The most recent figures in our data, for the year 2019, show that the vast majority of parents had at least one adult child living within a distance of 50 km, at age 80 (Fig. 2). The figure for mothers is 86.0% and for fathers 82.0%. At age 65, the figures are rather similar, with 83.8% of women and 80.6% of men living within 50 km of their nearest child. We also found that many older parents had at least one adult child living very nearby, within 1 km. At age 80, this figure is 27.7% for mothers and 23.8% for fathers, while at age 65 the equivalent figures are 29.0% and 26.1%, respectively.

Mothers and fathers aged 65 and 80, percentage with nearest child within 1 km, 1–5 km, 5–25 km, 25–50 km and > 50 km, in 2019

Since we focus on the consequences for older men and women of not having adult children nearby, we also want to highlight the even more vulnerable group of older people without any adult children. In 2019, 15% of all 65-year-old women and 20% of all 65-year-old men, as well as 12% of all 80-year-old women and 15% of all 80-year-old men, had no adult children over the age of 25, as illustrated in Fig. 9 in Appendix. This category includes both the childless and those who still have children under 25, and mothers and fathers whose children do not live in Sweden.

When looking at the overall trend in intergenerational proximity for the whole period, we found large variations in the median distances over time, as illustrated in Fig. 3, but also a considerable increase in the median distances to the nearest child for people aged 65 from the 1990s to the 2010s for both men and women. For fathers, the median distance increases from 3 km in the mid-1990s to 5 km in 2007, and a substantial increase was also observed for women. Moreover, we also found a slight decrease in distances for both genders during the last years of the 2010s. However, for mothers aged 80, we noticed almost the opposite pattern, with decreasing distances until 2011, followed by a slight increase. For fathers aged 80, the trend is less pronounced, but a noticeable increase was found from 2010, an increase that mirrors the trend found 15 years earlier, when these cohorts reached the age of 65. These figures indicate that we can foresee an extensive increase in intergenerational distances for cohorts approaching 80, in the coming decades.

Median distance in km to nearest child for mothers and fathers aged 65 and 80, 1990 −2019

The results from the logistic regressions for the whole period 1990—2019 with year of observation as one of the independent variables, do not show any distinct trend, only a minor increase in the likelihood over the three decades to have an adult child within 50 km at age 80 for mothers (B = 0,021, p < 0,0001Footnote 1) and even smaller for fathers (B =  = 0,008), and at age 65 for mothers (B = 0,001, p < 0.001) and for fathers a somewhat larger increase in the likelihood for fathers (B = 0,004), when controlling for covariates (see Figs. 10 and 11 in Appendix).

To further analyse the trends over time, we performed logistic regressions for the whole period, 1990–2019 (see Figs. 12, 13, 14 and 15 in Appendix), with the three time periods (1990–1999, 2000–2009, 2010–2019) included as independent variables as well as other covariates. These analyses revealed results that are highly consistent with the descriptive findings on trends over time also when including the control variables in the analysis and show that fathers who reached the age of 65 in the 1990s, were more likely to have their nearest child living within 50 km compared to the 2000s (B = −0.091) and the 2010s (B = −0.102), while for mothers no significant difference over time was found. However, the likelihood to have a child within 1 km decreased from 1990 to 2000s, for both mothers (B = −0.045) and even more for fathers (B = −0.121), and to the 2010s, for mothers (B = −0.049) and for fathers (B = −0.106).

We found opposite results for parents aged 80, with substantially higher probabilities of having a child living within 50 km for mothers in the 2000s (B = 0.241) and especially in 2010s (B = 0.415) and also for fathers, although less pronounced, in the 2000s (B = 0.161) and in the 2010s (B = 0.171), as well as within 1 km, during the latter period than during the first.

Figure 3 also reveals gender differences and, consistent with previous studies, we found on average longer distances for fathers. Additionally, we observe gender differences in the trends, with changes first appearing among fathers, due to the larger age gap between fathers and their adult children. Thus, the fathers in both age-groups – 80 and 65 – have younger children than women the same age groups, and the distances are therefore influenced by different cohort-specific migration patterns in the children’s generation.

Figure 4 shows the percentage of mothers and fathers whose nearest child lives within 1 km, 1–50 km and further than 50 km. The patterns revealed are in line with the observed variations in distances found in the figures on median distances. The trend over the entire period is a decrease in the percentage whose nearest child lived further than 50 km away, for fathers and especially for mothers aged 80. However, since 2010, there has been a slight increase in this percentage, accompanied by a similar decrease in the percentage whose nearest child lived within 1 km. For both fathers and mothers aged 65, the percentage whose nearest child lived further than 50 km away had increased, while simultaneously the percentage whose nearest child lived within 1 km had decreased. These figures illustrate that the percentage of parents at age 65 with adult children living close by began to decrease around year 2000, when the parents born in the late 1930s reached 65, and arrived at its lowest point around 2010, when the cohorts born in the 1940s reached 65. Furthermore, the figures for parents aged 80 show a decline in the percentage with adult children living close by, starting around 2015, when the cohorts born in the late 1930s reached 80. Hence, the same cohort who had fewer children close by at age 65 also had fewer children nearby at age 80. This illustrates that the distances to adult children only change marginally for older adults as they age. Moreover, the figures reveal that the patterns established, through migration earlier in the family life course, prevail when the parents are getting older.

Percentages of mothers and fathers aged 65 and 80 with nearest child within 1 km, 1–50 km, and further than 50 km, 1990–2019

In summary, these figures do not reveal any long-term trends of shorter or longer intergenerational distances but demonstrate substantial fluctuations over time from cohort to cohort, reflecting previous, cohort-specific migration patterns. Hence, the results also indicate that the preconditions for physical contact between parents and adult children have varied, and will continue to vary, over time. For the Swedish baby-boom cohort born in the 1940s, and now at an age when care is increasingly needed, distances to adult children are longer than for cohorts born earlier.

In the logistic regressions, for the period 2010–2019 (see Figs. 5, 6, 7, 8, and Table 3 in appendix), we found a substantially higher probability for parents aged 80 living in metropolitan regions to have at least one child living within 50 km compared to those living in other urban areas, for mothers (B = −1.22) and fathers (B = −1.24), and especially compared to those residing in rural areas, for mother (B = −1.76) and fathers (B = −1.83). However, for parents aged 80 living in metropolitan areas the likelihood of having a child within 1 km is slightly lower compared to those living in other urban areas, especially for mothers (B = 0.11) but also for fathers (B = 0.04) or living in rural areas, for mothers (B = 0.17) and for fathers (B = 0.09). If living in the same town or village as an adult child, it is more likely that those aged 80 will live at a very short distance from them. However, for parents aged 65, we found similar results across all distances, with a much higher probability in metropolitan areas of having the nearest child living within 50 km than in urban areas, for mothers (B = −1.38) and for fathers (B = −1.18) and in rural areas, for mothers (B = −1.99) and for fathers (B = −1.76). Similar results were found for having a child within 1 km, with a lower probability in urban areas, for mothers (B = −0.17) and for fathers (B = −0.17), and an even lower probability in rural areas, for mothers (B = −0.23) and for fathers (B = −0.22).

Logistic regressions, having at least one child within 50 km at age 80, 2010–2019, coefficients and CI

Logistic regressions, having at least one child within 1 km at age 80, 2010–2019, coefficients and CI

Logistic regressions, having at least one child within 50 km at age 65, 2010–2019, coefficients and CI

Logistic regressions, having at least one child within 1 km at age 65, 2010–2019, coefficients and CI

Moreover, in the analyses for the most recent period 2010–2019 (see Figs. 5, 6, 7 and 8), we found a strong association between parents’ education and nearness to an adult child. There was a notably lower probability of having an adult child within 50 km among those with secondary education compared to those with a primary education, at age 80 for mothers (B = −0.43) and for fathers (B = −0.53), with the lowest probability for those with university education, for mothers at age 80 (B = −1.02) and for fathers (B = −1.10). These results are more or less the same for parents in both age groups, 65 and 80, and for the probability of having a child within 1 km. These results are in line with findings from previous research and are expected outcomes, because the children of the well-educated are more likely to enrol in higher education themselves and thus more likely to move away from their parents and end up living farther afield. For men, the effect is marginally stronger among parents aged 80, but at age 65 the opposite gender difference was found.

In line with the findings on education, we noticed a higher probability that parents with low income at age 80 would have a child within 50 km, compared to the higher income quartiles, especially in comparison to the highest income quartile, for mothers (B = −0.45) and for fathers (B = −0.39). Similar results were found for having an adult child within 1 km. However, at age 65 the association was apparent only for having a child within 1 km. Furthermore, we found some gender differences, but since we have defined income using the individual share of the total household income, the observed gender disparities result from income differences between men and women living alone.

When comparing married parents to those who never married, or were divorced or widowed, we found a rather dispersed pattern, with substantial gender differences. At age 65, married fathers were more likely to have a child within 50 km than fathers who had never married (B = −0.17) and especially compared to those who were divorced (B = −0.46). Similar results were found for having a child within 1 km for fathers at age 80, with lower probability for those never married (B = −0.34) and those who are divorced (B = −0.63).

For women at age 65, the pattern is almost the opposite, with married mothers being less likely to have a child within 50 km, compared to mothers never married (B = 0.12) or widowed (B = 0.17), but not compared to those who are divorced (B = −0.13). For mothers at age 80, we found it less likely for those married to have a child within 50 km than for those who were widowed (B = 0.36). But when comparing the married to those never married or divorced, no significant results were found. And similar patterns are found for having a child within 1 km both for women at age 65 and 80. However, we also found a significant higher probability for never married mothers at age 80 to have a child within 1 km (B = 0.20).

In the logistic regression with interaction variables, region type and periods, we found that for mothers aged 65, the probability to have an adult child within 50 km is lower for those living in urban and especially rural areas in the second and especially in the third period, meaning that the regional gradient increases over time (see Figs. 16 and 17 in Appendix).

For men and women aged 80, living in less urbanised areas, we found that the probability of having an adult child living within 50 km is lower in the third period and for men also in the second period, compared to the first. We observed a less distinct patterns when analysing different groups by education level. However, for women aged 80 with a university education we also find a significant lower probability of having an adult child within 50 km in the third period. While we found increasing distances among people with only primary education in the descriptives, this effect disappeared when we controlled for other covariates, including region type.

Regarding the influence of sibship composition among adult children, logistic regressions confirm the result from descriptive analyses and show that mothers and fathers with only daughters were less likely to have a child within 50 kms compared to those with mixed sibships or only sons. As expected, the number of children was strongly associated with having at least one child living nearby (see Table 3 in appendix).