Introduction
Sweden is known for its universalistic policies, which provide relatively generous and equalising social and economic benefits to people of all ages. The pension programme, a key component of the welfare system, has contributed to a stable financial situation for the average pensioner (Hagen et al. Reference Hagen, Laun and Palme2022). Additionally, the increased labour force participation of women in more recent cohorts has led to higher pensions among women, enhancing their financial resources (Statistics Sweden 2014). At the same time, the welfare of younger generations has changed in recent decades owing to societal challenges such as an unstable labour market and rising housing costs (Grander Reference Grander2023; Jansson Reference Jansson2020). As a result, the financial support provided by older adults to their children and grandchildren may have become more frequent and more important.
However, despite the generally advantageous position of older adults in Sweden, socio-economic inequality has increased in this population (Rehnberg et al. Reference Rehnberg, Östergren, Fors and Fritzell2022), partly owing to changes in the pension system (Hagen et al. Reference Hagen, Laun and Palme2022). This trend is part of a general reduction in public benefits that has increased overall societal inequality (Blomqvist and Palme Reference Blomqvist and Palme2020; Rostgaard et al. Reference Rostgaard, Jacobsen, Kröger and Peterson2022). Thus, while there is growing resource availability among older women, socio-economic inequality within the older population is increasing. Consequently, women’s contributions to financial transfers to younger generations are likely to grow, while socio-economic disparities in transfers are expected to widen over time.
To address these issues, this study examines financial transfers between older parents and their adult children and grandchildren in Sweden over two decades, focusing on gender and socio-economic patterns. Understanding the levels, changes and group differences in family financial transfers can shed light on the role of social protection benefits, such as pensions, and how societal changes shape family responses to, and the reproduction of, socio-economic inequalities.
Background
Financial circumstances of older parents and their children and grandchildren
The universal public pension system in Sweden plays a crucial role in ensuring financial security for older adults and is a cornerstone of the welfare state, aimed at reducing old-age poverty and providing a decent standard of living in retirement. Over time, the average pensioner in Sweden has benefited from a stable and progressive pension system relative to previous cohorts of older retirees and compared to working-age individuals (Hagen et al. Reference Hagen, Laun and Palme2022).
While Sweden did not experience a deep economic crisis during the Great Recession between the years 2007 and 2009, as seen in many other countries (Palme Reference Palme, Ólafsson, Daly, Kangas and Palme2019), the financial situation of those of working age still became unstable. For example, single households, with or without children, had an increased risk of poverty, which has partly been explained by less-generous income replacement policies in combination with higher thresholds to qualify for unemployment benefits (Alm et al. Reference Alm, Nelson and Nieuwenhuis2019). Younger adults also faced challenges entering the labour market owing to high youth unemployment and increased demand for a better-educated skilled workforce (Jansson Reference Jansson2020). Additionally, access to homeownership and rented housing was limited, leading to housing inequalities (Grander Reference Grander2023). These developments could partly explain why income disparities among individuals aged 20–34 years peaked between the years 2008 and 2014 (Hagen et al. Reference Hagen, Laun and Palme2022). Younger adults’ unpredictable life circumstances have led to a greater dependency on parents, either by relying on their financial resources for buying a home or by living longer in the parental home (Tenants’ Association 2021). With increasing life expectancy, it has been suggested that grandparents will assume a more substantial role within multi-generational families (Bengtson Reference Bengtson2001). Increasing financial support from grandparents to grandchildren may be part of that role.
Social inequality in old age
The financial situation of older adults is strongly tied to their previous participation in the labour market. In Sweden, as in many Western countries, female participation in both higher education and the workforce has steadily increased throughout the 20th century. Three out of four women born in the mid-1920s were in the workforce during their 50s, while the equivalent share for those born in the mid-1940s was nine in ten (Statistics Sweden 2014). While the financial situation for women has improved in absolute terms, women often have lower pensions than men owing to more prevalence of part-time work and lower lifetime incomes. Despite the slightly reduced gender pay gap observed in Sweden’s working population in recent decades, this change has not yet benefited retired women (Hagen et al. Reference Hagen, Laun and Palme2022). Furthermore, despite progress made by women in the workforce, significant gender disparities in financial resources persist between older men and women in Sweden, which exhibits one of the highest rates of gender inequality in old-age poverty within the European Union (European Commission 2021).
During the last decades, inequalities in resources, such as financial means, health, social relations and care access, have expanded in the general population (Berkman et al. Reference Berkman, Kawachi and Glymour2014; Fritzell et al. Reference Fritzell, Hertzman, Bäckman, Borg, Ferrarini, Nelson, Nolan, Salverda, Checchi, Marx, McKnight, Tóth and van de Werfhorst2014) and also among older adults (Blomqvist and Palme Reference Blomqvist and Palme2020; Fors et al. Reference Fors, Illinca, Jull, Kadi, Phillips, Rodrigues, Vafaei, Zolyomi and Rehnberg2022; Rehnberg et al. Reference Rehnberg, Östergren, Fors and Fritzell2022; Rostgaard et al. Reference Rostgaard, Jacobsen, Kröger and Peterson2022; von Saenger et al. Reference von Saenger, Dahlberg, Augustsson, Fritzell and Lennartsson2023). Societal changes have also shifted the labour market from an industrial to a service-based economy, consequently altering the class structure within the workforce. Over the past 50 years, the share of unskilled workers, self-employed individuals and farmers has decreased, while non-manual workers have become more prevalent. This shift reflects a diminished need for industrial skills in relation to a growing demand for non-manual service workers and professionals in a knowledge-based economy (Fritzell and Lundberg Reference Fritzell and Lundberg2007).
More recently, the participation of older workers in the labour force has increased, though significant differences persist by occupational status. From 2005 to 2018, non-manual workers extended their working lives more than manual workers, further widening disparities and contributing to financial inequality in old age (Fransson and Söderberg Reference Fransson and Söderberg2018). These trends underscore the continued relevance of occupational-based social class as a key indicator of socio-economic resources at the end of one’s working life (Shahbazian and Bihagen Reference Shahbazian and Bihagen2022).
Welfare system changes are closely tied to shifts in social disparities in society. In Sweden, the pension system, a key part of the welfare structure, has shifted from a traditional defined benefit model to a contribution-based system, emphasising individual choice and privatisation (Hagen Reference Hagen2013). This shift has raised concerns about weakening the system’s universalistic goal of ensuring adequate and equitable benefits, potentially increasing financial inequality in old age (Blomqvist and Palme Reference Blomqvist and Palme2020). Occupational pensions, provided by employers, form an essential part of the contribution-based system. It has especially benefited those in non-manual occupations, thereby intensifying financial inequality among older adults and between different cohorts of the ageing population over the last 30 years (Hagen et al. Reference Hagen, Laun and Palme2022).
Intergenerational financial transfers
Through recorded history, the multi-generational family has been important in determining resource availability for individuals across the lifespan into old age (Blieszner and Voorpostel Reference Blieszner, Voorpostel, Bengtson and Settersten2016). In contemporary Western societies, flows of financial and social resources go in both directions, yet net economic support tends to flow downwards, that is, from older to younger generations (Tur-Sinai and Lewin-Epstein Reference Tur-Sinai and Lewin-Epstein2020). Although the downward flow of resources diminishes with parental age, even those aged over 70 years make significant net contributions to their adult children (Albertini et al. Reference Albertini, Kohli and Vogel2007; Lennartsson et al. Reference Lennartsson, Silverstein and Fritzell2010). Parental intergenerational transfers have been explained from several theoretical perspectives (for an overview see Kohli and Künemund Reference Kohli, Künemund, Bengtson and Lowenstein2003). One recurring aspect involves the altruistic motivations that arise when resources flow from more privileged to less privileged individuals within families. Viewed through the lens of intergenerational financial transfers, this dynamic corresponds to the assumption that the wellbeing of adult children (and grandchildren) is important for older parents (Laitner Reference Laitner, Rosenzweig and Stark1997) and that families function as social welfare agents in the form of ‘efficient’ distributors of resources to the most vulnerable individuals with the greatest need (Kohli and Künemund Reference Kohli, Künemund, Bengtson and Lowenstein2003).
Another perspective that highlights the role of behaviour, resources and needs in intergenerational transfers is the Informal Care Model proposed by van Groenou and De Boer (Reference van Groenou and De Boer2016). As the demand for care from adult children to ageing parents increases, barriers limiting the provision of care become important for older family members to address. These barriers may include financial costs related to travel, purchasing goods or services for parents, or reducing participation in the paid labour market. Drawing on reciprocity theory – which suggests that individuals provide support with the expectation of receiving benefits in return, either immediately or in the future (Silverstein et al. Reference Silverstein, Conroy, Wang, Giarrusso and Bengtson2002) – parents may provide financial transfers as compensation for the care they receive from their adult children.
However, family transfers do not stand alone but should be considered in the context of each country’s filial norms, cultural values and welfare state programmes (Connidis and Barnett Reference Connidis and Barnett2018; Kohli Reference Kohli and Silverstein2004). From welfare regime theory (Esping-Andersen Reference Esping-Andersen1990) one would first assume that financial transfers would be marginal in countries where the state has a more prominent role in the provision of welfare. On the contrary, comparative studies have found clear links between public and private transfers, where more-generous welfare regimes predict a higher likelihood of intergenerational transfers, for example financial resources, than less-generous welfare systems (Brandt Reference Brandt2011). However, the monetary levels are usually lower than in less-generous welfare states (Brandt and Deindl Reference Brandt and Deindl2013). The importance of family-oriented values, often linked to a sense of duty or obligation, also reflects welfare regime profiles. In Nordic countries, citizens usually prioritise state support over family support compared to other European countries (Albertini and Kohli Reference Albertini and Kohli2013).
Motives for providing resources can also vary depending on gender and socio-economic belonging. Older women, given their greater investment in family ties, tend to have closer relationships with younger generations (Silverstein et al. Reference Silverstein, Gans and Yang2006) and are more likely to help them financially, driven by altruistic motives when children are in need compared to older men who tend to reason in more reciprocal terms (Baeriswyl et al. Reference Baeriswyl, Girardin and Oris2022; Kohli Reference Kohli and Silverstein2004). Financial transfers can also function to reproduce status within diverse socio-economic groups. Drawing on European data, Albertini and Radl (Reference Albertini and Radl2012) demonstrated that parents from higher social classes provided more financial support to their adult children, even after adjusting for income, education and wealth, compared to their counterparts from lower social classes. In other words, looking through the altruistic lens, the fear of children falling behind when in need was interpreted as more important among persons in higher rather than lower social strata. While theories provide frameworks for the interpretation of findings, this study does not aim to validate or test particular theoretical claims. Regardless of the welfare context and the possible motives for giving, intergenerational transfers of resources, or the lack of such transfers, play an essential part in the life chances of people of all ages (Bengtson Reference Bengtson2001).
Distribution of downward financial transfers
Earlier Swedish studies have revealed that at the start of the millennium approximately one in four adults aged 60 or older had provided financial support to either a child or a grandchild (Fritzell and Lennartsson Reference Fritzell and Lennartsson2005; Lennartsson et al. Reference Lennartsson, Silverstein and Fritzell2010). In a comparative study using data from persons aged 50 years or older, Sweden had the highest occurrence (32 per cent) of financial transfers to adult children compared to all other European countries (Schenk et al. Reference Schenk, Dykstra and Maas2010). In the United States, parental financial support for younger adults, particularly students, has been on the rise, reflecting a prolonged transition into adulthood (Aquilino Reference Aquilino2005). This trend aligns with altruistic motivations, as most downward financial transfers are targeted towards those in need (Albertini et al. Reference Albertini, Kohli and Vogel2007; Fingerman et al. Reference Fingerman, Miller, Birditt and Zarit2009; Wong et al. Reference Wong, Liao, Etherton-Beer, Baldassar, Cheung, Dale, Flo, Husebø, Lay-Yee, Millard, Peri, Thokala, Wong, Chau, Chan, Chung and Yeoh2020). Similar patterns were found for grandchildren in the US, where grandparents were more likely to financially support adult grandchildren when their parents were unemployed or studying (Huo et al. Reference Huo, Kim, Zarit and Fingerman2018). A German study found that the likelihood of financial transfers to grandchildren increased between 1996 and 2002, while transfers to children decreased to some extent. However, a higher percentage of older persons made financial transfers to children (23 per cent) compared to grandchildren (17 per cent) (Hoff Reference Hoff2007).
When examining downward intergenerational financial transfers, we will also consider grandchildren, an area that has received less attention in previous studies. The involvement of this younger generation in our analysis is deemed essential for enriching our understanding of multi-generational relationships, as emphasised by Remle (Reference Remle, Settersten and Angel2011).
Given the financially differentiated development for older adults and their younger family members during the last two decades in Sweden, we hypothesise that:
1: Downward intergenerational financial transfers have increased, regarding both the occurrence and the amount provided, reflecting a more positive overall financial situation in the older generation, in particular, compared to young adults.
Gender differences in downward financial transfers
The few studies in Sweden and Europe assessing the importance of gender among older adults for the occurrence of downward financial transfers found no significant differences between women and men (Fritzell and Lennartsson Reference Fritzell and Lennartsson2005; Kohli Reference Kohli and Silverstein2004; Lennartsson et al. Reference Lennartsson, Silverstein and Fritzell2010). The living situation of the parent, and consequently the resources availability, rather than gender itself, has been one of the main explanations for non-differential gender findings (Fritzell and Lennartsson Reference Fritzell and Lennartsson2005). Given previous research demonstrating both financial advances made by older women and the persistence of gender disparities during the last two decades in Sweden, we hypothesise that:
2a: Downward intergenerational financial transfers are gender equal in the occurrence of providing and unequal in the amounts provided, where women provide lower amounts since women still have fewer financial resources than men.
2b: Women are increasingly common providers of downward financial transfers and provide higher amounts in more recent cohorts than women in previous cohorts, owing to women’s increasing financial resources.
Socio-economic differences in downward financial transfers
International studies have consistently highlighted the importance of higher socio-economic position in providing financial support to both children and grandchildren, measured as income and wealth (Albertini and Radl Reference Albertini and Radl2012; Wong et al. Reference Wong, Liao, Etherton-Beer, Baldassar, Cheung, Dale, Flo, Husebø, Lay-Yee, Millard, Peri, Thokala, Wong, Chau, Chan, Chung and Yeoh2020), education (Wong et al. Reference Wong, Liao, Etherton-Beer, Baldassar, Cheung, Dale, Flo, Husebø, Lay-Yee, Millard, Peri, Thokala, Wong, Chau, Chan, Chung and Yeoh2020) and, in Sweden, occupational class (Fritzell and Lennartsson Reference Fritzell and Lennartsson2005). With increasing social inequality, families in higher socio-economic positions are continuously more likely to share their resources downward, while those in lower socio-economic positions are not. Thus, social class differences might be widening over time.
Indeed, the socio-economic gap in parental financial investments in children has been growing among the working population (<65 years) in the US between 1975 and 2014 (Schneider et al. Reference Schneider, Hastings and Labriola2018). The same pattern was revealed among a slightly older population (50–75 years) in the US, linking increases in income inequalities within states to increases in socio-economic differences in downward financial transfers (Floridi Reference Floridi2024). A Norwegian study revealed that the economic upper class, comprising the top 10 per cent of executives, managers and financial brokers, played a significant role in driving socio-economic disparities in both the proportion and the amount of financial support provided to the younger generation aged 24–31 years (Hansen and Wiborg Reference Hansen and Wiborg2019). The same patterns were observed in a Swedish study, where the wealthiest parents typically had children who performed better in various aspects related to their social and economic wellbeing, with a notable impact on their income and wealth (Hällsten and Thaning Reference Hällsten and Thaning2021). In addition, Kalmijn (Reference Kalmijn2024) identified with Dutch data that the stratification process was strongest in early adulthood compared to other stages in the lifecourse. The socio-economic background of older parents is therefore a key factor in shaping the resources available to adult children and their families. However, it is important to recognise that intergenerational financial transfers themselves play a substantial role in the development of increasing wealth disparities among younger generations. This was evidenced in a multinational study including France, Spain, Great Britain and the US (Palomino et al. Reference Palomino, Marrero, Nolan and Rodríguez2022).
Two aspects guide the last hypothesis, which considers socio-economic differences in financial transfers. First, older adults have increasingly worked in middle and higher non-manual occupations (Fritzell and Lundberg Reference Fritzell and Lundberg2007), while income inequality among the older population has increased, partly owing to changes in the pension system. Second, in recent decades there have been increasing differences in life circumstances between generations within families. Younger generations often face more unstable financial situations compared to older ones, prompting parents with resources to provide greater financial support. We, therefore, hypothesise that:
3a: Downward intergenerational financial transfers are shaped by socio-economic inequalities in the occurrence and the amount provided, where older parents belonging to higher social classes are more likely to be providers and to transfer higher amounts than parents in lower social classes.
3b: In recent cohorts, a larger proportion of older parents from higher social classes, compared to those from lower social classes, are providing financial transfers to their children and grandchildren. Furthermore, higher social classes are increasing the amounts they contribute over time.
The overall aim of this study is to examine intergenerational financial transfers between 2002 and 2021. More specifically, we examine changes in (a) the proportion of intergenerational financial providers among older parents and the amounts transferred, and (b) the variation of intergenerational transfers by gender and social class.
Materials and methods
Design and respondents
This study is based on repeated cross-sectional data from the 2002, 2011 and 2021 data collection waves of the Swedish Panel Study of Living Conditions of the Oldest Old (SWEOLD), a national sample of persons aged 77 years or older living in Sweden at the time of the interview. The SWEOLD draws respondents from the Swedish Level of Living Survey (LNU), a study of approximately 1 in 1,000 of the Swedish population aged 15–75 years (Fritzell and Lundberg Reference Fritzell and Lundberg2007). The SWEOLD sample includes all individuals who were initially in the LNU sample but dropped owing to the age ceiling restriction. The 2011 and 2021 waves of the SWEOLD were complemented with an additional national sample of women and men, residents of Sweden at the time of the interview, stratified by gender and five-year age groups. The additional samples were obtained using the personal identification numbers issued to all Swedish residents by the state. The response rate was 84.4 per cent in 2002, 86.2 per cent in 2011 and 63.6 per cent in 2021.
The SWEOLD study uses different interview methods to avoid high non-response rates because of poor health or impaired cognition. The study is based on questionnaires, and data collection was primarily done face-to-face in 2002 and 2011, although, when preferred, interviews were carried out via telephone. In 2021 the interviews were primarily carried out via telephone owing to the Covid-19 pandemic. If the participant was unable to answer questions, for example because of cognitive impairment or physical frailty, indirect (or mixed) interviews were carried out with the person’s spouse/partner, carer or another close person. In SWEOLD 2011 and 2021, postal questionnaires were used when the participant was not able to interview via telephone, for example owing to hearing problems or because of indirect interviews. Informed consent was obtained before each interview. Ethical approval for the SWEOLD study was provided by the Karolinska Institutet Regional Committee for Research (03-413), the Regional Ethical Review Board in Stockholm (2010/403-31/4) and the Ethical Review Agency (2019-06324; 2021-00393; 2021-05031; 2024-03705-02). For more detailed information about the study, see Lennartsson et al. (Reference Lennartsson, Agahi, Hols-Salen, Kelfve, Kareholt, Lundberg, Parker and Thorslund2014).
The analytic sample is restricted to those with living children who have complete information on the dependent and main independent variables. The final sample consisted of 1,982 older parents (see Figure 1). There was an increase (p = 0.006) in the proportion having children across the three waves from 83.9 per cent to 89.3 per cent (see Figure S1 in the online supplementary material).
Figure 1. Flowchart of the final analytic sample in 2002, 2011 and 2021.
Dependent variables
The two dependent variables described (1) the occurrence of financial transfers, regardless of amount, from older parents to younger family members, measured with transfers to children and/or grandchildren taken together, and (2) the flow of financial transfers amounting to a minimum sum of 10,000 SEK (Swedish krona) from older parents to younger family members, with children and/or grandchildren taken together. The first variable was measured via the item: ‘Have you, during the last 12 months, given any financial support or gifts of a value of 5,000 SEK or more to anyone outside your household?’, where those who replied ‘yes’ were then asked ‘To whom?’.
When a child or grandchild was mentioned, the older parent was considered a provider of downward financial transfers. Since the first question related to a threshold of 5,000 SEK in total to anyone outside of the household (including ‘other persons’ and ‘charitable organisations’), the sum provided to children and grandchildren could be less than 5,000 SEK. Therefore, the first dependent variable captures the occurrence of downward providers of any sum. The second variable was measured via the final item: ‘Can you give an approximate value, for the last 12 months?’.
Questions regarding providing transfers were asked in the same way in all interview years, but the minimum value was higher in 2011 (6,000 SEK) and 2021 (7,000 SEK), accounting for inflation.
The sums given to children were collected at an aggregated level in 2002, meaning that they were not child-specific. In 2011 and 2021, however, the answers were child-specific and derived from categorical sums using the median value for each child receiving financial support. The sums given to grandchildren were collected on an aggregated level for all years where the median value for each categorical sum was used. The above estimations were standardised for comparability across all years. Although sensitivity analyses (not shown) indicate that this assessment may have underestimated actual values, altering the threshold within each category – whether lower or higher – did not significantly predict the final estimation of sums. The final dichotomous variable used to represent transferred amounts was constructed using a threshold of over 10,000 SEK. Maintaining the same cut-off in all three waves was considered acceptable owing to the low inflation during the relevant years (Statistics Sweden 2023).
Independent variables
Key independent variables were the parental gender and household social class. The measure of social class followed the official Swedish ‘SEI classification’ (Andersson et al. Reference Andersson, Erikson and Wärneryd1981), which is based on several dimensions of, and has many similarities with, the EGP-1 classification (Erikson et al. Reference Erikson and Goldthorpe1992). Social class was measured by the respondents’ and his/her partner’s occupations. The dominant social class was then used for the household as a whole assuming that some social classes have greater influence than others on the general attitudes and behaviour patterns of the household (Erikson Reference Erikson1984).
In this study, the household class was represented by four groups: manual workers, lower non-manuals, intermediate and higher non-manuals, and self-employed and farmers. For 13 individuals, gender-stratified imputation of class based on the average years of education was performed.
Additional variables used to describe the study sample were the respondents’ age group, living situation, self-rated health, whether the respondent has a cash margin, whether the respondent has grandchildren and the mean number of children. Living situation distinguished those who were living alone, including people living in nursing homes, from those who were living in the same household with other(s). Self-rated health distinguished those with good, bad, or neither good nor bad health. The cash margin, a financial buffer, measured the respondent’s ability to generate a set sum within a week, standardised to 12,000 SEK (2002), 14,000 SEK (2011) and 16,000 SEK (2021). Individuals had a cash margin if they could obtain the funds from personal resources, like withdrawing from a bank or selling stocks. The number of children was based on how many living children the participant had at the time of the interview. To see if the incoming flow of financial transfers in the household had changed during the study period, we included a variable measured by the question: ‘Have you during the last 12 months received any financial support or gifts of a value of 5,000 SEK or more from anyone outside your household?’.
Analyses
Descriptive statistics aimed at presenting the characteristics of the analytical sample are presented in Table 1. Further descriptive analyses display the proportion of givers with 95 per cent confidence intervals in terms of occurrence (Table 2) and giving more than 10,000 SEK (Table 3) to children and/or grandchildren taken together. These tables show the distribution of providing financial transfers for the pooled sample, stratified by gender and social class as well as differentiated by year. The distribution of financial transfers given to children and grandchildren separately and taken together is shown in additional figures (Figures 2a, 3b). A descriptive analysis presents the proportion of older parents providing to their children and grandchildren combined by amounts in Swedish currency (SEK) for all years (Figure 3a). To test for statistically significant changes in the pooled sample over the years in total occurrence and amounts transferred as well as gender and social class differences, logistic regression analyses were conducted (Tables 2–3). To visualise these gender and social class differences over the years, interaction analyses (year X gender; year X social class) were performed with logistic regression analyses presented as predicted probabilities with 95 per cent confidence intervals (Figures 2b–c; 3c–d). No control variables were added to the analyses, since the study aimed to descriptively investigate change over time rather than to explain such changes. Since some individuals had participated in more than one of the three sample waves, the statistical tests were performed with robust standard errors adjusting for clustering. Weights were used in the analyses to compensate for the unequal probability of inclusion in the sample, depending on gender and age (owing to stratified sampling of additional samples, as presented earlier). Data were analysed using Stata 17.0 for Windows.
Figure 2. (a) The proportion (%) giving financial resources to children and/or grandchildren taken together and separately for the years 2002, 2011 and 2021. (b) The predicted probability of giving financial resources to children and/or grandchildren taken together by gender for the years 2002, 2011 and 2021. (c) The predicted probability of giving financial resources to children and/or grandchildren taken together by household class for the years 2002, 2011 and 2021.
Figure 3. (a) The proportion (%) given to children and/or grandchildren taken together by amounts in Swedish krona (SEK) for the years 2002, 2011 and 2021. (b) The proportion (%) giving more than 10,000 SEK to children and/or grandchildren taken together and separately for the years 2002, 2011 and 2021. (c) The predicted probability of giving more than 10,000 SEK to children and/or grandchildren taken together by gender for the years 2002, 2011 and 2021. (d) The predicted probability of giving more than 10,000 SEK to children and/or grandchildren taken together by household class for the years 2002, 2011 and 2021.
Table 1. Characteristics of the analytical sample of parents with living children aged 77 years and older in 2002, 2011 and 2021 (n = 1982)
Notes:
a In 2002, 64 persons did not receive the ‘Self-rated health’ question when indirectly interviewed.
b Missing values varied between 1 and 14.
c This question was not asked in the postal questionnaires in 2011 (n = 78). Percentages are weighted using sample weights.
Table 2. Total and gender and household class stratified descriptive statistics with 95% confidence intervals of giving financial support to children and/or grandchildren as reported by older parents in the pooled sample and in the years 2002, 2011 and 2021 (%)
Notes: Percentages are weighted using sample weights. Logistic regression was used to assess differences (1) by gender and household class in the pooled sample, (2) over the years in the total group and by gender and household class. Statistically significant values are in bold (p < 0.05). Ref = reference category.
Table 3. Total and gender and household class stratified descriptive statistics with 95% confidence intervals of giving more than 10,000 SEK to children and/or grandchildren as reported by older parents in the pooled sample and in the years 2002, 2011 and 2021 (%)
Notes: Percentages are weighted using sample weights. Logistic regression was used to assess differences (1) by gender and household class in the pooled sample, (2) over the years in the total group and by gender and household class. Statistically significant values are in bold (p < 0.05). Ref = reference category.
Results
Characteristics of the sample
The characteristics of the analytical sample are presented in Table 1. A majority of the respondents were women, aged 80–84 years. More than half of the sample were living alone in 2002 and 2011, while a smaller proportion were doing so in 2021. Most respondents considered their health to be good or neither good nor bad, while the group with poor self-rated health was almost twice as prevalent in 2011 (13.6 per cent) as in 2021 (7.7 per cent). The average number of children varied between 2.26 and 2.47, while more than nine out of ten older parents had grandchildren across the years. Regarding the socio-economic factor, the share of manual workers and self-employed and farmers diminished over time, while those belonging to intermediate and higher non-manual occupations doubled between 2002 and 2021. There was a lower proportion of people without a cash margin in 2021 than in 2002. Overall, at least one in four older parents provided financial support of a total minimum sum of 5,000 SEK to anyone outside of the household with the highest share found in 2011 (30.9 per cent). Few parents received financial support from outside the household, with the highest number of recipients observed in 2011 (n = 25). A more detailed description of the number of recipients receiving support from their children is provided in Table S2.
Proportions, changes and gender and social class differences in the occurrence of downward intergenerational financial transfers
Figure 2a presents the proportion of older parents providing financial resources to children and/or grandchildren taken together, and for children and grandchildren separately. While a higher proportion of parents provided financial transfers to children (16.4 per cent) in 2002 than to grandchildren (12.8 per cent), no difference was found in 2021. The share of those giving to grandchildren increased during the study period and was 17.2 per cent in 2021.
As presented in Table 2, 25.5 per cent of the respondents in the pooled sample reported giving financial support to children and/or grandchildren. Parents were more likely to give in 2011 (29.5 per cent) compared to 2002 (22.1 per cent), while no significant change was found between 2002 and 2021 (24.6 per cent).
Starting with gender, no significant difference was found between women (24.0 per cent) and men (27.8 per cent), while an increase in giving in 2011 compared to 2002 was found for both genders. In the year 2021, women maintained a similar level of giving as in 2011, contrary to men, who gave to a lesser extent. To visualise the interaction between year and gender, Figure 2b shows the predicted probability of giving for both women and men over time. The results suggest that women have increasingly become financial providers over time, while men in the most recent cohort exhibit similar levels of financial support as the oldest cohort.
Sensitivity analyses (Table S3) showed no statistically significant results for gender differences in giving. Among those living alone, more men than women provided financial transfers. Similarly, cohabiting women showed an increasing tendency to give during the study period, and in 2021, women (31.8 per cent) slightly exceeded men (26.7 per cent) in providing financial transfers.
Table 2 shows that social class is important for the likelihood of giving. Intermediate and non-manuals (30.7 per cent) and self-employed and farmers (27.3 per cent), respectively, were more likely to give than manual workers (18.4 per cent). Social class-stratified analyses showed no statistically significant changes in giving by year. However, as for gender differences, the year 2011 suggested that intermediate and non-manuals as well as self-employed and farmers became more common givers than in previous years. Figure 2c shows the predicted probability of giving for different social classes over time. Although the likelihood of giving differs by social class across the study period, this difference is statistically significant only in 2021. In 2021, manual workers were on average 15.6 per cent (95 per cent CI; 8.7, 22.5) likely to provide financial transfers to a child and/or a grandchild, while intermediate and upper non-manuals were almost twice as likely (28.6 per cent) to do the same (95 per cent CI; 23.1, 34.0) (see Table S5).
Amounts transferred to children and/or grandchildren
Figure 3a shows the proportion who provided downward financial transfers by categories of amounts in Swedish krona for all three years. The higher values were less common than the lower values. In 2011, the values of 1–5,000 and 5,001–10,000 SEK were more often transferred compared to other years. In 2002, it was more common to give 1–5,000 SEK (6.1 per cent) than to give more than 100,000 SEK (0.4 per cent), while in 2021 giving such amounts was equally common (4.0 per cent and 3.7 per cent, respectively). The results suggest a trend of rising amounts transferred over consecutive data collection waves. To analyse this further, Figure 3b presents the proportion of parents who gave more than 10,000 SEK to children and grandchildren together and separately. The proportion of givers to children and/or grandchildren increased (p = 0.004) from 7.9 per cent in 2002 to 13.4 per cent in 2021. While it was more common to provide a higher amount to children than to grandchildren, the difference narrowed over time. Indeed, the likelihood of giving to grandchildren increased across the last decades (p = 0.003).
Proportions, changes and gender and social class differences in downward financial transfers of higher amounts
As presented in the pooled sample in Table 3, one in ten respondents reported giving more than 10,000 SEK to children and/or grandchildren. Parents were more likely to give higher amounts in 2021 (13.4 per cent) compared to 2002 (7.9 per cent) and men (13.5 per cent) were twice as likely to give higher amounts than women (7.8 per cent). Results from the gender-stratified analyses show that women were twice as likely to give in 2021 (12.8 per cent) than in 2002 (6.3 per cent). Figure 3c shows the predicted probability of giving more than 10,000 SEK for women and men over time. In 2011, there were gender differences that did not exist in the previous or the following year. Overall, the figure also shows that it was women who increased their likelihood of giving higher amounts in the last decade.
Table 3 shows that social class is important for the likelihood of giving higher amounts. Intermediate and non-manuals (13.2 per cent) and self-employed and farmers (11.1 per cent), respectively, were more common givers of financial support than manual workers (5.9 per cent). Social class-stratified analyses showed no statistically significant changes in giving by year. However, in 2021, all social classes but manual workers showed a greater tendency to give financial support compared to other years. Figure 3d shows the predicted probability of giving for social classes over time. No significant differences were found; however, the social class pattern seemed to change with time. Manual workers and lower non-manuals were on average 5.3 per cent and 3.8 per cent likely, respectively, to give higher amounts in 2002. The difference between these two social classes widened and was reversed in 2021 where lower non-manuals had a higher likelihood of giving (11.1 per cent) than manual workers (7.2 per cent).
Discussion
Over the last decades, various societal changes have occurred that are likely to influence downward intergenerational financial transfers. These include generational differences in life opportunities (Grander Reference Grander2023; Jansson Reference Jansson2020), increasing resource availability among a large portion of older women (Statistics Sweden 2014) and increasing socio-economic inequalities (Berkman et al. Reference Berkman, Kawachi and Glymour2014; Blomqvist and Palme Reference Blomqvist and Palme2020; Fritzell et al. Reference Fritzell, Hertzman, Bäckman, Borg, Ferrarini, Nelson, Nolan, Salverda, Checchi, Marx, McKnight, Tóth and van de Werfhorst2014; Rostgaard et al. Reference Rostgaard, Jacobsen, Kröger and Peterson2022). The expanding ageing population is becoming increasingly crucial for sustaining family life and resources (Blieszner and Voorpostel Reference Blieszner, Voorpostel, Bengtson and Settersten2016). Consequently, intergenerational financial transfers can be viewed as responses to these societal shifts and may potentially perpetuate inequalities across generations. This study aims to explore the occurrence and magnitude of intergenerational financial transfers in Sweden between 2002 and 2021, with a focus on gender and social class differences.
We found that at least one in four parents provided financial resources to a child or a grandchild. The proportion of providers varied between 22 per cent and 30 per cent across the years with an increasing tendency; however, this was not statistically significant. The most apparent change related to who was receiving transfers. While it was previously most common to provide financial support to adult children, a larger share of receivers in 2021 consisted of grandchildren.
The results also indicate that while the lower amounts were standard in previous years, it has become as common to provide the highest as the lowest amount in 2021. The likelihood of providing more than 10,000 SEK has nearly doubled over time and around 13 per cent of older parents did so in 2021. Hypothesis 1 concerning supposed increases in the occurrence of downward financial transfers over time was partly confirmed, since this pattern was found when the receivers were grandchildren. These findings echo and complement research conducted in Germany up until 2002, which also showed increasing giving patterns to grandchildren (Hoff Reference Hoff2007). Regarding increases in the amounts transferred, this aspect of Hypothesis 1 was also confirmed. Parents provided larger financial transfers in 2021 than they did 20 years earlier. This growth in both the occurrence and the amount of transfers is largely attributed to grandparents’ increased involvement with grandchildren, with the rise in transfer amounts being particularly notable between 2011 and 2021.
Several possible explanations exist for the overall increases in support directed towards grandchildren. First, the older population has become increasingly advantaged in terms of having good health, living with a partner (Dahlberg et al. Reference Dahlberg, von Saenger, Naseer, Lennartsson and Agahi2024), being more likely to have worked in non-manual occupations and having a cash margin. Second, the mean number of children was higher in 2011, which could lead to more parents being financially involved in their offsprings’ lives. Finally, more-unstable living conditions among younger generations since 2011 (Alm et al. Reference Alm, Nelson and Nieuwenhuis2019; Jansson Reference Jansson2020) could have motivated altruistic behaviour from older generations (Kohli and Künemund Reference Kohli, Künemund, Bengtson and Lowenstein2003; Laitner Reference Laitner, Rosenzweig and Stark1997).
The notable rise in transfers to adult children and grandchildren in 2011, followed by a decrease in 2021, may have additional explanations. One could be attributed to the removal of the gift tax in Sweden in 2004 (Henrekson and Waldenström Reference Henrekson and Waldenström2016), potentially encouraging grandparents to provide lower amounts to a larger number of receivers in 2011 than in the previous wave. Indeed, sensitivity analyses show that it was more than twice as common (14 per cent) to give up to 5,000 SEK to grandchildren in 2011 compared to other years (see Figure S3). The effect of the removal of the gift tax might have diminished by 2021. The decrease observed in 2021 could also be attributed to the Covid-19 pandemic. From a reciprocal standpoint, limitations on social interaction with family members may lead to fewer financial transfers downward (Lennartsson et al. Reference Lennartsson, Silverstein and Fritzell2010; Silverstein et al. Reference Silverstein, Conroy, Wang, Giarrusso and Bengtson2002). In addition, while younger adults were particularly at risk of unequal living conditions around 2008 and 2014 (Hagen et al. Reference Hagen, Laun and Palme2022), the pandemic contributed to lower consumption patterns in the total population during 2021 (Roos Reference Roos2020), possibly decreasing the need for financial support. In summary, similar to the situation in the US, older parents in Sweden appear to play a role in extending the transition to adulthood for their younger family members, particularly in response to societal changes (Aquilino Reference Aquilino2005). This suggests that, despite the general belief in Nordic countries that the state, rather than the family, should provide support when needed (Albertini and Kohli Reference Albertini and Kohli2013), older generations still feel a sense of duty or obligation to help, regardless of the amount transferred.
Although not part of our research aim, findings on financial transfers from outside the household to older individuals suggest that, in 2011, some parents may have required greater support, or some providers may have been more generous compared to previous or subsequent years. This could point to gaps in welfare protection following the financial crises of 2007–2009 (Palme Reference Palme, Ólafsson, Daly, Kangas and Palme2019).
Gender differences in financial transfers and changes over time
Focusing on gender, findings confirmed Hypothesis 2a that downward financial transfers were gender equal in the likelihood of providing; however, they were gender unequal in terms of the amounts provided, where women gave lower amounts than men. Moreover, there was a growing tendency among women to contribute more over time, while men generally maintained their previous levels. Notably, the development among men was primarily attributed to a decrease in providing in 2021, whereas women continued to contribute at similar rates as in 2011. Additionally, women tended to provide higher amounts more frequently in 2021 than in previous years, while men’s increase happened between 2002 and 2011. In sum, the findings substantiate Hypothesis 2b suggesting that, in more recent cohorts, women are more frequent contributors and provide higher amounts compared to women in earlier cohorts.
These results complement previous research (Fritzell and Lennartsson Reference Fritzell and Lennartsson2005; Kohli Reference Kohli and Silverstein2004; Lennartsson et al. Reference Lennartsson, Silverstein and Fritzell2010) by showing that patterns in both the occurrence and the amounts provided vary by gender, where women seem to make use of their increasing financial resources by supporting younger family members. Our gender-specific findings also prompt questions regarding the potential changes in gender dynamics in future intergenerational financial transfers. If women are becoming more important for these transfers, possibly owing to increasing financial bargaining power within the household (Andreoni et al. Reference Andreoni, Brown and Rischall2003; Wiepking et al. Reference Wiepking, Einolf and Yang2023) and their stronger family commitments (Silverstein et al. Reference Silverstein, Gans and Yang2006), how will the transfer behaviour of men develop? Future research should explore whether women’s financial contributions in old age are less reciprocal than those of men. If this is the case, the observed development of intergenerational financial transfer in this study would suggest a future decline in care support from adult children to older fathers, given that the financial needs of younger generations may have been met ‘for free’ by mothers or grandmothers (Baeriswyl et al. Reference Baeriswyl, Girardin and Oris2022; Kohli and Künemund Reference Kohli, Künemund, Bengtson and Lowenstein2003). Since there are gender differences in income in old age, our results also suggest that women and men might differ in how they prioritise sharing their financial resources downwards even when differences in resource availability exist.
While many older women have improved their financial wellbeing compared to previous cohorts, this does not apply to everyone. Sensitivity analyses, consistent with previous studies (European Commission 2021; Jansson Reference Jansson2020), showed that women more often than men lacked financial resources such as a cash margin (see Table S1). Future studies should assess the financial vulnerability of older women to inform policies that improve their disadvantaged status.
Social class differences in financial transfers and changes over time
Our study confirms Hypothesis 3a, showing that downward financial transfers are shaped by social class inequalities. Parents from higher social classes are more likely to provide support and larger amounts compared to those from lower classes, aligning with both international (Albertini and Radl Reference Albertini and Radl2012) and Nordic findings (Fritzell and Lennartsson Reference Fritzell and Lennartsson2005; Hansen and Wiborg Reference Hansen and Wiborg2019). Our additional findings partially support Hypothesis 3b, showing a statistically significant distinction in these social-class differences of financial transfers in 2021 between manual workers and intermediate and upper non-manuals, unlike in previous years. This implies a widening socio-economic gap in the occurrence of financial support for children and grandchildren in Sweden. Hypothesis 3b also stated that the provision of higher amounts would be increasingly more frequent among higher social classes over time, which could not be confirmed. While intermediate and upper non-manuals consistently provided higher amounts more often than manual workers across the years, differences and changes over time were non-significant, possibly owing to small sample sizes. However, Aal social classes, except for manual workers, showed an increasing tendency to provide a higher amount over time. Even lower non-manuals seemed to provide larger sums to younger family members compared to two decades ago. Additional analyses reveal that the social-class gap was primarily driven by higher average contributions from intermediate and upper non-manuals (see Figure S4), reflecting the growing importance of socio-economic status in downward financial investments, as observed in younger populations in both the US (Schneider et al. Reference Schneider, Hastings and Labriola2018) and Europe (Hansen and Wiborg Reference Hansen and Wiborg2019; Kalmijn Reference Kalmijn2024).
Our study adds to existing knowledge by showing that parents aged 77 years or older in the Swedish welfare state display social-class differences in intergenerational financial transfers that seem to increase. These findings could partly be explained by the increasing difference in financial resource availability between social classes in old age (Albertini Reference Albertini2013), which is partly owing to social-class differences in prolonged working life (Fransson and Söderberg Reference Fransson and Söderberg2018) and the enhanced probability of having capital income among the higher social strata in late life (Gustafsson et al. Reference Gustafsson, Johansson and Palmer2009). Results indicate that, as fewer people work in manual occupations (Fritzell and Lundberg Reference Fritzell and Lundberg2007), a growing proportion of older adults will hold non-manual jobs, as confirmed by this study. Additionally, the occupational based part of the pension system has heightened the risk of income inequality in old age, favouring non-manual workers and thereby increasing financial disparities among older adults (Hagen et al. Reference Hagen, Laun and Palme2022).
When younger family members need financial support, older adults from higher social strata provide it to maintain the family’s social status, aligning with the concept of status reproduction (Albertini and Radl Reference Albertini and Radl2012). Consequently, an increasing number of families with older parents will likely uphold and widen the socio-economic gap through financial transfers, contributing to greater inequality among younger generations (Fritzell and Lennartsson Reference Fritzell and Lennartsson2005; Palomino et al. Reference Palomino, Marrero, Nolan and Rodríguez2022).
There is most likely an interaction between gender and social class in financial transfers. Women have increased their education and labour force participation and are thereby changing the gender composition within social classes. In addition, they differ in their giving preferences and family attachments compared to men (Baeriswyl et al. Reference Baeriswyl, Girardin and Oris2022; Kohli and Künemund Reference Kohli, Künemund, Bengtson and Lowenstein2003; Silverstein et al. Reference Silverstein, Gans and Yang2006), which could also change with social-class belonging. This will be important to address in future research in studies with larger sample sizes. It is also essential to further understand what drives changes in both the likelihood of providing financial support and the amounts provided and to disentangle differences between children and grandchildren as receivers. Finally, to fully grasp intergenerational inequalities in Sweden, measuring wealth is essential, as studies without this data, including the current one, may underestimate the extent of transmissions (Hällsten and Thaning Reference Hällsten and Thaning2021).
Strengths and limitations of the study
This study utilises SWEOLD, a nationally representative sample of older individuals with consistently high response rates across all data collection waves. Therefore, the respondents represent the population well, encompassing frail and institution-based individuals. The response rate was lower during 2021 than in previous data collection waves, which could be a result of a generally lower tendency to respond to questionnaires and an increasing fear of fraud. The study’s uniqueness also lies in the availability of information about financial transfers to both children and grandchildren and information on the amounts given. Still, there is a limitation in the crudeness of the values transferred, since we do not have an exact sum and we do not capture those providing less than 5,000–7,000 SEK (i.e. the threshold for transfers in and out of the household). Owing to the small sample sizes, the results concerning the socio-economic differences in transfers should be taken with caution.
A potential limitation of the study is the ten-year interval between each wave of data collection. Another limitation is the challenge of disentangling gender differences in downward intergenerational financial transfers. Although the respondents are answering a question about whether they have provided financial support, there is a possibility that this is a joint decision within the household, which makes it difficult to discern the individual responsible for the decision to give (Andreoni et al. Reference Andreoni, Brown and Rischall2003; Wiepking and Bekkers Reference Wiepking and Bekkers2010). To address this issue, sensitivity analyses were conducted in which the analyses were stratified by the respondents’ living situation. These analyses revealed a more pronounced gender difference in providing financial support among those living alone compared to cohabiting parents, with non-cohabiting men giving more often than non-cohabiting women. These findings align with European research suggesting that individuals living alone, especially women, are less likely providers of financial support (Tur-Sinai and Lewin-Epstein Reference Tur-Sinai and Lewin-Epstein2020). Furthermore, sensitivity analyses indicated a growing tendency among cohabiting women to provide downward financial transfers during the study period, exceeding cohabiting men in the proportion of those providing financial support in 2021.
Conclusion
This study confirms and complements previous research by demonstrating an increase in downward intergenerational financial transfers among older parents in Sweden, in terms of both occurrence and amounts provided. The findings underscore the substantial contribution to grandchildren in this trend, while also shedding light on evolving gender and social-class patterns in these transfers. From a policy standpoint, it is crucial to recognise the scale and the social distribution of intergenerational financial transfers and their impact on societal inequality. With a progressively ageing population and its unequal distribution of financial resources, there is a significant risk of life opportunities being unevenly distributed for future generations.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0144686X24000825.
Acknowledgements
This study was accomplished within the context of the Swedish National Graduate School for Competitive Science on Ageing and Health (SWEAH), which is funded by the Swedish Research Council.
Author contributions
All authors participated in the study conception and design. Isabelle von Saenger performed the statistical analysis and drafted the manuscript. All authors contributed to the interpretation of results and read and approved the final manuscript.
Financial support
The research was supported by funding from the Swedish Research Council (2019-02895), the Swedish Research Council for Health, Working Life and Welfare (2016-07206) and the Marianne and Marcus Wallenberg Foundation (2016.0017). Financial sponsors played no practical role in the execution of this study.
Competing interests
The authors declare no conflict of interest.
Ethical standards
Ethical approval for the SWEOLD study was provided by the Karolinska Institutet Regional Committee for Research (03-413), the Regional Ethical Review Board in Stockholm (2010/403-31/4) and the Ethical Review Agency (2019-06324; 2021-00393; 2021-05031; 2024-03705-02).
Open access
