Transfers of time and money among elderly Europeans and their … · 2005. 9. 29. · Marco Albertini, Martin Kohli and Claudia Vogel (European University Institute, Florence, and

1

�

�

�

Transfers of time and money among elderly Europeans and their children:

The impact of welfare regimes

Marco Albertini, Martin Kohli and Claudia Vogel

(European University Institute, Florence, and Research Group on Aging and the Life Corse, Free University of Berlin)

Paper prepared for the 1st SHARE-ELSA-HRS User Conference

26-28 September 2005 Lund, Sweden

Preliminary draft version Please do not quote without permission

Comments welcome

��

2

��

Social support and financial transfers between adult generations in the family are an

important part of the intergenerational link in contemporary societies. They are critical not only for

the family as such – how the family distributes its resources among and assures the well-being of its

members – but also for the broader issues of welfare policy, social inequality, and social integration.

We argue that the frequency and intensity of intergenerational social support and financial

transfers is influenced by factors both at the macro and at the micro level. Macro-level factors

comprise the publicly regulated benefits and obligations, such as pension coverage and levels,

public care provisions to the elderly, legal obligations of intergenerational support, or legal

regulation and taxation of bequests. Their patterns have been addressed in terms of welfare and

family regimes. Micro-level factors comprise the relevant structural, institutional and cultural

attributes of individuals and families.

In this paper we analyze the frequency and intensity of time and money transfers between

elderly parents and their children across the ten European countries covered by SHARE. We (i) give

a brief descriptive overview of the the frequency and intensity of money and time transfers between

parents and children, (ii) examine how differences among countries are related to welfare regimes,

and (iii) open the black box of micro-level mechanisms for explaining this relation.

�

3

�

Introduction The “generational contract” is the most important and also the most contentious pillar of

contemporary welfare systems. Much of the discussion on how to reform it is still truncated,

however, by focussing on its public dimension only, especially on old-age pensions (e.g., Esping-

Andersen, 2002; Myles, 2002). The transfer of resources between generations in the family is an

essential part of the “generational contract”. The patterns of exchange of time and money between

adult family generations matter not only for the well-being of individuals and families but also for

the broader issues of social policy, social inequality and stratification, and social integration (Kohli,

2004; cf. also Attias-Donfut, 1995, 1997; Esping-Andersen, 1999, 2002; Spilerman, 2000; Szydlik

2000; Myles, 2002; Künemund et al., 2003; Kohli, Künemund & Vogel., 2005).

A key question is how “public” (state- and market-provided) and “private” (family) transfers

interact. According to the conventional story of modernization the emergence of the modern welfare

state and of the nuclear family has crowded out the traditional intergenerational family transfers.

The results of most recent studies, however, point to the opposite conclusion: (i) Substantial

transfers between adult generations in the family beyond the nuclear households still exist, with a

net downward flow of resources from the elderly to their adult descendants. (ii) These

intergenerational resource flows in the family in many ways depend on and are “crowded in” by

public support; in other words, it is the welfare state provisions that enable the family to provide

transfers (Kohli, 1999; Künemund & Rein, 1999; Attias-Donfut & Wolff, 2000; Kohli et al., 2000a,

b; Attias-Donfut et al., 2005).

Assuming that adult children have in the past financially supported their elderly parents1,

this support has indeed vanished; in this respect, the welfare state has crowded out family support

(cf. Reil-Held, 2005). But in other respects, family support has on the contrary been encouraged and

enabled by the new welfare arrangements. As an example, the financial support given by elderly

parents to their adult children is often possible only on the basis of pension incomes.

So far, research on family transfers has almost exclusively been limited to single-country

studies. Now, the availability of comparative data through SHARE opens up new research avenues.

The first question to be asked from the comparative evidence is for similarities: To what extent are

the results found in the single-country studies valid across most or all of the countries – in other

words, to what extent do modern Western societies show a common transfer pattern? The second

1 This assumption is to some extent debatable. Social historians (e.g., Ehmer, 2000) as well as evolutionary anthropologists (e.g., Low, 1998) now point out that the support for the elderly by their children has been less pronounced than conventionally thought, and that there has been a pervasive pattern of help from elderly parents to their adult offspring.

4

question is for differences, which are to be expected along such dimensions as demographic and

economic structure, institutional regulation of transfers, and cultural patterns of values and beliefs.

A powerful base-line is provided here by the comparative welfare regime literature. Can we speak

of transfer regimes, and if this is the case, do they follow the typology of welfare regimes? The third

question is for the explanation of between-country differences, which requires analysis at the micro-

level. Many between-country differences may simply be due to compositional differences; in other

respects we should expect differential impacts of individual conditions such as health on transfer

patterns according to the prevailing welfare regime.

The aim of our paper is thus, first, to document cross-country similarities in the frequency

and intensity of intergenerational transfers2, second, to analyze their relationship with welfare

regimes types; and third, to start explaining this relationship by exploring the micro foundations of

the cross-regime differences observed at the macro level.

A theoretical model of intergenerational solidarity

In the sociological literature a range of factors that would affect intergenerational family

transfers and their national patterns have been put forth, for example, the demographic structure of

families, the educational and occupational status balance between the sexes and the generations, the

legal obligations of intergenerational support, the legal regulation of gift and inheritance taxation,

the instruments of family policy, the values and traditions that make up a family culture.

Differences in the specific set of factors examined are often closely correlated to the theoretical

approach adopted. Thus, behind the different combinations of explanatory factors used in the

literature, there is also a large variation in the theoretical models of intergenerational solidarity

(Kohli, 2004; Szydlik, 2004). Furthermore, explanations of intergenerational family transfers differ

according to which specific characteristics of the exchange are considered – e.g., qualitative versus

quantitative dimensions, social support and help in kind versus financial transfers.

This complexity can however be reduced by taking account of different levels of

aggregation and different categories of explanatory factors. For the sake of simplicity, we

distinguish here between the micro (individual and family) and the macro (anything above) level

only, and between three broad categories: structural, institutional and cultural factors 3 (Kohli, 2004;

2 We refer throughout to the two recent studies on this topic that are also based on SHARE data (Attias-Donfut et al., 2005; Kohli, Künemund & Vogel, 2005). 3 These distincitions are not the only ones possible. If we conceive of an “institution” as that set of rules, values, norms and practices which regulate social relations and behaviors of (groups of) individuals (Gallino, 1978), institutional factors could be viewed as a combination of the other two categories.

5

see Figure 1 for some examples). It is clear that, as in all such schemes, these groupings of factors

are somewhat arbitrary. Thus, for example, in explanations of intergenerational transfers one

important structural factor at the micro level is gender. However, why this factor matters in the

exchange process and why its relevance varies from place to place depends on the values and

traditions of individuals and communities. Of particular importance for comparative research is how

the macro factors are interrelated and packaged into patterns such as welfare regimes. The latter

term often refers to a specific class of institutional factors but can also be used more broadly to

include structural and cultural factors as well.

[Figure 1 about here]

Differences and changes in intergenerational family transfers need to be explained both (i)

by differences and changes in institutional, cultural and structural factors at the macro and micro

level and (ii) by differences and changes in how these factors are connected to intergenerational

exchange. Thus, for example, cross-country differences in the frequency and intensity of social

support may result from differences in the distribution of certain institutional, cultural or structural

factors (e.g. in country A there are more unhealthy elderly people than in country B); but they may

also result from differences in the way in which these characteristics affect intergenerational

transfers (e.g. in country A the effect of parents’ health on support received is lower than in country

B because the level of public service provision is higher)4.

Data and variables

Our empirical analyses are based on the first wave (Release 1) of the Survey of Health,

Ageing and Retirement in Europe (SHARE). SHARE is a longitudinal, multidisciplinary and cross-

national survey representing the population of individuals aged 50 and over in Europe5. The first

wave of SHARE took place in 2004 with ten participating countries: Austria, Denmark, France,

Germany, Greece, Italy, the Netherlands, Spain, Sweden and Switzerland. The number of

interviewed individuals for Release 1 is 22777.

SHARE contains detailed information on social support and financial transfers from the

perspective of adults aged 50 and above. In this paper we consider the following information: (i)

frequency and amount of financial or material gifts or support (other than for shared housing and

4 In the following part of the article we will sometimes refer to the first type of explanation as “compositional effect”. 5 On the methodological aspects of the survey see Börsch-Supan et al. (2005), Chapter 7. Regarding data quality Börsch-Supan reports that “SHARE represents the population of individuals aged 50 and over in Europe well […]. Comparisons with three prominent other European surveys, the quarterly European Union Labour Force Survey (EU-LFS), the European Community Household Panel (ECHP) and the European Social Survey (ESS) show that the SHARE data produces very similar distributions of key concepts such as employment, income, education and health” (2005: 21).

6

food) of at least 250 EUR (or the equivalent in local currency) from/to someone within or outside

the household; (ii) frequency and amount (in hours) of social support given or received in any of

three forms – personal care, practical household help and help with paperwork – from/to someone

outside the household,; (iii) frequency and amount (in hours) of looking after grandchildren., All

these questions refer to the twelve months prior to the interview. For couples, questions on financial

transfers, both given and received, and on social support received are addressed to only one member

who is asked to answer for both members of the couple (“Have you or your spouse/partner…”),

whereas questions on social support given are addressed to all individuals included in the sample.

There is one exception: where the two members of the couple are not informed about each other’s

finances, information on financial transfers is asked from both of them individually.

The SHARE transfer data refer to exchanges between respondents and all other individuals

(parents, children, grandparents, grandchildren, friends, neighbours, etc.): In the present paper we

restrict ourselves to the exchange between respondents and their children (both their own and those

of their current spouses/partners). As to financial transfers, we focus on those inter vivos, and

exclude bequests.

As a consequence of the peculiar forms of SHARE data collection on social support and

financial transfers, the construction of our dependent variables is rather complex. Here is a brief

description:

Financial transfers received/given from/to children: The SHARE questionnaire asks the

respondents whether they have received or given financial transfers of at least 250 EUR (or the

equivalent in local currency), in the prior twelve months, to someone inside or outside the

household, and then, information is asked on the type and amount of transfers received from/given

to a maximum of three different persons. As mentioned above, in a couple usually only one of the

two partners is asked these questions. In a minority of cases where finances are managed separately,

both partners are asked individually. The variable is set equal to 1 if the respondent or his/her

partner – in case of separate finances: the respondent only – have received/given such a transfer and

if one of the three transfer givers/receivers he/she mentions is his/her child.

Social support received/given from/to children: SHARE asks the respondents whether

they have received or given social support, in the prior twelve months, to someone outside the

household, and then, information is asked on the type and amount of help received from/ given to a

maximum of three different persons. It should also be remembered that the information on receiving

help is asked from only one of the two partners of a couple. Thus, in the following analyses, the

variable “received social support” is equal to 1 if the respondent has positively answered the

question and if one of the three helpers he/she mentions is his/her child. The same value of the

7

variable is imputed to the non-responding partner. An analogous process is repeated for the variable

“given social support”; however, since the two members of the couple respond separately to this

question the partners’ values on this variable are independent.

Social support given to children including looking after grandchildren: SHARE asks the

respondents whether they have, occasionally or regularly, looked after any of their grandchildren

without the presence of the parents. This value, then, is set to 1 if respondents give a positive

answer to this question or if the variable “given social to support to children” is positive.

Amount of financial transfers given/received: When asked about the amount, some

respondents refuse to answer or encounter difficulties in remembering exact amounts. In these

cases, a series of brackets unfolds, with the respondent being asked whether the amount is similar

to, lower or higher than or in-between these bracket values. This information can be used to reduce

the number of missing values. We proceed as follows: When the answer is “about” one of the

proposed values we impute the same amount; if the answer lies between the low and the mid or the

mid and the high entry points we impute the average of the two points; if the answer is “less than

the low entry point” we impute an amount equal to ½ of the entry point; if the answer is “more than

the high entry point” we impute 1.5 of the point. After this transformation, we sum up all transfers

given/received to/from any child. For the majority of cases where only one of the two partners of a

couple is asked this question on behalf of both partners, each of the two partners is imputed with

half the value of the sum across all children.

Amount of social support given/received: For each person to/from whom the respondent

has given/received social support information is collected on the frequency of this help and its

average amount in hours. Respondents can give/receive help “almost daily”, “almost every week”,

“almost every month” and “less often”. Depending on the answer to this question respondents are

asked about the number of hours on a “typical day/week/month” or “in the last twelve months”. We

express this value as the number of hours of support in the last twelve months, and sum it up for all

children involved in the exchange. Thus, for example, if help is provided on a weekly basis and the

average number of hours in a typical week is 2, we multiply by 52 and get a value of 104 hours per

year. For couples – where in all cases only one of the two partners of a couple is asked the question

on support received but on behalf of both partners – each of the two partners is imputed with half

the value of the sum across all children. All individual values have been top coded at a maximum

value of 8760 hours per year (24 hours a day).

8

From the elderly to their offspring: The downward flow of resources at the European level

As shown above, previous scientific literature has found much evidence, in a number of

different countries, of a net downward flow of resources from the elderly to their children, in the

opposite direction of the public transfers through the old-age security system. SHARE data show

that this is indeed a general pattern, both for inter vivos financial transfers and for social support

(Table 1). Resource transfers from parents to children are much more frequent and usually also

more intense than those from children to parents. In the ten European countries considered here, 21

percent (with a confidence interval at the .05 level of 20-22 percent) of the respondents have given

financial support to their children in the previous twelve months, while only 3 percent of the

respondents (c.i. 2-3) has received financial support. For social support, the picture at first sight is

somewhat different, with 9 percent having given and 16 percent (c.i. 15-17) having received such

support. But if looking after grandchildren – which can be critical for young mothers’ labor force

participation and thus for their ability to combine parenthood and gainful work – is included, the

downward direction of help is reaffirmed: 46 percent (c.i. 45-48) of elderly parents have given help

to their offspring. The average intensity of the help provided by parents to their children is also

higher than the opposite flow: 602 hours of social support per year received (c.i. 468-736) versus

902 hours per year given (c.i. 841-964), and 1470 � of financial transfers received (c.i. 948-1992)

versus 2914 � given (c.i. 2584-3244).

[Table 1 about here]

The story varies to some extent with age. SHARE’s age range of 50 years or more comprises

several distinct life phases,. There are people in the later years of their working careers, most of

them in good health and living with a partner, with some of them having their early adult children

still at home. There are those in the transition to or the first years of retirement, with a large

majority also still able to live their lives independently. Finally, there is the older group, where

among women the majority live without spouse or partner but most of them have children and

grandchildren not far away (cf. Kohli, Künemund & Lüdicke, 2005); many of this group have

increasing health difficulties Patterns of intergenerational transfers reflect these different situations.

While in the youngest group (50-59 years) only 7 percent of respondents receive social support (c.i.

6-9), among those aged 70 years or more this proportion increases to 28 percent (c.i. 26-30). Social

support given decreases from 12 percent among those up to 59 years to 6 percent among the 70+,

and if looking after grandchildren is included, from 62 to 29 percent, which means that in the latter

view even among those aged 70+ receiving and giving occur at equal frequency. With financial

transfers, there is a net downward flow for all three age groups, even though it is less marked

among the older ones. Regarding the amounts of support and the balance between receiving and

9

giving, the results are similar. While the oldest individuals tend to give fewer hours of social

support than the two other age groups, there are no significant differences in the average balance of

financial transfers. The SHARE results thus confirm that there is a net downward flow of resources

from parents to their adult offspring. It is most pronounced among the youngest group, but even for

the oldest group the balance remains equal or even somewhat positive.

Is there evidence of transfer regimes? And are they related to welfare regimes?

Structural, institutional and cultural factors do not vary independently among countries; they

tend to occur in packages. Is it possible to identify a small number of such combinations of the

factors that regulate intergenerational family transfers, in other words are there different transfer

regimes? As we have shown elsewhere, no studies exist so far that have done for the transfer field

what the comparative welfare state literature has achieved with the creation of broadly accepted

typologies of welfare regimes (Kohli, 2004: 272).

In line with welfare state research (Esping-Andersen, 1990, 1999), the first step for the

identification of transfer regimes would be to study the national packages of structural features,

social institutions and cultural values which regulate intergenerational exchange (input). The second

step would be to analyze the effects of these packages for the qualitative and quantitative aspects of

intergenerational transfers (output). In the present article, we will go directly to the second step and

examine the output side, i.e., the frequency and intensity of social and financial exchange between

elderly parents and their offspring. We will also analyze the correlation between these transfer

regimes and welfare regimes.

Why should transfer and welfare regimes be correlated, in other words, why should

countries usually grouped in the same welfare regime type have similar transfer patterns? Three

answers can be put forth: (i) The geography of welfare regimes as proposed by Esping-Andersen in

many respects corresponds to an geography of structures, institutions and values related to family

solidarity (Lesthaeghe, 1995; Mason & Jensen, 1995; Pinelli, 1995; Reher 1998). This

correspondence seems to be better in the case of the Anglo-Saxon (‘Liberal’) and the Nordic

(‘Social democratic’) welfare regimes, whereas the Continental European (‘Conservative’) group is

more heterogeneous. The goodness of fit of the welfare regime geography with the issues of family

solidarity can be increased by distinguishing in the Continental group those countries belonging to

the Southern European model identified by many authors (for example, Leibfried, 1992; Lessenich,

1995; Ferrera, 1996). (ii) As argued above, a number of macro level factors connected with the

specific configuration of welfare regimes are relevant for intergenerational transfers, including

those institutional regulations usually grouped together under the heading of ‘family policy’

10

(Gauthier, 1996; Kaufmann et al., 2002). Thus, for example, we would expect that intergenerational

exchange is more frequent and more intense in those countries – such as Austria, Germany, Italy

and Spain - where there is a legal obligation of intergenerational support, i.e., a prescription that

parents (or children) be responsible for their children (or parents) in case of need, so that social

assistance will not be granted even to adults if their parents or children can support them. The

correlation between Esping-Andersen’s (modified) welfare regime geography and that of family

policies is only very partial, however; it is especially again the Continental European group that is

marked by rather diverse family policy regimes. (iii) Welfare regimes can have an impact on the

relation between structural characteristics at the micro level and exchange behavior. As an example,

if we examine the relationship between the health status of an individual and the likelihood of

receiving social support and financial help from his/her children or parents, we would expect that,

everything else being equal, the weaker the public provision of care and health care the stronger this

relation. In other words, we would expect that the role of an individual’s need and opportunity

factors in explaining intergenerational transfers is stronger where welfare state provisions are

weaker.

The range of countries in SHARE does not allow for a full test of the welfare regime impact.

There are no countries fully belonging to the liberal cluster. Table 2 gives an overview of the

transfer results for the three remaining clusters.6 It should be noted that Switzerland is usually

classified as lying somewhere between the liberal and the conservative model, and the Netherlands

between the conservative and the social democratic one.


The first result is that the general pattern found in Table 1, of a net downward flow of

resources, also applies to each country separately. The downward flow is especially clear for

financial transfers; but it is also valid for social support if looking after grandchildren is included. In

this basic sense, all countries are similar. Beyond this, there are similarities among countries

according to welfare regimes but also some differences within them.

Financial transfers between parents and children. In all countries elderly people who

receive some financial help from their children represent a small minority: the proportion varies

from less than one percent in Demark and in Switzerland to slightly more than 8 percent in Greece.

Between-country differences are slight, with the exception of Greece which is a clear outlier.

Variation across countries is larger in the average amounts of financial transfers received but

country patterns are difficult to discern.

6 Switzerland and Greece, set in italics, are not contained in Esping-Andersen’s analysis.

11

The story is different for the giving of financial help. In the Southern European group – with

the exception of Greece which is at the level of the Continental cluster – the frequency of financial

transfers from parents to children is lower: 16 percent in Italy (c.i. 14-19) and 9 percent in Spain

(c.i. 7-10)7. At first glance these two countries seem also to have higher mean transfer values (per

donor) but the differences are not significant. The legal obligation of intergenerational support does

not seem to play a consistent role here .

Social support to/from children outside the household. With regard to the balance in the

amount of support given and received to and from children, SHARE data show that the average

elderly persons are net receivers in France, Germany, Greece and Spain, whereas they are net givers

in Demark and the Netherlands8. As noted above, however, when the time devoted to looking after

grandchildren is taken into consideration, the average elderly persons are net givers in the relation

with their children in all ten countries.

The country patterns of social support received only partially follow the proposed distinction

between southern, continental and Nordic welfare regimes. For the frequency of support the

regimes seem not to matter; the proportion of those receiving help is significantly higher in

Denmark, Sweden, Austria, Germany and Greece than in France, the Netherlands, Switzerland,

Italy and Spain. The intensity of social support received, on the other hand, corresponds well with

the regime clusters. The lowest group in terms of the average number of hours of help received is

that made up of Denmark, Sweden and the Netherlands; Austria, France and Germany are in the

middle group, and Italy, Spain and Greece in the highest one9.

The correspondence between welfare regimes and intergenerational family transfers is

perfect with regard to the proportion of social support givers among elderly people, and somewhat

less so for its intensity. Denmark and Sweden are the two countries in which elderly people are

most likely to give social support to children living outside the household. This downward flow of

time resources is significantly less frequent in Continental European countries10, and least frequent

in southern Europe. As to the average yearly amount of hours given to children, the pattern is

reversed: The Southern European countries are highest, the Nordic ones lowest, and the Continental

ones somewhere in-between: while confidence intervals for France and the Netherlands overlap

7 The confidence intervals for these two countries do not overlap with any other except for Switzerland. 8 In all other cases the confidence intervals range from negative to positive values. 9 As to confidence intervals, Greece and France are somewhere between the middle and the highest group. The confidence interval for Switzerland is so large that it is impossible to clearly locate this country in one of the three groups. 10 France and Switzerland are located at the bottom of this group. In fact, confidence intervals for these two countries very slightly overlap with that for Italy.

12

with those for the Nordic countries, that for Germany overlaps with those for the Mediterranean

countries11.

Social support to children outside the household including looking after grandchildren.

When looking after grandchildren is included, the regime pattern in frequency of support almost

disappears but still holds for its intensity.

In general, our analysis of country patterns in intergenerational family transfers suggests the

existence of a north-south gradient.. Denmark and Sweden are the countries in which the exchange

of time and money is more frequent but with the lowest intensity. On the opposite side, Italy and

Spain show the lowest proportion of elderly givers and receivers, but the highest average value of

the exchange. The differences correspond to the three welfare state groups; there is thus some

evidence for a correlation between transfer and welfare regimes.

One explanation for the observed differences in social support12 is clearly to be found in the

different rates of co-residence. As shown in Table 3, in Southern Europe co-residence of elderly

parents with their children is much more widespread than in northern Europe (see also Kohli,

Künemund & Lüdicke, 2005). There are two simple ways of controlling for this compositional

effect. The first one is that of reproducing the statistics in Table 2 only for those individuals who

have children but live alone (i.e. do not co-reside neither with a partner nor an adult child); the

second one is that of recalculating all statistics regarding exchange frequency by considering

parent-child co-residence per se as an existing resource-exchange from parents to children and

viceversa. The first solution provides numbers and patterns which are not significantly different

from those shown in Table 2. The second solution, applied by Kohli, Künemund and Vogel (2005),

generates a completely different picture. in which Southern European countries have much higher

proportions of support, higher than those in the Continental and Nordic countries. These results

suggest that co-residence is the Southern European way of transferring resources from parents to

children and viceversa. This is the norm, and when it happens that an elderly parent remains alone

he/she is less likely to give or receive help than an elderly parent in the Continental or Nordic

countries. On the other hand, in the relatively few cases in which resource exchange does take place

between non co-residing parents and children it tends to be much more intense than in other

countries, thus probably resembling what in the ‘normal’ families occurs within the household. In

the Nordic countries, where intergenerational cohabitation is rare, family support tends to revolve

around separate households and to be less intense.

11 Austria’s confidence interval slightly overlaps with those for Greece and Denmark. Switzerland’s confidence interval is again so large that it overlaps with those for all other countries. 12 Regarding financial transfers it should be remembered that all of them to/from persons living inside and outside the household are taken into account in the SHARE questionnaire.

13


Many of the observed between-country differences remain problematic, however. Why are

there differences for financial transfers as well – despite the fact that transfers both within and

outside the household are taken into account? Are there other ‘compositional effects’ that, together

with co-residence, can provide better explanations? What else can explain the systematic

differences in the intensity of the exchange? An answer to this question needs to be found in the

micro-level factors related to transfer behavior. Therefore, in the next section, following an

“analytical sociology” approach we will start exploring the micro-foundations of country patterns

(Coleman, 1990; Hedstrom & Swedberg, 1998; Abell, 2003; Barbera, 2004).

Opening the black box: Explaining the relation between welfare regimes and transfer patterns

Our analysis so far has largely confirmed the results of previous single-country studies: (i)

substantial transfers between adult generations in the family beyond the nuclear household still

exist; (ii) there is a net downward flow of resources from the elderly to their adult descendants. It

has also shown that a correlation exists between welfare regimes and broad general country patterns

in the frequency and intensity of social support and financial transfers between parents and their

offspring. The correlation is such that it provides tentative evidence for the crowding in as well as

the crowding out hypothesis: In the stronger welfare states of the North parents support their

children more frequently but with less intensity.

What remains to be explored now is the explanation of these between-regimes differences.

Many of them may simply be due to compositional differences – for example in economic and/or

health conditions of the elderly population, or in patterns of household composition and family co-

residence. Others may be due to the differential impact of individual conditions – such as health

status or gender – on transfer patterns according to the prevailing welfare regime.

In the following analyses we will focus on the explanation of four of our dependent

variables introduced so far: frequency and amount of financial transfers from parents to children,

and frequency and amount of social support given to children including looking after grandchildren.

In addition to the variable “welfare regime”, we will concentrate on one group of explanatory

factors: the structural micro level determinants, and more specifically, respondents’ demographic

and socio-economic characteristics.13 We need also to consider possible compositional effects.

13 The variables used are economic status (logarithm of per capita wealth, employment status, house property), education, demographic characteristics and living arrangements (age, number of children , co-residence with at least one adult child, co-residence with a partner), health status and use of public care services (self perceived health

14

For each of the four dependent variables we will therefore test three different models. In the

first model we will estimate the effects of individuals’ characteristics on the dependent variable.

The second model introduces dummies accounting for the welfare regimes, thus checking whether,

after controlling for relevant compositional effects, between-welfare regimes differences still hold.

In the last regression model we introduce those interaction factors whose regression coefficients, in

the previous analyses, have been significant.

After having controlled the welfare regime effect for the most relevant characteristics of the

respondents, it is time to start opening the black box. If the effect remains after controlling for the

usual compositional suspects, what are the social mechanisms that explain it? We do not aim for a

complete explanation here but for clarifying some select hypothetical links among several kinds of

factors that we test by introducing interaction terms into our models. Our hypotheses are the

following:

1. The effect of individuals’ economic well-being on giving help depends on the

generosity of public welfare provisions (interaction between welfare regime and the

logarithm of per capita wealth). Previous research has shown that parental economic

well-being (with regard to income and wealth) has a strong influence on the

likelihood and the amount of transfers to children. A generous welfare system (both

in terms of financial transfers and personal care services and facilities) may free up

more individual resources for transfers. Moreover, since their well-being is less

market-dependent (commodified) people living in more generous welfare states may

be more prone to give of their private resources to their children than elderly people

living in a state with poorer public provisions.

2. The effect of health differs according to the welfare regime (interaction between

welfare regime and health). This hypothesis is analogous to the preceding one. Poor

health, generally, puts elderly people in a situation of vulnerability both in physical

and economic terms. Being frail means to be largely dependent on the support

bought on the market or provided by the family or the state. A welfare state

characterised by low public provision of personal care and health care forces people

to turn to the family or the market. People living in weaker welfare states may

therefore be more prone to keep their resources for themselves (instead or

transferring them to their children) in order to be able to buy care and health care

services on the market.

European version, dummy codifying for having been in a nursing home overnight in the 12 month before the interview or for having received professional or paid care at home), and social support received from children (dummy codifying for help received, amount of help received).

15

3. Since welfare regime geography largely corresponds to the geography of cultural

values for family arrangements and behavior ordered along a North-South gradient,

the effect of gender on transfers may change accordingly (interaction between

welfare regime and gender). As shown in previous literature gender is a relevant

explanatory factor for money and especially time transfers from parents to children.

It has also been widely shown that differences in gender roles follow a north-south

gradient14.

4. It is often assumed that intergenerational family transfers follow an exchange or

reciprocity logic (do ut des);. Given the higher cultural valuation of family solidarity

in the South, the strength of the exchange or reciprocity logic in intergenerational

transfers may be lower (interactions between welfare regime and both the dummy

codifying for social support received from children and the amount of this support).

[Tables 4 and 5 about here]

First we analyse the determinants of whether financial transfers have been given or not

(frequency). If we just consider the characteristics of the respondents Table 4 shows that being

younger, having higher wealth and education, having a partner, being male, having good health, and

still being in employment have a significant positive impact on the likelihood of providing financial

resources to offspring. There is also evidence for reciprocity: for parents who have received social

support form their children the odds of making a financial transfer to their children increase by a

factor of 1.44, holding all other variables constant. The primary aim of this paper is addressed by

the two following models. They demonstrate that the welfare regime effect exists even after

controlling for a large number of potential compositional effects: people living in Southern and in

Continental Europe are less likely to make a financial transfer than people in the Nordic countries,

coherent with what we have found in the descriptive statistics. However, contrary to what we have

assumed (Hypothesis 1), model 3 indicates that the positive effect of wealth on the likelihood of

giving financial resources to children is weaker in the Southern European countries with weak

welfare provisions than in the northern ones. This would seem to suggest that in the South other

factors such as culturally anchored values of family solidarity take precedence over socio-economic

ones.

Table 5 presents the results for the amount of financial transfers from parents to children.

The most striking results are the following: (i) While clear country and regime patterns could not be

discerned in the descriptive account (Table 3), after controlling for respondents’ characteristics

financial transfers tend to be significantly higher in the Continental and Southern European

14 For a broader discussion of motives for giving see Kohli and Künemund (2003).

16

countries than in the Nordic ones. (ii) The amount of social support received from children in the

previous twelve months is positively related to the amount of money given to children, and this

relation of exchange or reciprocity is stronger in the two Nordic countries than in the other ones

(Hypothesis 4). (iii) The strength of the relation between wealth and amount of financial transfers to

children is not significantly different across welfare regimes.

In sum, the welfare regime effect on frequency and intensity of financial transfers from

parents to children is confirmed after controlling for of the most relevant structural (socio-

economic) characteristics of respondents. Moreover, while the gender effect and the health status

effect do not differ across regimes, the impact of wealth and of amount of social support received

varies across countries, partly in line with our hypotheses.

[Table 6 and 7 about here]

In Table 6, the correlates of giving or not giving social support to children outside the

household (including looking after grandchildren) are modelled, similarly to Table 4, by

multivariate logistic regression. The likelihood that parents have given, in the previous twelve

month, some kind of social support to their children is negatively affected by age, poor health and

being still in employment (versus being retired or homemaker); on the other hand, higher wealth,

co-residence with a partner and being female have a significant positive effect. There is no

significant relation between giving and receiving social support to and from one’s children. As the

second model shows, even after controlling for respondents’ characteristics the welfare regime has

again a significant impact on the likelihood of giving help to children. In particular, there is a

marked positive effect of the Nordic welfare regime on this variable – a clear indication that family

solidarity, rather than being crowded out, is crowded in by a strong welfare state. In contrast to what

we have found for financial transfers, none of the interaction terms we have tested has a significant

effect on giving social support to offspring.

The analyses of the determinants of the amount of support to children (Table 7) yield a

number of interesting results. The first model presented in Table 7 shows that, with the exception of

wealth, all variables with a significant effect on the likelihood of giving support also have a

significant effect – and with the same direction – on the amount of support given. Thus, while age

and poor health negatively influence the intensity of giving, being female, being a pensioner or a

homemaker, and co-residing with a partner all are conditions which, everything else being equal,

lead to a larger transfer of time resources to offspring. Once again, receiving social support from

children is not significantly related to the intensity of giving support. Model 2 demonstrates that,

even after controlling for respondents’ characteristics, there exist systematic differences between

welfare regimes. Combined with the findings from our previous analyses, these results confirm that

17

in the Nordic countries transfers from parents to children are more frequent but less intense. The

third regression model suggests some explanations for these differences: (i) In the Southern

European countries there is a significant positive effect of wealth on the amount of time transfers.

(ii) The gender effect is stronger in the continental countries – but interestingly not in the southern

countries – than in the Nordic ones; so the hypothesized North-South gradient is only partially

confirmed. (iii) The effect of having received help from children – i.e., the exchange or reciprocity

effect – is now significant in two of the three cases: positive in the Nordic and negative in the

Southern European countries. (iv) Similarly, the amount of help received has a positive impact on

giving in Sweden and Denmark and a negative one in Continental Europe.

Conclusions

Our analysis of family transfers presented here largely confirms, at the European level, what

has been found in the previous literature based on single-country studies. Substantial transfers of

resources between parents and their offspring still exist and, in general, there is a net downward

flow of resources from the older to the younger generation, both by inter vivos financial transfers

and by social support. Resource transfers from the parents to their children are much more frequent

and also usually much more intense than those in the opposite direction. The positive balance

decreases with age but even those beyond age 70 clearly remain net givers.

Our results also demonstrate the existence of consistent country patterns in the frequency

and intensity of intergenerational family transfers. Resource transfers from parents to children are

less frequent but more intense in the Southern European countries than in the Nordic ones, with the

Continental European countries being somewhere in-between these two “extremes”. This welfare

regime effect still holds after controlling for the most relevant individual characteristics of the

parents.

Our regression models are also a first tentative to open the black box of cross-country

differences, and to identify the micro level mechanisms explaining the macro level findings. The

results indicate that (i) the positive impact of wealth on the likelihood of making a financial transfer

is lower in Southern Europe than in the other countries; (ii) the positive relation between amount of

social support received from children and amount of money transferred to them is stronger in the

Nordic countries; (iii) the relation between parental economic well-being and the amount of social

support given to children is negative in the Nordic countries and positive in the Southern ones; (iv)

the correlation between receiving social support from children and amount of help given to them –

in other words, the exchange or reciprocity logic – is positive in the Nordic countries, whereas it

18

goes in the opposite direction in the Southern European ones; (v) the amount of help received from

children has a negative impact on the amount of help given in Continental Europe, whereas the

opposite is true for Northern Europe. These results, partly in line with our hypotheses, need

additional investigations in order to be fully interpreted, but they give us a first indication of where

the micro-foundations of cross-country and cross-regime differences may be found.

19

Figure 1: Exemplary factors affecting intergenerational family transfers and their patterns across countries

Macro level (above family) Micro level (family, dyads and

individuals)

Structural factors Demographic structure of families and

households; labor force structure;

income and wealth distribution

Family and household composition;

educational and occupational status of

parents and children; income and

wealth status

Institutional

factors

Legal obligations of intergenerational

support; gift and inheritance taxation;

family and social security policies

Marriage/cohabitation arrangements;

household division of labor

Cultural factors Religious traditions; family and gender

values; age and generation values

Values, beliefs, attitudes and cultural

practices of families, parents and

children

�

20

Table 1: Frequency and intensity of intergenerational transfers by age group

� ��

��

��

��

��

��

��

��

��

� ��

��

��

��

��

��

��

�

��

��

��

�

��

��

��

�

��

��

��

��

��

��

��

�

!��

��

�"��

!��

��

��

�

�#��

!��

��

�$��

Age group Up to 59 7 425 12 651 62 850 1 1672 28 2642 244 691 2420 60-69 11 265 11 536 60 933 3 1605 22 3172 145 747 2568 70 + 28 751 6 504 29 912 4 1282 13 3161 - 497 83 2229 TOTAL 16 602 9 569 46 902 3 1470 21 2914 - 171 469 2423 ��

��

��

��"��

��

��#��

��

��

�$��

��

��

21

Table 2: Frequency and intensity of intergenerational transfers by country

� ��

��

��

��

��

��

��

��

��

��

� ��

��

��

��

��

��

��

�

��

��

��

�

��

��

��

�

��

��

��

��

��

��

��

�

!��

��

�"��

!��

��

��

�

�#��

!��

��

�$��

Nordic Regimes Denmark 20 135 20 247 60 382 ,9 1331 28 2858 67 264 2718 Sweden 17 182 17 218 52 388 1,2 745 32 1391 15 274 1300 Continental European Regimes Austria 21 320 11 405 45 820 3,9 740 25 2754 -65 516 2245 France 12 486 9 285 50 742 1,3 2956 22 3362 -124 447 2895 Germany 23 457 12 508 44 689 3,6 882 27 2203 -153 250 1879 Netherlands 13 134 13 298 59 471 ,9 1139 22 2854 91 381 2603 Switzerland 12 360 11 535 43 673 1,3 1584 21 9788 87 473 8981 Southern European Regimes Greece 24 596 7 676 45 1647 8,3 1072 25 2509 -298 820 1630 Italy 12 1498 7 1449 44 1443 1,6 3230 16 3436 -451 766 2816 Spain 12 829 3 894 40 1338 2,9 1169 9 3493 -351 699 2160 ��

��

��

��"��

��

��#��

��

��

�$��

��

��

22

Table 3: Co-residence of elderly people with children and partners by country

� ��

�

��

��

�

��

��

Nordic Regimes�

%

�� &� ��

'(�

� )� �#�

Continental European Regimes�

�� *� ��

+�� *� �&�

��,� �"� ��

-��.�� &� �*�

��

Southern European Regimes�

��

/��,� �)� �$�

'�� "0� �$�

�

�

23

Table 4: Frequency of financial transfers to children (Logistic regression, odds ratios) Model 1 Model 2 Model 3 Age 0.972** 0.973** 0.973** Sex (Ref.: Men) Women 0.864 0.844* 0.848* Family status (Ref.: Not living in couple) Living in couple 1.218* 1.217* 1.208* Number of children 1.025 1.009 1.009 Children in HH (Ref.: No) Yes 0.872 0.991 0.984 Wealth (log) 1.151** 1.160** 1.254** Housing status (Ref.: Tenant) Owner 1.017 1.011 1.012 Education (Ref.: No education) ISCED 1 1.541* 1.582* 1.579* ISCED 2 2.053** 1.924** 1.945** ISCED 3 3.194** 2.824** 2.860** ISCED 4 3.271** 2.896** 2.863** ISCED 5 4.541** 3.956** 3.985** ISCED 6 6.488** 6.080** 6.057** Other education 3.855** 3.323** 3.372** Employment status (Ref.: Employed) Retired 0.910 0.926 0.921 Unemployed 0.664* 0.662* 0.667* Sick or disabled 0.750 0.749 0.753 Homemaker 0.712** 0.777 0.769* Other status 0.746 0.759 0.754 Health status (Ref.: Good and better) Less than good health 0.761** 0.773** 0.776** Help from care services (Ref.: No help) Received help from care services 1.084 1.027 1.036 Help from children (Ref.: No help) Received help from children 1.437** 1.366** 1.370** Amount of help from children 1.000 1.000 1.000 Welfare regime (Ref.: Nordic) Continental European regime 0.720** 1.334 Southern European regime 0.507** 2.804 Interaction terms Wealth (log)*Continental regime 0.946 Wealth (log)*Southern regime 0.858* * significant at 5%; ** significant at 1%

24

Table 5: Amount of financial transfers to children (OLS regression, coefficients) Model 1 Model 2 Model 3 Age 0.001 0.001 -0.000 Sex (Ref.: Men) Women -0.082 -0.072 -0.067 Family status (Ref.: Not living in couple) Living in couple -0.403** -0.399** -0.393** Number of children -0.004 0.002 0.007 Children in HH (Ref.: No) Yes 0.173* 0.133 0.128 Wealth (log) 0.139** 0.138** 0.141** Housing status (Ref.: Tenant) Owner 0.059 0.061 0.062 Education (Ref.: No education) ISCED 1 -0.163 -0.170 -0.172 ISCED 2 -0.197 -0.177 -0.200 ISCED 3 -0.208 -0.176 -0.182 ISCED 4 -0.340 -0.290 -0.301 ISCED 5 0.159 0.194 0.188 ISCED 6 0.411 0.408 0.403 Other education 0.627* 0.646** 0.642** Employment status (Ref.: Employed) Retired -0.056 -0.070 -0.060 Unemployed 0.004 0.012 0.019 Sick or disabled -0.164 -0.158 -0.137 Homemaker 0.081 0.044 0.043 Other status 0.409 0.394 0.400 Health status (Ref.: Good and better) Less than good health -0.063 -0.068 -0.071 Help from care services (Ref.: No help) Received help from care services 0.061 0.082 0.045 Help from children (Ref.: No help) Received help from children -0.086 -0.067 -0.096 Amount of help from children (*100) 0.021** 0.020** 0.095** Welfare regime (Ref.: Nordic) Continental European regime 0.154* 0.160* Southern European regime 0.257* 0.276** Interaction terms Amount of help from children (*100)*Continental regime

-0.100*

Amount of help from children (*100)*Southern regime

-0.100**

Constant 5.777** 5.596** 5.605** R-square 0.08 0.08 0.08 * significant at 5%; ** significant at 1%

25

Table 6: Frequency of social support to children including looking after grandchildren (Logistic regression, odds ratios) Model 1 Model 2 Age 0.916** 0.916** Sex (Ref.: Men) Women 1.383** 1.371** Family status (Ref.: Not living in couple) Living in couple 1.797** 1.803** Number of children 1.034 1.030 Children in HH (Ref.: No) Yes 0.990 1.025 Wealth (log) 1.088** 1.091** Housing status (Ref.: Tenant) Owner 1.004 1.004 Education (Ref.: No education) ISCED 1 0.949 0.935 ISCED 2 0.987 0.959 ISCED 3 1.154 1.119 ISCED 4 1.443 1.296 ISCED 5 1.138 1.090 ISCED 6 0.936 0.929 Other education 0.602 0.574 Employment status (Ref.: Employed) Retired 1.841** 1.880** Unemployed 0.877 0.894 Sick or disabled 1.271 1.296 Homemaker 1.648** 1.734** Other status 1.728* 1.777* Health status (Ref.: Good and better) Less than good health 0.675** 0.682** Help from care services (Ref.: No help) Received help from care services 0.711* 0.703** Help from children (Ref.: No help) Received help from children 1.116 1.106 Amount of help from children 1.000 1.000 Welfare regime (Ref.: Nordic) Continental European regime 0.689** Southern European regime 0.657** * significant at 5%; ** significant at 1%

26

Table 7: Amount of social support to children including looking after grandchildren (OLS regression, coefficients) Model 1 Model 2 Model 3 Age -16.328** -17.406** -17.410** Sex (Ref.: Men) Women 109.776** 129.077** 57.268 Family status (Ref.: Not living in couple) Living in couple 173.692** 172.562** 174.352** Number of children -3.378 2.222 0.737 Children in HH (Ref.: No) Yes 95.305 25.484 25.379 Wealth (log) 7.201 1.950 -49.364* Housing status (Ref.: Tenant) Owner -6.993 -2.108 -3.438 Education (Ref.: No education) ISCED 1 9.976 -15.654 -24.854 ISCED 2 -23.328 34.884 18.402 ISCED 3 -52.541 41.745 35.757 ISCED 4 495.738 571.556 578.671 ISCED 5 -106.530 -7.344 -7.449 ISCED 6 -206.461 -141.436 -157.508 Still in education -347.310** -427.779** -480.452** Other education -128.487 -32.309 -42.437 Employment status (Ref.: Employed) Retired 225.231** 224.865** 221.372** Unemployed -97.435 -94.407 -91.861 Sick or disabled 127.006 107.639 113.133 Homemaker 210.277** 153.698* 175.032** Other status 246.785 231.674 249.043 Health status ( Ref.: Good and better) Less than good health -75.491* -90.329** -83.789* Help from care services (Ref.: No help) Received help from care services -97.101* -46.841 -50.553 Help from children (Ref.: No help) Received help from children -25.842 4.462 104.451* Amount of help from children (*100) 7.996 6.832 7.228* Welfare regime (Ref.: Nordic) Continental European regime 99.890** -491.927 Southern European regime 347.826** -345.590 Interaction terms Wealth (log)*Continental regime 50.351 Wealth (log)*Southern regime 67.173* Gender*Continental regime 111.122* Gender*Southern regime 8.874 Received help from children *Continental regime

-19.886

Received help from children *Southern regime

-329.697**

Amount of help from children (*100)*Continental regime

-9.10*

Amount of help from children (*100)*Southern regime

8.20

Constant 1,163.876** 1,053.053** 1,632.854** R-square 0.04 0.05 0.06 * significant at 5%; ** significant at 1%

27

References

Abell, P. 2003. "On the Prospects for a Unified Social Science: Economics and Sociology." Socio-Economic Review 1:1-26.

Attias-Donfut, C. 1995. Les Solidarités Entre Générations. Vieillesse, Familles, État . Paris: Nathan.

Attias-Donfut, C., J. Ogg, and F. C. Wolff. 2005. "European Patterns of Intergenerational Transfers." To appear in European Journal of Ageing.

Attias-Donfut, C. and F. C. Wolff. 2000. "The Redistributive Effects of Generational Transfers ." Pp. 22-46 in The Myth of Generational Conflict: The Family and the State in Ageing Societies, eds. S. Arber and C. Attias-Donfut. London: Routledge.

Barbera, F. 2004. Meccanismi Sociali: Elementi Di Sociologia Analitica. Bologna: il Mulino.

Börsch-Supan, A. 2005. "Introduction." Pp. 7-28 in Health, Ageing and Retirement in Europe. First Results From the Survey of Health, Ageing and Retirement in Europe, eds. A. Börsch-Supan, A. Brugiavini, H. Jürges, J. Mackenbach, J. Siegrist and G. Weber. Mannheim: Mannheim Research Institute for the Economic of Aging.

Börsch-Supan, A., A. Brugiavini, H. Jürges, J. Mackenbach, J. Siegrist and G. Weber (eds.). 2005. Health, Ageing and Retirement in Europe. First Results From the Survey of Health, Ageing and Retirement in Europe. Mannheim: Mannheim Research Institute for the Economics of Aging.

Coleman, J. S. 1990. Foundations of Social Theory. Cambridge Mass.: The Belknap Press of the Harvard University Press.

Ehmer, J. 2000. "Ökonomische Transfers und emotionale Bindungen in den Generationenbeziehungen des 18. und 19. Jahrhunderts." Pp. 77-96 in Generationen in Familie und Gesellschaft, eds. M. Kohli and M. Szydlik. Opladen: Leske + Budrich.

Esping-Andersen. G. 1999. Social Foundations of Post Industrial Economies. Oxford: Oxford University Press.

———. 1990. The Three Worlds of Welfare Capitalism. Cambridge: Polity Press.

———. 2002. "Towards the Good Society, Once Again? ." Pp. 1-25 in Why We Need a New Welfare State, eds. G. Esping-Andersen, D. Gallie, A. Hemerijck and J. Myles. Oxford: Oxford University Press.

Ferrera M. 1996. "Il modello Sud-Europeo di Welfare State" Rivista Italiana di Scienza Politica 1:67-101.

Gallino, L. 1978. Dizionario di Sociologia. Torino: UTET.

Gauthier, A. H. 1996. The state and the family: A comparative analysis of family policies in industrialized countries. Oxford: Clarendon Press.

Hedström, P. and R. Swedberg (eds.). 1998. Social Mechanisms. Cambridge: Cambridge University Press.

28

Jürges, H. 2005. "Methodology." Pp. 349-61 in Health, Ageing and Retirement in Europe. First Results From the Survey of Health, Ageing and Retirement in Europe, eds. A. Börsch-Supan, A. Brugiavini, H. Jürges, J. Mackenbach, J. Siegrist and G. Weber. Mannheim: Mannheim Research Institute for the Economics of Aging.

Kaufmann, F.-X., A. Kuijsten, H.-J. Schulze and K. P. Strohmeier (eds.). 2002. Family life and family policies in Europe: Problems and issues in comparative perspective. Oxford: Oxford University Press.

Kohli, M. 1997. "What Comes After Enterprise-Centered Social Protection? The Case of East Germany." Pp. 115-34 in Enterprise and Social Benefits After Communism, eds. M. Rein, B. L. Friedman and A. Wörgötter. Cambridge: Cambridge University Press.

———. 1999. "Private and Public Transfers Between Generations: Linking the Family and the State." European Societies 1:81-104.

———. 2004. "Intergenerational Transfers and Inheritance: A Comparative View." Pp. 266-89 in Intergenerational Relations Across Time and Place (Annual Review of Gerontology and Geriatrics, Vol. 24), ed. M Silverstein. New York: Springer.

Kohli, M. and H. Künemund. 2003. "Intergenerational transfers in the family: What motivates giving?" Pp. 123-142 in Global Aging and Challenges to Families, eds. V. L. Bengtson and A. Lowenstein. Hawthorne, NY: Aldine de Gruyter.

Kohli, M., H. Künemund, A. Motel and M. Szydlik. 2000. "Families Apart? Intergenerational Transfers in East and Weast Germany." Pp. 88-99 in The Myth of Generational Conflict: The Family and the State in Ageing Societies, eds. S. Arber and C. Attias-Donfut. London: Routledge.

Kohli, M., H. Künemund and J. Lüdicke. 2005. “Family structure, proximity and contact.” Pp. 164-170, 207-231 Health, Ageing and Retirement in Europe. First Results From the Survey of Health, Ageing and Retirement in Europe, eds. A. Börsch-Supan, A. Brugiavini, H. Jürges, J. Mackenbach, J. Siegrist and G. Weber. Mannheim: Mannheim Research Institute for the Economic of Aging.

Kohli, M., H. Künemund and C. Vogel. 2005. Intergenerational Family Transfers in Europe: A Comparative Analysis. Paper for the Research Network on Ageing at the 7th European Sociological Association (ESA) Conference, Torun, Sept. 9-12.

Künemund, H., A. Motel-Klingebiel and M. Kohli. 2005. "Do private intergenerational transfers increase social inequality in middle adulthood? Evidence from the German Aging Survey." Journal of Gerontology: Social Sciences, 60B: S30-S36. Künemund, H. and M. Rein. 1999. "There Is More to Receiving Than Needing: Theoretical Arguments and Empirical Explorations of Crowding in and Crowding Out." Ageing and Society 19:93-121.

Leibfried, S. 1992. "Towards a European Welfare State: On Integrating Poverty Regimes in the European Community." Pp. 245-280 in Social Policy in a Changing Europe, eds. Z. Ferge and J. E. Kolberg. Frankfurt: Campus Verlag.

Lessenich, S. 1995. "Espana y Los Tres Mundos de Estado del Bienestar." Universitat Pompeu Fabra Working Paper no. 95/9.

29

Lesthaeghe, R. 1995. "The Second Demographic Transition in Western Countries: An Interpretation." Pp. 17-62 in Gender and Family Change in Industrialised Countries, eds. K. Oppenheim Mason and A. M. Jensen. Oxford: Clarendon Press.

Low, Bobbi S. 1998. "The Evolution of Human Life Histories." Pp. 131-161 in Handbook of Evolutionary Psychology: Ideas, Issues, and Applications, eds. C. Crawford and D. L. Krebs. Mahwah, NJ: Erlbaum.

Mason, K. O. and A. M. Jensen, (eds.) 1995. Gender and Family Change in Industrialised Countries. Oxford: Clarendon Press.

Myles, J. 2002. "A New Social Contract for the Elderly?" Pp. 130-172 in Why We Need a New Welfare State, eds. G. Esping-Andersen, D. Gallie, A. Hemerijck and J. Myles. Oxford: Oxford University Press.

Pinelli, A. 1995. "Women's Condition, Low Fertility, and Emerging Union Patterns in Europe." Pp. 82-101 in Gender and Family Change in Industrialised Countries, eds. K. Oppenheim Mason and A. M. Jensen. Oxford: Clarendon Press.

Reher, D. 1998. "Family Ties in Western Europe: Persistent Contrasts." Population and Development Review 24:203-34.

Reil-Held, A.. 2005. Crowding Out or Crowding In? Public and Private Transfers in Germany. Mannheim: Mannheim Research Institute for the Economics of Aging.

Spilerman, S. 2000. "Wealth and Stratification Processes." Annual Review of Sociology 26:497-524.

Szydlik, M. 2004. "Inheritance and Inequality: Theoretical Reasoning and Empirical Evidence." European Sociological Review 20:31-45.

———. 2000. Lebenslange Solidarität? Generationenbeziehungen zwischen erwachsenen Kindern und Eltern. Opladen: Leske+Budrich.

Documents

Transfers of time and money among elderly Europeans and their … · 2005. 9. 29. · Marco Albertini, Martin Kohli and Claudia Vogel (European University Institute, Florence, and