Instrumental support exchanges among kin and non-kin in light of personal configurations

Abstract As exchanges of instrumental support between kin and non-kin remain essential to buffer the impact of critical life events, we consider the characteristics of personal configurations that may enhance or hinder them. Personal configurations vary in terms of their composition and two aspects of their structure: density and centrality. These dimensions are investigated to uncover whether they influence the type of instrumental support being exchanged (financial, material, and care) and the likelihood of their being reciprocal. Drawing on a representative sample of individuals living in Switzerland, results show that overall instrumental support is present in all personal configurations irrespective of their composition, but financial support is more prevalent in configurations based on parent–adult child relationships. Furthermore, configuration structures characterized by density of emotional support are positively associated with giving care support and with reciprocity, while those characterized by density of conflict are associated with giving less instrumental support overall.


Introduction
The development of welfare policies in Switzerland as in most Western societies following World War II has decreased the need for family instrumental support by filling the unexpected "holes" in individuals' trajectories (e.g., following accidents, disease, unemployment, and maternity leave) [Kohli (2007), Wall et al. (2018)]. Correlatively, states have progressively taken over some tasks traditionally carried out within families (e.g., childcare, education, and elderly care). Nevertheless, exchanges of instrumental support between family members and beyond remain essential to ease family transitions, buffer the impact of critical life events, and prevent vulnerabilization [Spini et al., (2017)]. For instance, the current pandemic crisis reactivates the need for family instrumental support [Widmer et al. (2020)]. Extensive research has been devoted to resources such as instrumental support that are directly or indirectly embedded in personal relationships, which individuals may get access to and further invest in or draw on when needed [Wellman and Wortley (1990), Kadushin (2002), Van Der Gaag and Snijders (2005), Ryan et al. (2008)]. Instrumental support can be of various types such as financial (loans, money transfers), material (providing goods), or care (such as domestic help, transportation, caregiving services) [Swartz (2009)]. It is distinguished from emotional support (such as motivation, guidance, closeness) as it responds to more specific needs [Cohen (2004)].
Instrumental support is usually associated with the Western nuclear family centered on the heterosexual couples and their minor children, which has been conceptualized as an economically independent and self-reliant unit of support [Parsons (1949)]. This focus on the nuclear family has been challenged by demographic shifts (e.g., increase in the life expectancy and co-survivorship creating a larger supply of intergenerational ties) and the changes in nuclear families themselves (greater marital instability and single parenting) [Bengtson (2001), Swartz (2009)]. In this context, the support potential of a wide range of relationships is brought to light, such as instrumental support stemming from intergenerational family relationships as well as from non-kin relationships. Regarding intergenerational family relationships, research shows their latent availability under different concepts such as "family national guard" [Hagestad (1996)], "latent kin network of support" [Riley and Riley (1993)], "support bank" [Antonucci and Jackson (1989)], or "kin reservoir" [Coenen-Huther et al. (1994)]. As for non-kin relationships, research has acknowledged the role of friends and other non-related individuals in providing instrumental support [Smart (2010), Suanet and Antonucci (2017)], especially when individuals do not experience a nuclear family setting [e.g., same-sex couples, see Weeks et al. (2001); singlehood, see Bellotti (2008); shared living arrangements, see Heath et al. (2017)], experience a critical life event like an intimate relationship breakdown [Aeby and Heath (2020)], and explore career development opportunities [McDonald and Mair (2010)].
Many individuals find themselves isolated and unable to rely on family or friendship supports [McPherson et al. (2006), Nicolaisen and Thorsen (2017)]. Individuals tend to be acquainted with alters sharing similar circumstances [McPherson et al. (2001)] and, therefore, those with lower socioeconomic backgrounds may be unable to get access to the support required when needed (especially financial). Similarly, migration contexts reduce the availability for some types of instrumental support (especially care) [Lubbers et al. (2010), Wyss and Nedelcu (2020)]. Besides these factors, the explanation for the lack or, availability of instrumental support may also lay on the characteristics of the web of interdependencies that individuals are embedded in. Indeed, instrumental support exchanges between two individuals do not happen in a social vacuum, but are embedded in personal configurations [Elias (1978[Elias ( , 1983, Widmer (2010)]. These personal configurations may vary in terms of their composition and structure. Composition refers to the type of ties; for instance, whether individuals develop their personal configurations around their couple, siblings, and/or their friends ]. Structure relates to the interconnectedness of personal configurations; for instance whether individuals and their configuration members are densely or loosely connected to one another [Coleman (1988), Burt (2002)].
Moving beyond the nuclear family, this paper explores the characteristics of personal configurations that may enhance or hinder instrumental support exchanges. More precisely, we assess the impact of the composition and structure of personal configurations on instrumental support exchanges, looking at the type of instrumental support (financial, material and care) and on the likelihood of such exchanges to be reciprocal. Indeed, to understand instrumental support exchanges, one has to consider their embeddedness in a wider web of interdependencies. In the following two sections, we first explore instrumental support exchanges among kin and non-kin, and then the characteristics of personal configurations drawing on the configurational approach.
2. Instrumental support exchanges among kin and non-kin Individuals have access to a set of relationships that are likely to vary in commitment and givenness. While relationships with family members are at the heart of instrumental support [Bonvalet and Ogg (2007)], friend and work-related relationships also represent a significant source of support [Allan (2008), McDonald and Mair (2010), Conkova et al. (2018)]. There is a lasting debate in most Western societies about whether there is a process of suffusion between different sets of values bringing the roles of family and friend closer together [Pahl and Spencer (2004), Allan (2008), Wall and Gouveia (2014)]. Some values are commonly associated with the family, such as obligation and hierarchy, and others to friends, such as choice and equality. Some authors highlight the similarities between friend and family roles, with friends playing family-like roles and family members playing friend-like roles [Pahl and Spencer (2004), Wall and Gouveia (2014)]. However, other authors suggest that friends and family retain some specialized roles and that different principles of solidarity still exist [Allan (2008)]. For instance, loans and money transfers are said to be more prevalent among family members and, in particular, between parents and their adult children [Bonvalet and Ogg (2007)], which is partially explained by the persistent norm of filial responsibility [Swartz (2009)]. Correlatively, instrumental support not implying money or goods such as care support may be more evenly widespread. Finally, female family members are the most active in participating in these exchanges; they tend to provide and receive more support than men [Swartz (2009)].
In case of need and in their everyday life, most individuals get help from their personal relationships, but they are also expected to provide support to their close ones. Thus, instrumental support exchanges have a tendency to be reciprocal. Reciprocity is a key principle, ordering exchanges among individuals [Molm et al. (2007), Girardin, 2017]. Anthropologists have studied the mechanisms of gift exchange, emphasizing the act of giving creating a social bond between giver and receiver and a moral obligation to reciprocate [Mauss (2007[Mauss ( ) [1923]. This creates chains of exchanges of goods and services. These are outside the formal market and governed by culturally grounded informal norms. The principle of reciprocity implies the exchange of goods and services. Reciprocity takes place over time, and often a favor is returned much later in another form. It is mostly unspoken, but normatively expected, and individuals who fail to comply may be excluded or exclude themselves from further exchanges, especially low-income individuals [Offer (2012)]. Other exceptions to this reciprocity principle are found within the parent-child relationships as money transfers tend to go from parents to their adult children [Bonvalet and Ogg (2007)].
The structure of reciprocity depends on two structural dimensions: direct (only two actors) vs. indirect (more than two) reciprocation of benefits, and unilateral versus bilateral flow of benefits. The combination of these two further translates into three forms of exchange: direct negotiated (based on an agreement), direct reciprocal, and indirect generalized [Molm et al. (2007)]. Generalized indirect exchange produces greater social solidarity than restricted direct exchange, which is related to the issue of risk management and building trust, and to the salience of conflict in relationships [Molm et al. (2007)]. In the case of personal configurations where individuals know each other, we see chain-generalized reciprocity in contrast to pure-generalized reciprocity (unknown recipients and benefactors).

Composition and structure of personal configurations
The composition of personal configuration reflects individuals' embeddedness in society and, thus, is closely related to life-course stages, transitions, and events [Antonucci et al. (2010), Levy and Widmer (2013), Aeby et al. (2019)]. The configurational approach accounts for the interdependencies individuals develop over their life by participating in multiple daily interactions. The notion of configuration was first coined by Elias (1978Elias ( , 1983 and later adapted to the study of family and personal relationships (see Introduction chapter). Individuals have a personal configuration composed of alters they consider as important; those alters may come from the same or different sociability settings, be related or not, and know each other or not. Therefore, one may hypothesize that the instrumental support available to individuals will vary depending on the composition of their personal configurations.
The structure of personal configurations may be described as more or less interconnected. On the one hand, some configuration structures are characterized by a high density of relationships, meaning that the majority of configuration members are in direct ties with one another. When these dense relationships are supportive, the configuration produces bonding social capital that can give a sense of protection and togetherness but also of control and lack of autonomy [Coleman (1988), Burt (2002), Widmer (2010)]. When, in contrast, these dense relationships are conflictual, it means that individuals are embedded in a very tense environment in which all configuration members tend to antagonize each other on a regular basis. On the other hand, configuration structures can be characterized by a high centrality of respondents, meaning that they mediate most relationships between configuration members. If those relationships are supportive, the configuration produces bridging social capital [Coleman (1988), Burt (2002), Widmer (2010)]. In this case, individuals who serve as a bridge between configuration members benefit from their central position to have power over flows of resources [Cook and Emerson (1978), Burt (2002), Cornwell (2011)]. However, in conflictual relationships, the respondent is at the center of tensions [Widmer (2010), Widmer et al. (2018)]. Frequently, supportive and conflictual relationships co-exist in the same personal configurations, and this generates triadic ambivalence [Widmer andLüscher (2011), Sapin et al. (2016), Aeby et al. (2020)]. While kin and non-kin may provide support, the level of co-dependence between family members is overall higher, and this explains the higher level of ambivalence in relationships with partners, parents, offspring, and siblings in comparison with those with extended family members, friends, and acquaintances Pillemer (1998), Fingerman et al. (2004)]. Interestingly, even in times of financial struggle, being connected can play an efficient buffering role and contribute to subjective well-being, whereas the satisfaction of the socially isolated can "catch up" to some degree if they are doing well financially [Richards (2016)]. Therefore, one may hypothesize that instrumental support available to individuals varies depending on the structure of their personal configurations.

Summary and hypotheses
Instrumental support exchanges between two individuals are embedded in personal configurations. In this study, we explore the characteristics of personal configurations that may enhance or hinder those exchanges. Personal configurations distinguish themselves from one another according to their composition and two aspects of their structure (density and centrality). These dimensions are investigated to uncover whether they influence the type of instrumental support being exchanged (financial, material, and care) and the likelihood of such exchanges being reciprocal. More precisely, considering the type of instrumental support, we propose the following hypotheses: (1a) As instrumental support exchanges arise in situations of manifest and specific needs, exchanges of care support are more frequent than financial or material support exchanges. (1b) As loans or money transfers are more common between parents and adult children, personal configurations based on close family members are more likely than personal configurations based on non-kin to provide or be given financial or material support. (1c) As configuration structures characterized by density of emotional support create a more protective environment, they are associated with more instrumental support exchanges overall, whereas configuration structures characterized by density of conflict are associated with less instrumental support exchanges overall.
Considering reciprocity, our hypotheses are: (2a) As reciprocity is an overarching principle, ordering exchanges among individuals, it is more common than unidirectional exchanges. Nevertheless, reciprocity does not systematically take place within the same type of support (material for material), but at a more global level (care for material for instance). (2b) As reciprocity is normatively expected and failing to reciprocate may lead to exclusion, it is systematic among personal configurations based on non-kin members, as these relationships are taken less for granted. In families, reciprocity is likely to unfold over larger periods of time and to be indirect (e.g., adult children getting financial help from their parents and later on financially helping their children). (2c) As configuration structures characterized by density of emotional support create a more normatively controlling environment, they are associated with more reciprocity, whereas configuration structures characterized by density of conflict are associated with less reciprocity. was made through the Swiss Federal Statistical Office, and a survey institute conducted face-to-face interviews using a computer-assisted personal interviewing method.

Respondents
There were 755 respondents who provided valid configuration data and reported at least one configuration member. Women composed 51% of the sample; 52% of respondents belonged to the 1950-55 cohort and 82% included Swiss citizens. Respondents with vocational education added up to 64%; furthermore, 20% had a tertiary education, 10% lower secondary education, and 7% upper secondary education. Among the respondents, 48% were employed full-time (80-100%), 22% were employed part-time (less than 80%), 15% were self-employed, 7% stayed at home, and 7% were in other situations.

Methodological approach
We first estimated the distribution of dyadic instrumental support exchanges between respondents and their configuration members (support relating to overall instrumental, financial, material, and care). Next, we considered whether these were reciprocal or unidirectional with the respondents being either mainly providers or recipients of help. For both aspects, we assessed the association with the whole configuration composition and structure by means of logistic regressions, including a set of sociodemographic covariates. All computations were made using the R statistical environment (R Development Core Team, 2020).

Instrumental support in dyadic exchanges
Dyadic instrumental support exchanges between respondents and their configuration members were investigated through three types of support: financial, material, and care. Financial support encompasses lending or giving money, inheritance or donation, offering partnership or passing on a business, giving a house or land. Material support encompasses giving clothes and food, helping to buy furniture and appliances, and with housing. Care support encompasses helping with the housework, in illness, taking care of alters, errands, helping in small repairs at home, and transport of alters. Both financial and material support requires more wealth than care support, but on a quite different scale. It was asked whether the respondents received or gave any of the three kinds of support to their configuration members. Partners and children aged under 25 were excluded because such exchanges are linked to sharing a common family household. Therefore, to measure such exchanges, it was necessary to eliminate cases that did not meet this criterion. We kept all respondents who had at least one configuration member outside the nuclear family; as for those entirely focused on the nuclear family, we kept those who had at least one child aged 25 or above (n = 672).

Configuration composition
Respondents were first asked to name "the individuals who, over the past year, have been very important to you, even if you have not gotten along well with them." They could name up to 20 important alters (size ranged from 1 to 17 with a mean of 3.9 and a standard deviation of 2.15). For each of them, respondents were asked to define their relationship, choosing from an extensive list of 51 personal ties printed on a showcard.
A typology of personal configurations was built based on the different personal ties cited by the respondents to account for the diversity of arrangements. We used a standard two-step procedure combining principal components analysis and hierarchical clustering analysis [for details of the procedure see Widmer et al. (2013), Aeby et al. (2019)]. The labels of the configurations were chosen to reflect the standpoint of the focal person, i.e., the respondents in relation to their configuration members. Labels represent the main orientation of the configurations, but do not exclude the presence of other ties. We ended up with seven personal configurations, by order of importance: Female friends and children based on female friendship and children, often without a partner (25%); Nuclear centered on the partner and minor and adult children (23%); Parents including mothers and fathers (13%); Siblings including sisters and brothers (12%); Male friends based on male friendship, often including a partner in contrast to the Female friends and children configuration (11%); Kinship including collaterals, siblings-in-law and other relatives (9%); Professional and non-kin including colleagues and other non-kin individuals (7%).

Configuration structure
To describe the structure of personal configuration, respondents were asked to report emotional support relationships, received and/or given, existing among all of their configuration members. The question started with, "Among the persons you have just mentioned, who could give you emotional support if needed?" and was followed by, "And what about the first (second, third,… nth) person you mentioned? Who could give him or her emotional support if needed (yourself included)?" Regarding conflict, respondents were asked to report who, among their configuration members, could anger or bother them and/or the other configuration members. Similarly, the question started with, "Among the persons you have just mentioned, who could anger you (annoy you)?" and was followed by, "And what about the first (second, third,… nth) person you mentioned? Who could anger him or her (annoy him or her) (yourself included)?" This approach goes beyond dyadic exchanges (between respondents and their configuration members) as it considers the whole configuration holistically. Following previous validation studies [ Widmer (2010)], configuration analysis measures were applied to investigate density and centrality in personal relationships [Wasserman and Faust (1994), Borgatti et al. (2013)]. Density within the whole configuration corresponds to the number of actual connections reported by the respondent, divided by the maximum theoretical number of connections in a given configuration. The overall mean of support density was 0.62, much higher than the overall mean of conflict density (0.36). Betweenness centrality examines the extent to which an actor stands between all other actors within the same configuration [Freeman (1978)]. It is computed as the number of shortest paths between configuration members that pass through this actor, corresponding here to the respondent. The overall mean of betweenness centrality was 0.26 for support and 0.11 for conflict.

Control variables
We included several control variables in the analyses, namely sex (women or men), birth cohort (1950-1955 or 1970-1975), education (lower secondary, upper secondary, vocational, and tertiary), and citizenship (being a Swiss national or not).

Results
In this section, we first estimate the prevalence of different types of instrumental support and then assess the direction of these exchanges, examining who provides or benefits from them. In both cases, we measure the influence played out by the personal configurations in the exchanges of support.
6.1 Financial, material, and care supports in personal configurations Results show that circa 40% of the respondents (n = 271) were engaged in dyadic instrumental support exchanges with their configuration members (partners and children under 25 excluded). It implies that circa 60% of the respondents (n = 401) did not report such exchanges. Overall, a third of the respondents were involved in receiving (n = 213) and another third in giving (n = 226) at least one of the three kinds of support to their configuration members. Care support was the most common (26.2% reporting receiving care and 28% reporting giving care), whereas financial (10.3% reporting receiving care and 11.6% reporting giving care) and material support (11.5% reporting receiving care and 15% reporting giving care) were more exceptionally received and given, which confirms hypothesis 1a regarding the prevalence of care support.
We then looked at the absence or presence of instrumental support exchanges in the personal configurations across the full sample of 672 respondents. In all personal configurations, instrumental support exchanges took place, which means that no type of personal configuration was excluded from such exchanges. At the highest end, 58.2% of respondents with a Parent configuration were involved in such exchanges, whereas at the lowest end it was 22% of respondents with a Male friend configuration.
Evaluating the subsample of respondents who were involved in instrumental support exchanges (n = 271), we look closely at the distribution of the different types of support across personal configurations (see Table 1). Receiving varied between 19.2% in the Nuclear configurations and 54.1% in the Parent ones, while giving varied between 18% in Male friend configurations and 41.5% in Kinship ones. Those with the highest percentage of respondents involved in receiving financial and material support was the Parent configuration. Following the same downward direction from parents to children, the configurations with the highest percentage of respondents involved in giving financial support were the Nuclear ones. It shows the predominance of parents-adult children solidarity.
In configurations composed of non-kin ties, for instance Professional and non-kin and Female friends and children configurations, instrumental support exchanges also took place. While financial support was less common, it did happen, indicating that non-kin ties may be mobilized if needed under certain circumstances. Care support was widespread in all types of personal configurations with the exception of the Male friend ones. Getting back to Hypothesis 1b, according to which personal configurations based on family members (especially parents and adult children) are more likely than non-kin to provide or be given financial or material support, the data confirm the prevalence of such exchanges in configurations including parent and adult children. Moreover, exchanges of financial and material support were well-spread in Female friends and children configurations as in Kinship configurations.
We then performed a series of logistic regressions on receiving (yes, no) and giving (yes, no) instrumental support indicators to assess their association with configuration composition and structure (see Table 2). Four control variables were used, namely sex,  birth cohort, education, and citizenship. For configuration composition, we used the Nuclear configuration type as the reference as they represent a standard well-known kind. Results shown in Table 1 were corroborated. Indeed, in Parents configurations receiving and giving support overall was higher than in Nuclear configurations. Financial, material, and care supports were more frequently received by individuals in such configurations, showing downward as well as upward intergenerational solidarity. Kinship configurations were also positively associated with receiving support (overall, material, and care) and with giving support overall. Female friends and children configurations showed similar tendencies as Kinship ones, including for financial support. In contrast, most non-kin configurations (with the exception of Female friends and children) were more likely to be associated only with care exchanges. Considering the structure of personal configurations, density of emotional support was positively associated with giving care support, while betweenness of emotional support was negatively associated with receiving and giving financial support. Finally, density of conflict showed a negative association with receiving support overall and with care support in particular. Hypothesis 1c is confirmed, especially for care support. The negative association between centrality of emotional support and financial was not expected; we will return to that in Section 7.
Regarding control variables, the younger birth cohort was positively associated with receiving material support and negatively with giving financial and material support. This aspect underlines that intergenerational solidarity tends to be downward at younger ages.
To illustrate these results, we show the configurations of two individuals, a man and a woman, both born between 1970 and 1975, mentioning four important alters and having a Parent configuration. Individual 1 mentioned his partner, father, mother, and sister, while individual 2 mentioned her partner, friend, mother, and father. While in terms of composition, their configurations show similarities, they are quite different regarding their structure. The top parts a and b of Figure 1 illustrate the emotional support relationships, received and/or given, existing among all the configuration members, reported by these individuals (referred to as Ego). The configuration of individual 1 is characterized by a high density of support as all configuration members are connected (0.7) in contrast to the configuration of individual 2 (0.3). The bottom parts c and d of Figure 1 demonstrate the existing conflict relationships among all configuration members and themselves. Although individual 1 did not report any conflict (conflict density of 0), individual 2 reported them, especially between herself and her parents leading to a density of conflict of 0.2. When asked about instrumental exchanges, individual 1 reported receiving and giving care, financial, and material supports. In contrast, individual 2 only reported being able to receive material support, which shows the impact of the structure.

Reciprocity of exchanges of instrumental support in personal configurations
Reciprocity implies that individuals who provide support receive it at some point and vice versa. Indeed, among the 271 respondents involved in dyadic instrumental support exchanges, 61.8% were involved in reciprocal exchanges, while 21.7% were only support providers and 16.5% only support recipients (see Table 3). This underlines the overall prominence of reciprocity in such relationships. However, when investigating within each type of support, other patterns emerged. Reciprocity was low for financial support (13.1%), medium for material (32.6%), and high for care (65.5%). Giving or receiving financial support implies that the recipient is in a durable situation of need that makes it difficult or impossible to reciprocate. Consequently, some objects of reciprocity (e.g., money) are logically less required, but a global equilibrium is often achieved engaging other kinds of support, as the general pattern of reciprocity indicates. Those findings substantiate Hypothesis 2a about the prevalence of reciprocity over unidirectional exchanges, and the fact that reciprocity is higher at a global level than within each type of support considered separately.
In a second step, we performed a set of logistic regressions on reciprocity indicators to assess their association with configurations' composition and structure (see Table 4). Regarding the composition of personal configurations, Parents configurations were not statistically associated with reciprocity overall, but interestingly they were significantly less likely to be associated with providers of support only. Kinship and Professional configurations were three to four times more likely to be associated with reciprocity overall compared to Nuclear configurations. The former was, however, less likely to be associated with support-only providers, as it was the case for Female friends and children configurations. This confirms Hypothesis 2b, according to which reciprocity is more likely among personal configurations based on non-kin members.
Finally, configuration structures characterized by density of emotional support were positively associated with reciprocity. Conflict did not account for the directions of exchanges, indirectly underlying that reciprocity is expected independently of the quality of the relationships. Hypothesis 2c is confirmed regarding the positive impact of the density of emotional support, but not concerning the negative association of density of conflict. For control variables, the younger birth cohort was negatively associated with being a provider-only. This aspect confirms again that intergenerational solidarity tends to be downward at younger ages.

Discussion
We addressed the issue of instrumental support exchanges embedded in personal configurations. Overall, a third of the respondents reported being involved in instrumental support exchanges with their significant others outside the nuclear family, which means that two-thirds were not. This seemingly low percentage highlights that in Switzerland, as in most Western contemporary societies, instrumental support tends to be limited to situations of manifest need [Swartz (2009)]. This percentage has to be put in perspective with the Swiss context characterized by long-term socioeconomic stability and a welfare state providing basic social coverage [Obinger (1998), Wall et al. (2018]. However, this percentage is not a negligible one and it does require our full attention. As instrumental support exchanges between two individuals are embedded in wider webs of interdependencies, we considered the characteristics of personal configurations that may enhance or hinder them. The composition of personal configurations undoubtedly mattered. Indeed, instrumental support in personal configurations including parents and adult children was common. It confirms that support, especially financial, does not stop when children become adults and leave the parental nest [Bonvalet and Ogg (2007)]. Moreover, configurations giving room to extended kinship were a source of support. Irrespective of the exact type of tie, being related guarantees a space of safe exchanges governed by family solidarity rules. This highlights the fact that family relationships have a latent function embodying a "family national guard" [Hagestad (1996)] or a "support bank" [Antonucci and Jackson (1989)]. Latency is important outside the realm of the family as it has been underlined with the idea of fossil friendships that one can reactivate when needed [Pahl (2000)].
Personal configurations based on non-kin and especially those including female friendships proved to be significantly supportive. A process of suffusion can be witnessed [Pahl and Spencer (2004)], if financial support is put aside. The absence of partner and kin ties reinforces the importance of friendship. Friendship is prominent in contexts of singlehood, for instance [Bellotti (2008)]. Overall, all types of personal configurations could be a source of instrumental support, in particular for care support. This is an important finding as it shows that there is no type of personal configuration that is associated with an exclusion from support mechanisms. Nevertheless, as pointed above, the type and amount of support vary significantly depending on composition. One has to keep in mind that composition of personal configurations reflects individuals' overall embeddedness in society and may vary following important life-course stages, transitions, and events [Antonucci et al. (2010), Levy and Widmer (2013), Aeby et al. (2019)]. It implies that in case of crisis such as an intimate relationship breakdown, some close friends can reveal themselves disappointingly less supportive than expected [Aeby and Heath (2020)] or, contrastingly, that marital instability of adult children can positively reinforce support to parents and feelings of filial responsibility [Swartz (2009)]. It means that over the life course, individuals may have to re-evaluate their relationships and may drop conflictual ones and replace them with new supportive ones.
Reciprocity is a key feature of those instrumental support exchanges. It is an organizing principle of exchanges in society in general, as well as among close individuals [Molm et al. (2007), Girardin (2017)]. However, it did not systematically happen within the same type of support, which emphasizes the importance of global reciprocity in personal relationships, rather than direct reciprocity. Financial support often cannot be returned, but can be reciprocated by care services. While the percentage of reciprocity is high, some respondents found themselves in the position of a provider or recipient of support. This could be linked to differences in resources, as well as to the fact that in situations of a chain-generalized reciprocity where recipients and benefactors know each other [Molm et al. (2007)], some (younger) individuals have only experienced the interplay of support in one position of the chain yet, but they may experience it in a different way in the future. Regarding such exchanges, it is also interesting to highlight that we did not find any difference based on economic means, proxied by the level of education. Offer (2012) explained that individuals may withdraw from exchanges when they cannot reciprocate, rather than becoming exclusively a recipient. Being a recipient only is accepted for children toward parents, but less so for other types of ties. This intergenerational solidarity is accountable for some of the social inequalities being maintained over time [Chauvel (2006), Swartz (2009)].
Finally, the structure of personal configurations extended the understanding of instrumental support exchanges. Emotionally dense structures that provide bonding social capital [Coleman (1988), Widmer (2010)] proved to be associated with giving care and reciprocity overall. This protective but also controlling environment creates a collective normative pressure to ensure that individuals comply with their obligation of giving. This confirms the weight of support norms, existing in most Western societies with their specificities according to types of welfare regime [Widmer and Ganjour (2016)]. Contrastingly, in emotionally centralized structures producing bridging social capital where respondents represent a bridge between distinct configuration members [Burt (2002), Widmer (2010)], giving financial support was less frequent. This may be due to an impossibility to be simultaneously at the center of emotional and financial flows. In personal configurations producing bonding social capital, individuals were less likely to be providers only. This shows that individuals in emotionally dense supportive environment are less likely to feel drained and report an imbalance toward giving only. Personal configurations characterized by density of conflict were less associated with receiving, but this did not influence reciprocity exchanges in general. It underlines that reciprocity does not uniquely respond to supportive relationships, as most close relationships are characterized by ambivalence [Lüscher and Pillemer (1998), Aeby et al. (2020)]. Supportive and conflictual relationships tend to co-exist in personal configurations, while solely conflictual ones tend to be dropped [Uchino et al. (2004)].
This study has some limitations. First, the analyses for reciprocity were performed on a small number of individuals who were engaged in instrumental support exchanges. Second, the personal configurations were collected from single informants who gave their perception of the relationships involving themselves and their configuration members. Finally, regarding the differences between the type of supports and the type of ties, as we did not test each type of ties separately, but did it by configuration, we cannot provide a answer to the roles specifically played by friends. Despite these limitations, this study sheds new light on instrumental support exchanges by integrating them in a configurational approach that includes non-kin ties and by considering simultaneously the composition and structure of personal configurations.