There’s a comforting assumption built into how most people picture their own old age: whatever social advantage you build up earlier in life — the friend group, the standing dinners, the group chat that never quite dies — should function like savings. Spend down slowly, draw on later, arrive at seventy still reasonably well provisioned. A long-running Swedish research effort, built from more than half a century of the same recurring national survey, suggests that advantage doesn’t function as savings at all. It functions more like a lease that expires right around the time a person might most want to renew it.
The data comes from ten combined waves of two linked Swedish surveys: the Level of Living Survey (LNU), which interviewed nationally representative samples of working-age Swedes in 1968, 1974, 1981, 1991, 2000 and 2010, and the Swedish Panel Study of Living Conditions of the Oldest Old (SWEOLD), which picks up the same population once they age out of LNU at 75, surveyed in 1992, 2002, 2011 and 2021.
Combined, the two datasets span fifty-three years, from 1968 to 2021. Erika Augustsson, Roger Keller Celeste, Stefan Fors, Johan Rehnberg, Carin Lennartsson and Neda Agahi, of Karolinska Institutet’s Aging Research Center, used the full run — 32,136 observations, ages 15 to 97, birth cohorts spanning 1905 to 1984 — for a 2025 paper in the Archives of Gerontology and Geriatrics, tracking how often people reported seeing friends across both age and generation at once.
A rising tide across generations — right up until old age
The first finding runs against the usual pessimism about modern friendship. Each successive ten-year birth cohort was more likely to report frequently visiting or being visited by friends than the cohort born before it, and the gap wasn’t small: adults born in 1975–84 were roughly 23 percentage points more likely to report frequent friend contact than a cohort born forty years earlier, even after the researchers adjusted for differences in education, income, marital status and health across generations. Each generation, in other words, appears to be building denser, more active friendships than the one before it, not thinner ones.
That trend held at nearly every stage of life the researchers could measure it at — with one significant exception. Once people reached old age, the gap between generations shrank dramatically or disappeared altogether. The researchers describe this directly as a convergence: whatever extra social contact a later-born generation had been racking up relative to its predecessors through young adulthood, established adulthood and midlife, that lead mostly failed to still be visible once the same generation aged into its seventies, eighties and beyond.
A generation’s friendships can look sturdier than every generation that came before it, right up until that generation gets old — at which point the sturdiness stops showing up in the data.
Where the lead quietly runs out
The researchers point to fairly ordinary mechanics behind the convergence, not anything mysterious. Retirement removes the daily workplace contact that quietly did a lot of the work of “staying in touch” without anyone having to schedule it. Friends within a generation tend to be roughly the same age, so bereavement and declining health start thinning a network from multiple directions at once as a cohort ages together, rather than one loss at a time. And reduced mobility, whether from health or from no longer driving, makes the low-effort contact that used to happen by default — the coffee that happened because two people worked two floors apart — much harder to replicate on purpose. None of these forces care how socially rich a generation’s earlier decades were. They apply once retirement, bereavement and reduced mobility start, roughly regardless of where a person started from.
It’s worth being precise about what the researchers can and can’t yet say. The convergence pattern is clearest among the generations who have already lived all the way through old age inside this fifty-three-year dataset — people born up through the 1940s and 50s. The most recently born cohorts in the study, the ones currently showing the largest friend-contact advantage of any generation on record, haven’t reached old age yet within the available data. The researchers are explicit that it remains an open question, not a confirmed finding, whether these cohorts will eventually converge the same way older generations did, or whether they’ll finally be the ones to carry a midlife advantage into old age.
A gap between expectation and reality, not just a decline
The paper’s own closing concern is less about the raw numbers than about what they might do to expectations. Because each generation has grown used to more frequent friend contact than the one before it, the researchers raise a specific worry: if that contact converges downward in old age the way it has for every generation studied so far, future cohorts of older adults may experience a steeper perceived drop than their parents did, arriving at old age expecting a level of friendship contact that generations before them never actually got to keep either. Separate research the authors cite suggests that expectations for social contact predict loneliness more strongly than actual contact levels do — which means the risk here isn’t just fewer visits from friends. It’s a widening gap between what a generation has come to expect and what old age, so far, has consistently delivered.
If reading this lands closer to home than it does as an abstract statistic, that’s worth paying attention to on its own terms. Loneliness in later life is common, not a personal failing, and it’s worth raising with a friend, a doctor, or a therapist rather than carrying alone.