- ties. It is also taken into account whether it is a household with a migration back- ground. For this, at least one parent must cite that they were born abroad or have no domestic citizenship. Finally, in order to analyse the infl uence of cultural contextual differences in greater detail, a model is estimated containing all of the individual and familial indi- cators of the basic model (with fourteen country dummies), the respective macro- indicator however is varied. These are the (1) classifi cation in four welfare state regimes as well as indicators that (2) refl ect the extent of social benefi ts, (3) meas- ure poverty and inequality, (4) consider the labour market situation as well as (5) take geographical circumstances into account. The countries are classifi ed in the four regime types conservative (AT, BE, DE, CH, FR, IE, NL), social democratic (DK, SE), familialistic (ES, GR, IT) as well as post-socialist (CZ, PL). Since no liberal social states are included in SHARE, unfortunately this type cannot be examined here. The extent of social benefi ts in the year 2006 is incorporated using social expenditures in general (per capita in current prices and adjusted for purchasing power in US dol- lars, OECD) as well as family expenditures (in percent of the GDP, OECD). The level of prosperity is indicated using the country-specifi c poverty rate (60 percent medi- an, new OECD scale) and the extent of income inequality using the Gini coeffi cients also from 2006 (UNDP 2006). Additionally, the social labour market situation is taken into account using the country-specifi c (youth) unemployment rates (according to Weltalmanach 2006). The signifi cance of family is operationalised using the percent- age of non-marital births in all births in a country in the year 2006 (OECD) as well as using an index derived from the SHARE data indicating familial obligation norms. For this, six individual items that indicate such norms were added and recoded so that high values indicate approval of the family and its responsibilities. Then, the average value was ascertained for each country and used as macro-indicator in Living Distances between Parents and their Adult Children in Europe • 247 the analyses. The bases are the following questions: (1) The following statements are related to the duties people may have in their family. Please tell us how much you agree or disagree with each statement: (a) Parents’ duty is to do their best for their children even at the expense of their own well-being; (b) Grandparents’ duty is to be there for grandchildren in cases of diffi culty (such as divorce of parents or illness); (c) Grandparents’ duty is to help grandchildren’s parents in looking after young grandchildren; (2) In your opinion, who – the family or the State – should bear the responsibility for each of the following…? (a) Financial support for older persons who are in need? (b) Help with household chores for older persons who are in need such as help with cleaning, washing? (c) Personal care for older persons who are in need such as nursing or help with bathing or dressing? Finally, the geographical circumstances are taken into account by recording the population density and the national area (Weltalmanach 2006). Other macro-indicators which are relevant for co-residence like the rate of residential property and the percentage of rent costs in the total expenditures of the households prove irrelevant for geographical proxim- ity or distance between generations and are therefore not taken into account in the following analyses (cf. Isengard/Szydlik 2010). 3.3 Methods Since the dependent variable, the living distance, is a categorical variable, ordered logit models are estimated (cf. e.g. Greene 2003; Agresti 2002). These are an en- hancement of the binary logit regression and are used when the dependent vari- able has an ordinal scale and has more than two categories that can be ranked hierarchically. The size of the intervals between the categories is insignifi cant. The corresponding regression equation is: Exactly one regression coeffi cient is estimated for each of the independent vari- ables, but a number of constants β 0i , which can be interpreted as threshold values between the categories of the dependent variable. The probability of falling in categories 1 to i of a dependent variable, in this case the living distance, with j categories results from it being correlated to the probabil- ity of falling in the category i+1 to j. The probability of an event (y) is calculated according to: The question of the living distance from each individual child was only answered once per household so the information has to be transferred to the non-responding partner. This, however, results in no variation in the responses since the living dis- tance to the parental home is a circumstance that cannot vary for both parents if they live together in one household and if it was not recorded when the parents 1... / 1... 0 1 1 2 2( | ) ...i i I i k kLogit Y X X X X . (1) 1Pr( ) ( < )i j j i j y i Pr k X u k . (2) • Bettina Isengard248 lived separately. For this reason, the information was regarded at the household level and the data were processed so that for each family one line per child exists in the data set. Since, however, multiple observations may exist per household, in these cases robust standard errors were estimated for clustered data according to households (cf. Bye/Riley 1989). Using an additional Stata ado (cf. Petersen 2009) it is also possible to calculate cluster-robust standard errors for two dimensions, how- ever only for linear regressions and logit, probit and tobit models. For this reason, additional models (not shown here) were estimated for the macro-indicators (Ta- ble 2 in section 4.2), which treat the living distance as a linear variable and estimate cluster-robust standard errors according to households and countries. The results proved to be very robust and show no notable differences compared with the one- dimensional observation. Mainly independent variables, which are additively linked, are included in the model estimations. However, an additional interaction term is used between the number of health restrictions (metric) and the relationship status of the parents (cat- egorical), which indicates a multiplicative correlation. The effects of the number of restrictions without and with partner each show the main effect; the interaction term denotes whether both groups differ signifi cantly from one another (cf. in detail Aiken/West 1991). 4 Empirical fi ndings 4.1 Living distances in Europe A fi rst look at the living distances between parents and their adult children reveals distinct differences in the European comparison, in particular between the north and the south (cf. Fig. 1). Especially co-residence occurs relatively frequently in the southern countries (Greece, Italy and Spain) as well as in traditionally Catholic Po- land. In Denmark and Sweden, by contrast, parents and their adult children rarely live together under one roof. But here, too, the generations (irrespective of co-res- idence) live close together, whereas the trend of geographical proximity is again more pronounced in the southern European countries where about 80 percent live in a radius of 25 kilometres, while in Denmark and Sweden but also in France, by contrast, it is only almost 50 percent. In France and Ireland large distances are far more probable than in many other countries of Europe. Germany, Austria and Switzerland lie within the European standard; about 40 percent of the parents and children live up to a maximum of fi ve kilometres apart. In the European average, which lies at roughly 50 percent, this is almost every second family. All in all, we see that the living distance increases from the south to the north. This might also be explained by the fact that co-residence plays a rather marginal role in Scandinavia where only 5 percent (in Denmark) and 6 percent (in Sweden) of generations live together under one roof. Living Distances between Parents and their Adult Children in Europe • 249 So what does geographical proximity look like if we only observe children who live outside the household or home? Excluding children which co-reside the pat- tern largely remain the same (cf. Fig. 2). In the southern countries the geographical distance is far lesser than in the north of Europe as well as in France and Ireland, whereas the Netherlands is an exception. Here, the percentage is comparable to that of Greece: slightly over 60 percent live within a maximum radius of fi ve kilome- tres apart. Austria and Switzerland lie within the European average; in Germany the living distances are somewhat greater, almost 30 percent of parents and children live more than 100 kilometres apart (of which approx. 11 percent live even more than 500 kilometres apart). In the Czech Republic, Belgium and Denmark the genera- tions rarely live more than 500 kilometres apart, perhaps due to the respective size of the country. Fig. 1: Living distances in Europe (in percent) 0 20 40 60 80 100 in percent Ø DK SE NL FR BE CH IE DE AT CZ ES PL IT GR Coresidence Up to 5 km Up to 25 km Up to 100 km Up to 500 km 500 km+ and abroad Source: SHARE, Wave 2, rel. 2.5.0, own calculations (weighted, except for Ireland). n=47.510. • Bettina Isengard250 Whether these differences also remain stable if we control for individual vari- ables and familial structures will be more closely examined in the following sec- tion. Furthermore the infl uence of cultural and institutional circumstances will be illuminated. 4.2 Determinants of geographical proximity or distance between the generations The correlations shown previously reveal that the degree of geographical proxim- ity or distance between parents and their adult children in Europe shows country- specifi c differences. These results, however, only comprise the living distances in a quantitative manner, and reveal nothing about the possible causes or backgrounds. In order to fi nd out what determines the living distance between the generations, or- dinal regression models are estimated, which take the personal, familial and cultural Fig. 2: Living distances in Europe, excluding co-residence (in percent) Source: SHARE, Wave 2, rel. 2.5.0, own calculations (weighted, except for Ireland). n=38.262 0 20 40 60 80 100 in percent Ø DK FR IE SE DE CH AT PL CZ BE GR NL IT ES Up to 5 km Up to 25 km Up to 100 km Up to 500 km 500 km+ and abroad Living Distances between Parents and their Adult Children in Europe • 251 contextual structures into account. In Table 1 the personal and familial infl uencing factors are included (controlling for countries), before then the country-specifi c dif- ferences are examined in greater detail (Table 2). It can be shown (cf. Table 1) that younger adult children (up to 29 years of age) live signifi cantly closer to their parents than older children (30 years and older). Chil- dren from the ages of 18 to 29 live particularly often close to their parents when the oldest parent is not older than 64 years. With an increasing age of the parents and primarily the children the geographical distance increases signifi cantly. This is also true when co-resident children are not taken into account, whereas here the effects distinctly weaken. In addition to age, familial structures also play a role, whereas, in line with the life cycle concept, the existence of an own family is a central variable. The relationship status of the child shows signifi cant correlations with geographi- cal proximity or distance. Nonetheless this has different effects depending upon whether co-residence is included or not. When children live in a relationship, the living distances are greater in the fi rst case and lesser in the second case. This is because although children usually leave their parental home when they move in together with a partner, they do not move far away (primarily, when children are born). With regard to gender, we only see that sons live less far away than daughters if there is only one parent left. This effect disappears however when co-residence is excluded, this is probably due to the known correlation that sons remain at “Hotel Mum” far longer than daughters. If co-resident children are excluded, the gender combined with the relationship status of the parents has no longer any infl uence. “Competing” family members strongly infl uence geographical proximity or dis- tance. If parents have at least two children the geographical distance for each child enlarges. This pattern is, however, broken when only those children are taken into account who no longer live under one roof with the parents. While a grandchild increases the geographical distance of the generations taking co-residence into ac- count, grandchildren lead to the generations moving closer together if the children (of the respondents) have already left the parental home. The health of the respond- ents combined with the existence of a relationship has no infl uence on the proxim- ity or distance of the generations if we control for other variables. Further results – not shown here – confi rm however that health has an impact on co-residence to such an extent that the probability increases with a rising number of restrictions if only one parent is still living or lives alone. If the parent has a partner, health restric- tions do not lead to an increase in children living with their parent. How do socio-economic variables affect proximity or distance? For education, the expected correlations are confi rmed: the better the parents’ education as well as their children’s education, the further apart the generations live. This is also as- sociated with the fact that both generations are generally more mobile and children already move further away for their studies. On the other hand, the economic situ- ation of the parental home, operationalised via the living on earnings, has hardly any infl uence on the living distance (but it does have an infl uence on co-residence). Things are different when looking at the labour force status of the children: un- employed children live more frequently near their parents (to a major part due to co-residence) than employed children. Economically inactive children (housewives • Bettina Isengard252 Tab. 1: Determining factors for geographical proximity or distance: Results of ordinal logit analysis Including co-residence Excluding co-residence Coeff. z-Value Coeff. z-Value Age and life course indicators Age (child up to 29 years/respondent up to 64 years) up to 29 years/65+ years .21* (2.08) .28** (2.96) 30+/up to 64 years .77** (21.12) .11** (2.82) 30+/65+ years .79** (21.80) .15** (4.04) Co-habitation, child (no) yes .65** (26.77) -.11** (-4.06) Gender child/partner, respondent (daughter-partner) daughter-alone .06+ (1.94) .05 (1.33) son-partner .03 (.93) -.01 (-0.47) son-alone .26** (5.66) .04 (0.80) Additional children, respondent (none) one .11** (3.02) .05 (1.14) two .25** (6.61) .14** (3.33) three and more .27** (6.73) .12** (2.61) Number of grandchildren, respondent (none) one .22** (7.53) -.27** (-4.06) two .04 (1.57) -.34** (-8.39) three and more -.02 (-0.46) -.33** (-10.81) Health, respondent (ADL + IADL) number of restrictions (without partner) -.01 (-1.18) .00 (0.10) IA effect: number*partner .02+ (1.91) -.00 (-0.06) number of restrictions (with partner) .01 (1.42) .00 (0.06) Socio-economic indicators Education, respondent (intermediate) low -.08** (-3.02) -.13** (-4.36) high .29** (8.55) .23** (6.75) Income situation (great difficulty) some difficulty -.07+ (-1.90) .04 (0.90) reasonable ease -.07+ (-1.85) .05 (1.16) ease -.00 (-0.05) .06 (1.11) Education, child (intermediate) low -.12** (-4.67) -.14** (-4.35) high .47** (18.66) .53** (19.87) Employment status, child (employed) unemployed -.44** (-8.88) -.04 (-0.66) in training -.64** (-9.36) .71** (11.94) homemakers .20** (5.81) .10* (2.22) (early) retirement -.15** (-2.91) -.02 (-0.39) other -.31** (-2.57) .65** (5.61) Origin effects Community size (villages and townships) (large) cities -.08** (-3.78) -.19** (-7.66) Migration background (no) yes .17** (3.87) .21** (4.71) N [HH-Cluster] 38.812 [17.486] 30.973 [15.357] Pseudo-R2 .06 .03 Note: Reference categories in italics. Signifi cance levels: ** p<.01, * p<.05, + p<.10. Z-values in brackets. The signifi cances for the main effect number of restrictions (with partner) were calculated based on the statistical post hoc probing (cf. in detail Aiken/West 1991). Models controlled for countries (results see Table 2). Source: SHARE, Wave 2, rel. 2.5.0, own calculations. Living Distances between Parents and their Adult Children in Europe • 253 or -husbands) by contrast live further away, which is linked to family cycle events (relationship and child). Finally, the origin effects also prove to be relevant. Children with a more highly educated background live further away from their parents than children from less- educated social classes. Yet while living in cities signifi cantly lessens the distance between the generations as expected, the generations live much further apart when they have a migration background. Based on the empirical fi ndings, it can be shown that individual and familial char- acteristics have a considerable infl uence on the geographical proximity or distance between generations. But what about country-specifi c differences? Table 2 con- tains the regression coeffi cients based on cluster-robust standard errors for various country-specifi c variables controlled for the individual and familial variables from the basic model. In general, the national differences found in the bivariate results are confi rmed here. In the northern European states, the geographical distance be- tween the generations is much greater than in the southern European countries. But how can these differences be explained? In order to illuminate the correlations more precisely, the results of the infl uence of various macro-indicators are shown below. The infl uence of the macro-indicators also proves signifi cant when controlling for individual variables and familial structures. In general: (1) the higher the social and family expenditure, the further apart the generations live from one another, (2) the poorer the national labour market situation, the closer parents and children live, (3) the more poverty and inequality, the lesser are the distances, (4) the greater the percentage of non-marital births and the lesser the value of family, the greater the geographical distance between the generations and (5) the larger the country, the greater the living distances and the more densely populated, the lesser the geo- graphical distance. All in all, the effects are stronger if co-residence is included; nonetheless the directions and signifi cances of the correlations remain stable. Only the country-spe- cifi c income inequality shows no effect for geographical proximity if co-residence is excluded and youth unemployment then has the effect that the living distances increase. This means that children frequently live with their parents when youth unemployment is high, but also that in countries with high rates they are apparently rather forced to move further away from their parents to escape the precarious job situation. The likelihood ratio test shows that the infl uence of the countries or types of welfare states are the greatest, but also that family expenditure as well as the social status of the family measured via the percentage of non-marital births and agreement with familial obligation norms are of signifi cance. If we only observe children who do not live with their parents, then the geographical circumstances are also important. Finally, we can note that overall the expected correlations were confi rmed. Since it is shown that age and family-cycle processes and, conditionally, also gender- specifi c differences have an infl uence on the geographical proximity between the generations (hypotheses 1a and 1b). But socio-economic structures also have an infl uence. In particular a better education of parents and children promotes greater • Bettina Isengard254 geographical distances. The income situation of the parents and the employment status of the child are, by contrast, less important. Hypothesis 2 is only partially confi rmed. Likewise, social, regional and ethnic origins play a role when explaining living distances. Children with a higher social status background from rural regions Tab. 2: Cultural contextual differences in Europe Note: Models controlled for the variables from Table 1, whereas each of the macro indica- tors were varied. 1 value was divided by 1,000 for the purpose of better portrayal. Reference categories in italics. Signifi cance levels: ** p<.01, * p<.05, + p<.10. z-values in brackets. (Number of cases: n = 38.812) Source: SHARE, Wave 2, rel. 2.5.0, own calculations. Including co-residence LR ratio test Excluding co-residence LR ratio test Coeff. z-value Coeff. z-value Country (Germany) 2704.73 940.20 Austria -.29** (-4.20) -.20** (-2.84) Sweden .48** (8.05) .34** (5.67) Netherlands -.10+ (-1.47) -.30** (-5.04) Spain -1.02** (-14.88) -.77** (-8.98) Italy -1.06** (-16.94) -.41** (-5.86) France .43** (6.69) .48** (7.59) Denmark .44** (7.44) .02 (0.33) Greece -.98** (-14.44) -.03 (-0.42) Switzerland .09 (1.47) -.09 (-1.25) Belgium -.51** (-9.40) -.54** (-9.79) Czech Republic -.34** (-5.70) -.12+ (-1.95) Poland -.84** (-13.67) -.13+ (-1.87) Ireland -.11 (-1.20) .02 (0.26) Country groups (continental Europe) 2078.20 245.84 Northern Europe .55** (18.21) .32** (10.12) Southern Europe -.93** (-27.55) -.26** (-6.40) Eastern Europe -.51** (-15.50) -.02 (-0.43) Social expenditure1 .15** (25.43) 905.94 .04** (6.61) 60.96 Family expenditure .48** (32.64) 1576.96 .20** (12.65) 228.32 Poverty rate -.10** (-28.44) 1278.20 -.03** (-8.87) 117.70 Income inequality (Gini) -.06** (-21.08) 644.52 -.00 (-0.76) 0.78 Unemployment rate -.08** (-16.28) 378.72 .02* (-2.73) 10.35 Youth unemployment rate -.03** (-15.63) 356.20 .01** (4.65) 30.31 Non-marital births .03** (30.02) 1442.66 .01** (8.93) 122.07 Familialism -.04** (-39.33) 1563.83 -.01** (-5.54) 30.67 Population density -.08** (-8.59) 92.76 -.16** (-15.83) 323.61 National area .04** (6.86) 71.05 .08** (12.12) 220.07 Living Distances between Parents and their Adult Children in Europe • 255 live further from their parents (than children from lesser-educated social classes and from urban regions (hypotheses 3a and 3c)). Contrary to the expectations, which are supported by previous empirical fi ndings from other countries, children with migra- tion backgrounds, who cite the availability of networks of relatives and friends from the country of origin as the reason for geographical proximity, live further from their parents. Therefore this analysis does not confi rm hypothesis 3b. This might be due to the fact that the older generation, for example those in retirement age return to their homeland while their children stay abroad. The empirically observed differences between the countries can be explained by referring to the type of welfare state. Social benefi ts have both a direct (hy- pothesis 4a) and an indirect infl uence on geographical distance through the vary- ing country-specifi c levels of prosperity, measured here via poverty and inequality (hypothesis 4b) the same applies to the importance of the family (hypothesis 4c). In addition to the cultural and institutional circumstances, however, geographical circumstances also play a role. The living distances rises with increasing national area (hypothesis 5a), while the settlement structure has the opposite effect: the more densely populated the country, the closer parents and their adult children live (hypothesis 5b). 5 Conclusions and outlook From a historical perspective, families are the most stable form of human co-exist- ence and, due to incisive demographic changes, in recent decades especially rela- tionships between parents and their (adult) children are gaining more signifi cance. Since most European countries demonstrate a trend towards fewer children, an ever lesser number of children spend a longer lifetime – due to the increased life ex- pectancy – together with their parents. This purely quantitative change in relation- ships is accompanied by a qualitative change and familial bonds are characterised by a lifelong solidarity. Against this background, the study of the geographical proximity or distance between the generations is not only of scientifi c interest, but it is also relevant for social policies since the living distance is a necessary prerequisite for many forms of intergenerational solidarity. While fi nancial transfers are even possible across great distances, timely support is often tied to geographical proximity: assistance in the household, with nursing or care of grandchildren, can only be provided if the givers and recipients do not live too far apart. Yet why do some parents and their adult children live closer together than others? What is it like in the single European countries? And how can country-specifi c differences be explained? As could be shown, based on the empirical analyses, the generations in all of the countries studied live relatively close to one another. Nevertheless there are personal, familial and cultural contextual differences, which are also important from a socio-political position. Not least due to processes of demographic change, the relationships between the generations are becoming more important and familial support networks often help where the state is not obligated to do so. The living • Bettina Isengard256 distance is an important prerequisite for this. At the individual and familial level, geographical proximity in Europe is promoted or impeded by (1) age and life course variables, (2) socio-economic factors and (3) origin effects. Not only individual, but also social inequality structures strongly infl uence the living distances between the generations. Against this background, co-residence cannot only be assessed as a direct reaction to socio-economic and social uncertainties – as known from the lit- erature – but also intergenerational proximity or distance as a whole. Since it can be shown that not only individual needs and opportunities of individuals but also contextual circumstances promote or impede proximity. The cultural and institutional structures can vary in each country and can partly be subject to welfare state infl uences. This results in both direct and indirect differ- ences in the geographical distances between the generations in these countries. Overall, it can be shown that in nations with generally better social benefi ts and specifi cally family-policy expenditures, parents and their adult children live further apart from one another. This might also be explained by the fact that in these coun- tries not only the familial obligation norm, but primarily the necessity to provide assistance and support decreases compared to weak welfare states. Since the geo- graphical distances become smaller with increasing poverty and income inequality at the societal level, this hypothesis is supported. This argument can be under- pinned with the example of co-residence, since this living arrangement increases in countries with the lowest welfare benefi ts. This enforces the assumption that co-residence is frequently not a voluntary living arrangement chosen by parents and children who want to live together, but rather that economic necessities or un- certainties exert infl uence. Finally, however, it should not be forgotten that in addition to personal needs and opportunities, existing family structures and social opportunity structures, an action-based perspective can also be of signifi cance. Ultimately, also the living dis- tance is (almost) always the result of an individual decision and subjective action structures play just as much a role as objective variables. The psychological and so- cial disposition of the generations is of signifi cance here. Emotional bonds between parents and children as well as personal values also infl uence the co-existence be- tween the generations. Acknowledgements I am very grateful for valuable comments on an earlier version of the paper from two anonymous reviewers and for very helpful hints from my colleagues of the research group Labour, Generation, Stratifi cation (AGES) Klaus Haberkern, Ronny König, Franz Neuberger and Marc Szydlik as well as from Martina Brandt, Christian Deindl, Corinne Igel and Tina Schmid. I would also like to thank the Swiss National Science Foundation for the fi nancial support. This paper uses data from SHARE release 2.5.0, as of May 24th 2011. The SHARE data collection has been primar- ily funded by the European Commission through the 5th framework programme (project QLK6-CT-2001- 00360 in the thematic programme Quality of Life), through the 6th framework programme (projects SHARE-I3, RII-CT- 2006-062193, COM- Living Distances between Parents and their Adult Children in Europe • 257 PARE, CIT5-CT-2005-028857, and SHARELIFE, CIT4-CT-2006-028812) and through the 7th framework programme (SHARE-PREP, 211909 and SHARE-LEAP, 227822). Additional funding from the U.S. National Institute on Aging (U01 AG09740-13S2, P01 AG005842, P01 AG08291, P30 AG12815, Y1-AG-4553-01 and OGHA 04-064, IAG BSR06-11, R21 AG025169) as well as from various national sources is gratefully ac- knowledged (see www.share-project.org for a full list of funding institutions). For details see Börsch-Supan/Jürges (2005). The data were adjusted for this study and an extensive plausibility check was conducted. References Aassve, Arnstein; Billari, Francesco C.; Mazzuco, Stefano; Ongaro, Fausta 2002: Leav- ing home ain’t easy: A comparative longitudinal analysis of ECHP data. Journal of European Social Policy 12: 259-276. Agresti, Alan 2002: Categorical Data Analysis. New York: Wiley. Aiken, Leona S.; West, Stephen G. 1991: Multiple Regression. Testing and Interpreting Interactions. Newbury Park, CA: Sage. Angel, Ronald; Tienda, Marta 1982: Determinants of Extended Household Structure: Cul- tural Pattern or Economic Need? In: American Journal of Sociology 87: 1360-1383. Aquilino, William S. 1990: The Likelihood of parent-child coresidence: Effects of Family Structures and Parental Characteristics. In: Journal of Marriage and the Family 52: 405-419. Aslund, Olof 2005: Now and forever? Initial and subsequent location choices of immi- grants. In: Regional Science and Urban Economics 35: 141–165 [doi: 10.1016/j.regsci- urbeco.2004.02.001]. Bengtson, Vern L. 2001: Beyond the Nuclear Family: The Increasing Importance of Mul- tigenerational Relationships in American Society. In: Journal of Marriage and Family 63: 1-16 [doi: 10.1111/j.1741-3737.2001.00001.x]. Bengtson, Vern L.; Robert A. Harootyan (Eds.) 1994: Intergenerational Linkages – Hid- den Connections in American Society. New York: Springer. Bengtson, Vern L.; Roberts, Robert E. 1991: Intergenerational Solidarity in Ageing Fami- lies: An Example of Formal Theory Construction. In: Journal of Marriage and Family 53: 856-870. Bertram, Hans 1997: Familienwandel und Generationenbeziehungen. In: Ebert, Sigrid u.a. (Eds.): Kinderstandort Deutschland. München: Profi l: 55-83. Billari, Francesco C.; Philipov, Dimiter; Baizán, Pau 2001: Leaving home in Europe: the experience of cohorts born around 1960. MPIDR Working Paper WP 2001-014. Ros- tock: MPIDR [http://www.demogr.mpg.de/Papers/Working/wp-2001-014.pdf, Sep- tember 2012]. Blaauboer, Marjolein; Mulder, Clara H.; Zorlu, Aslan 2011: Distances between couples and the man’s and woman’s parents. In: Population, Space and Place 17: 597-610 [doi: 10.1002/psp.648]. Börsch-Supan, Axel; Jürges, Hendrik 2005: Health, Ageing and Retirement in Europe – Methodology. Mannheim: Mannheim Research Institute for the Economics of Ageing [http://www.shareproject.org/uploads/tx_sharepublications/SHARE_BOOK_METH- ODOLOGY_Wave1.pdf, September 2012]. • Bettina Isengard258 Bordone, Valeria 2009: Contact and Proximity of Older People to their Adult Children: A Comparison between Italy and Sweden. In: Population, Space and Race 15: 359–380 [doi: 10.1002/psp.559]. Brandt, Martina 2009: Hilfe zwischen Generationen: Ein europäischer Vergleich. Wies- baden: VS Verlag. Bye, Barry V.; Riley, Gerald F. 1989: Model estimation when observations are not independent: Application of Liang and Zeger’s methodology to linear and lo- gistic regression analysis. In: Sociological Methods Research 17: 353–375 [doi: 10.1177/0049124189017004003]. Choi, Namkee G. 2003: Coresidence between unmarried aging parents and their adult children: Who moved in with whom and why? Research on Aging 25: 384-404 [doi: 10.1177/0164027503025004003]. Clark, Rebecca L.; Wolf, Douglas A. 1992: Proximity of Children and Elderly Migration. In: Rogers, Andrei (Eds.): Elderly Migration and Population Redistribution – A Com- parative Study. New York: Halsted Press: 77-96. Crimmins, Eileen M.; Ingegneri, Dominique G. 1990: Interaction and Living Arrange- ments of Older Parents and Their Children. In: Research on Aging 12: 3-35 [doi: 10.1177/0164027590121001]. Deindl, Christian 2010: Finanzielle Transfers zwischen Generationen in Europa. Wiesba- den: VS Verlag. Deindl, Christian; Isengard, Bettina 2011: Familiale Unterstützung und soziale Ungleich- heit in Europa. In: Berger, Peter A.; Hank, Karsten; Tölke, Angelika (Eds.): Reproduk- tion von Ungleichheit durch Arbeit und Familie. Wiesbaden: VS Verlag für Sozialwis- senschaften: 23-47. De Jong Gierveld, Jenny; van Tilburg, Theo G. 1999: Living arrangements of older adults in the Netherlands and Italy: Coresidence values and behaviour and their con- sequences for loneliness. In: Journal of Cross-Cultural Gerontology 14:1-24. Dykstra, Pearl A., van Tilburg, Theo G; De Jong Gierveld, Jenny 2005: Changes in Old- er Adult Loneliness: Results from a Seven-Year Longitudinal Study. In: Research on Aging 27: 725-747 [doi: 10.1177/0164027505279712]. Esping-Andersen, Gøsta 1998: Die drei Welten des Wohlfahrtskapitalismus. Zur Politi- schen Ökonomie des Wohlfahrtsstaates. In: Lessenich, Stephan; Ostner, Ilona (Eds.): Welten des Wohlfahrtskapitalismus. Der Sozialstaat in vergleichender Perspektive. Frankfurt a.M./New York: Campus: 19-56. Ferrera, Maurizio 1997: The Uncertain Future of the Italian Welfare State. In: West Euro- pean Politics 20: 231-249 [doi: 10.1080/01402389708425183]. Georg, Werner; Strzoda, Christiane; Zinnecker, Jürgen 1994: Determinanten des Aus- zugs junger Erwachsener aus dem Elternhaus. Eine Analyse mit Survivalmodellen. In: Zentralarchiv für Empirische Sozialforschung (Eds.): ZA-Information. Band 34. Köln: ZA: 106-124. Giannelli, Gianna Claudia; Monfardini, Ciara 2003: Joint Decisions on Household Mem- bership and Human Capital Accumulation of Youths: The Role of Expected Earnings and Labor Market Rationing. In: Journal of Population Economics 16: 365-385 [doi: 10.1007/s001480200119]. Giuliano, Paola 2007: Living Arrangements in Western Europe: Does Cultural Ori- gin Matter? In: Journal of European Economic Association 5: 927-952 [doi: 10.1162/ JEEA.2007.5.5.927]. Living Distances between Parents and their Adult Children in Europe • 259 Glaser, Karen; Tomassini, Cecilia 2000: Proximity of Older Women to Their Children: A Comparison of Britain and Italy. In: The Gerontologist 40: 729-737 [doi: 10.1093/ geront/40.6.729]. Glick, Paul C.; Lin, Sung Ling 1986: More Young Adults are Living with Their Parents: Who Are They? In: Journal of Marriage and the Family 48: 107-112. Greene, William H. 2003: Econometric Analysis. Upper Saddle River, NJ: Prentice Hall. Haberkern, Klaus 2009: Pfl ege in Europa: Familie und Wohlfahrtsstaat. Wiesbaden: VS Verlag. Hank, Karsten 2007: Proximity and Contacts Between Older Parents and Their Children: A European Comparison. In: Journal of Marriage and the Family 69: 157–173 [doi: 10.1111/j.1741-3737.2006.00351.x]. Hoff, Andreas 2006: Intergenerationale Familienbeziehungen im Wandel. In: Tesch- Römer, Clemens; Engstler, Heribert; Wurm, Susanne (Eds.): Altwerden in Deutsch- land: sozialer Wandel und individuelle Entwicklung in der zweiten Lebenshälfte. Wies- baden: VS-Verlag für Wissenschaft: 231-287. Höllinger, Franz; Haller, Max 1990: Kinship and Social Networks in Modern Societies: A Cross-Cultural Comparison Among Seven Nations. In: European Sociological Review 6: 103-124. Iacovou, Maria 2001: Leaving Home in the European Union. ISER Working Papers Num- ber 2001-18. Colchester: University of Essex [https://www.iser.essex.ac.uk/fi les/iser_ working_papers/2001-18.pdf, September 2012]. Igel, Corinne 2012: Großeltern in Europa: Generationensolidarität im Wohlfahrtsstaat. Wiesbaden: VS Verlag. Isengard, Bettina; Szydlik, Marc 2012: Living Apart (or) Together? Coresidence of Elderly Parents and Their Adult Children in Europe. In: Research on Aging 34,4: 449-474 [doi: 10.1177/0164027511428455]. Isengard, Bettina; Szydlik, Marc 2010: Koresidenz von Eltern und Kindern im Span- nungsfeld von Norm und gesellschaftlicher Herausforderung. In: Soeffner, Hans-Ge- org (Eds.): Unsichere Zeiten. Herausforderungen gesellschaftlicher Transformationen. Verhandlungen des 34. Kongresses der Deutschen Gesellschaft für Soziologie in Jena 2008. Wiesbaden: VS Verlag für Sozialwissenschaften. CD-ROM. Kalmijn, Matthijs 2006: Educational inequality and family relationships: Infl uences on contact and proximity. In: European Sociological Review 22: 1-16 [doi: 10.1093/esr/ jci036]. Keune, Maarten 2009: Mittel- und osteuropäische Wohlfahrtsstaaten im Vergleich: Ty- pen und Leistungsfähigkeit. In: Klenner, Christina; Leiber, Simone (Eds.): Wohlfahrts- staaten und Geschlechterungleichheit in Mittel- und Osteuropa: Kontinuität und post- sozialistische Transformation in den EU-Mitgliedsstaaten. Wiesbaden: VS Verlag für Sozialwissenschaften: 59-84. Kiernan, Kathleen 1999: Cohabitation in Western Europe. In: Population Trends 96: 25-32. Kohli, Martin; Künemund, Harald; Motel, Andreas; Szydlik, Marc 1997: Generationen- konstellationen, Haushaltsstrukturen und Wohnentfernungen in der zweiten Lebens- hälfte – Erste Befunde des Alterssurveys. In: Becker, Rolf (Eds.): Generationen und so- zialer Wandel. Generationsdynamik, Generationenbeziehungen und Differenzierung von Generationen. Opladen: Leske + Budrich: 157-175. Lauterbach, Wolfgang 1998: Die Multilokalität später Familienphasen – Zur räumlichen Nähe und Ferne der Generationen. In: Zeitschrift für Soziologie 27: 113-132. • Bettina Isengard260 Lauterbach, Wolfgang; Pillemer, Karl 2001: Social Structure and the Family: A United States–Germany Comparison of Residential Proximity Between Parents and Adult Children. In: Journal of Family Research 13: 68–88. Lawton, Leora; Silverstein, Merill; Bengtson, Vern L. 1994: Affection, Social Contact, and Geographic Distance Between Adult Children and Their Parents. In: Journal of Marriage and the Family 56: 57-68. Leibfried, Stephan 1992: Towards a European Welfare State? On Integrating Poverty Re- gimes into the European Community. In: Ferge, Zsuzsa; Kolberg, Jon E. (Eds.): Social Policy in a Changing Europe. Frankfurt a.M./New York: Campus: 245-280. Leopold, Thomas; Geißler, Ferdinand; Pink, Sebastian 2011: How Far Do Children Move? Spatial Distances After Leaving the parental Home. SOEP Paper Nr. 368. Berlin: DIW [http://www.diw.de/documents/publikationen/73/diw_01.c.371351.de/diw_sp0368. pdf, September 2012]. Lin, Ge; Rogerson, Peter A. 1995: Elderly Parents and the Geographic Availability of Their Adult Children. In: Research on Aging 17: 303-331 [doi: 10.1177/0164027595173004]. Madigan, Timothy J.; Hogan, Dennis P. 1991: Kin Access and Residential Mobility among Young Mothers. In: Social Science Quarterly 72: 615-622. Michielin, Francesca; Mulder, Clara H. 2007: Geographical distances between adult chil- dren and their parents in the Netherlands. In: Demographic Research 17: 655–678 [doi: 10.4054/DemRes.2007.17.22]. Mulder, Clara H. 2007: The family context and residential choice: A challenge for new research. In: Population, Space and Place 13: 265–278 [doi: 10.1002/psp.456]. Neyer, Gerda R. 2003: Family Policies and Low Fertility in Western Europe. Max Planck Institute for Demographic Research Working Paper No. 2003-021. Rostock: MPIDR [http://www.demogr.mpg.de/papers/working/wp-2003-021.pdf, September 2012]. OECD 1999: Classifying Educational Programmes: Manual for ISCED-97 Implementation in OECD Countries. OECD: Paris. Petersen, Mitchell A. 2009: Estimating Standard Errors in Finance Panel Data Sets: Com- paring Approaches. The Review of Financial Studies 22: 435-480 [doi: 10.1093/rfs/ hhn053]. Pettersson, Anna; Malmberg, Gunnar 2009: Adult children and elderly parents as mobil- ity attractions in Sweden. In: Population, Space and Place 15: 343-357 [doi: 10.1002/ psp.558]. Reher, David Sven 1998: Family Ties in Western Europe: Persistent Contrasts. In: Popu- lation and Development Review 24: 203-234. Rogerson, Peter A.; Burr, Jeffery A.; Lin, Ge 1997: Changes in Geographic Proxim- ity Between Parents and Their Adult Children. In: International Journal of Popula- tion Geography 3: 121-136 [doi: 10.1002/(SICI)1099-1220(199706)3:2<121::AID- IJPG60>3.0.CO;2-I]. Rossi, Peter H.; Rossi, Alice S. 1990: Of Human Bonding: Parent-Child Relations across the Life Course. New York: Aldine de Gruyter. Schimpl-Neimanns, Bernhard 2000: Soziale Herkunft und Bildungsbeteiligung. Empiri- sche Analysen zu herkunftsspezifi schen Bildungsungleichheiten zwischen 1950 und 1989. In: Kölner Zeitschrift für Soziologie und Sozialpsychologie 52: 636-669. Sciortino, Giuseppe 2004: Immigration in a Mediterranean Welfare State: The Italian Experience in Comparative Perspective. In: Journal of Comparative Policy Analysis 6: 111-129 [doi: 10.1080/1387698042000273442]. Living Distances between Parents and their Adult Children in Europe • 261 Shelton, Nicola; Grundy, Emily 2000: Proximity of Adult Children to Their Parents in Great Britain. In: International Journal of Population Geography 6: 181-195 [doi: 10.1002/1099-1220(200005/06)6:3<181::AID-IJPG181>3.0.CO;2-U]. Silverstein, Merril 1995: Stability and Change in Temporal Distance between the Elderly and Their Children. In: Demography 32: 29-45. Smits, Annika; Van Gaalen, Ruben I.; Mulder, Clara H. 2010: Parent–Child Coresidence: Who Moves in With Whom and for Whose Needs? In: Journal of Marriage and Family 72: 1022-1033 [doi: 10.1111/j.1741-3737.2010.00746.x]. Szydlik, Marc 2000: Lebenslange Solidarität? Generationenbeziehungen zwischen er- wachsenen Kindern und Eltern. Opladen: Leske + Budrich. UNDP (United Nations Development Programme) 2006: Human Development Report 2006. New York: New York [http://hdr.undp.org/en/media/HDR06-complete.pdf, Sep- tember 2012]. Weltalmanach 2006: Der Fischer Weltalmanach 2006. Zahlen, Daten, Fakten. Frankfurt a.M.: Fischer Taschenbuch Verlag. White, Lynn 1994: Coresidence and Leaving Home: Young Adults and Their Parents. In: Annual Review of Sociology 20: 81-102. Translated from the original text by the Federal Institute for Population Research, for information only. The reviewed and author’s authorised original article in German is available under the title “Der Apfel lebt nicht weit vom Stamm: Wohnentfernungen zwischen Eltern und ihren erwach- senen Kindern in Europa”, DOI 10.4232/10.CPoS-2013-09de or URN urn:nbn:de:bib-cpos-2013- 09de0, at http://www.comparativepopulationstudies.de. Date of submission: 12.09.2011 Date of Acceptance: 01.05.2012 Dr. Bettina Isengard ( ). University of Zurich, Institute of Sociology. Zürich. Switzerland. E-Mail: isengard@soziologie.uzh.ch URL: http://www.suz.uzh.ch/isengard.html © Federal Institute for Population Research 2013 – All rights reserved Published by / Herausgegeben von Prof. Dr. Norbert F. Schneider Federal Institute for Population Research D-65180 Wiesbaden / Germany Managing Editor / Verantwortlicher Redakteur Frank Swiaczny Assistant Managing Editor / Stellvertretende Redakteurin Katrin Schiefer Language & Copy Editor (English) / Lektorat & Übersetzungen (englisch) Amelie Franke Copy Editor (German) / Lektorat (deutsch) Dr. Evelyn Grünheid Layout / Satz Beatriz Feiler-Fuchs E-mail: cpos@bib.bund.de Scientifi c Advisory Board / Wissenschaftlicher Beirat Jürgen Dorbritz (Wiesbaden) Paul Gans (Mannheim) Johannes Huinink (Bremen) Marc Luy (Wien) Clara H. Mulder (Groningen) Notburga Ott (Bochum) Peter Preisendörfer (Mainz) Board of Reviewers / Gutachterbeirat Martin Abraham (Erlangen) Laura Bernardi (Lausanne) Hansjörg Bucher (Bonn) Claudia Diehl (Göttingen) Andreas Diekmann (Zürich) Gabriele Doblhammer-Reiter (Rostock) Henriette Engelhardt-Wölfl er (Bamberg) E.-Jürgen Flöthmann (Bielefeld) Alexia Fürnkranz-Prskawetz (Wien) Beat Fux (Zürich) Joshua Goldstein (Rostock) Karsten Hank (Köln) Sonja Haug (Regensburg) Franz-Josef Kemper (Berlin) † Michaela Kreyenfeld (Rostock) Aart C. Liefbroer (Den Haag) Kurt Lüscher (Konstanz) Dimiter Philipov (Wien) Tomáš Sobotka (Wien) Heike Trappe (Rostock) Comparative Population Studies – Zeitschrift für Bevölkerungswissenschaft www.comparativepopulationstudies.de ISSN: 1869-8980 (Print) – 1869-8999 (Internet)