Retirement appears to reshape men's and women's social lives in opposite directions. Australian data found retired men spending less time with friends and relatives outside the home, while retired women spent more — possibly because men lose more of their social world when they leave work.

Retirement removes more than work. It also removes the people a working day places within reach without anyone having to make a plan: the colleague at the next desk, the familiar face in the kitchen, the person who always asks whether you watched the match.

An Australian analysis found a striking gender split in what happened to social time outside the home. Retired men recorded less time with friends and relatives than non-retired men, while retired women recorded more than non-retired women.

This is one study, not settled consensus. It compared groups in Australian surveys from 2006 rather than following the same individuals through retirement, so the result is a finding from that dataset, not a universal rule about men and women.

What the researchers counted as social time

Roger Patulny’s analysis for the Australian Institute of Family Studies used two national datasets collected by the Australian Bureau of Statistics: the 2006 General Social Survey and the 2006 Time Use Survey. They captured related but distinct parts of social life.

The General Social Survey asked how often respondents had face-to-face contact with friends or family who did not live with them. The Time Use Survey asked people to keep diaries, allowing researchers to calculate how many minutes they spent in the presence of friends and relatives outside the household, household family, people such as colleagues and neighbours, or nobody else.

For this project, the General Social Survey contributed 5,833 people aged 45 to 74. The time-use analysis covered 5,959 diary days in the same age range, with each person generally completing two days. Most models defined retirees as people aged 65 or older who were not in the labour force; another version included younger respondents who described themselves as retired.

That is broader than a clean before-and-after account of leaving a career. Retirement can be planned or forced, early or late, financially comfortable or constrained. A person counted as working at 64 may also differ from one counted as retired at 67 in ways no regression can fully remove.

The diary figures moved in opposite directions

Among non-retired respondents, men and women recorded similar time with friends or family outside their household: 70.2 minutes a day for men and 75 minutes for women.

Retirement is where the figures separated.

Retired men recorded 53.3 minutes a day with outside-household friends and relatives, about 17 minutes less than non-retired men. Retired women recorded 102.6 minutes, nearly 28 minutes more than non-retired women and almost twice the retired-men figure.

The gender-and-retirement interaction remained after the researchers introduced demographic controls. It was also visible when they looked separately at partnered, single, and separated, divorced or widowed respondents. Marital status had a substantial relationship with social time, but it did not make the gender split disappear.

Other diary categories complicate any easy story about outgoing women and withdrawn men. Retired men reported nearly 15.7 hours a day in the presence of household family, up from about 13 hours among non-retired men. Retired women recorded more outside-household social time, but also more time alone, rising from roughly 4.2 to 6.4 hours.

The diaries describe a redistribution of time, not two personality types.

How often and how long are different questions

The other national dataset did not produce the same dramatic reversal. On the General Social Survey’s frequency scale, women had more face-to-face contact than men regardless of retirement status. Retirees of both genders also appeared to have somewhat more frequent contact than their non-retired counterparts after adjustment.

The gender-by-retirement interaction emerged clearly in the duration recorded by the time diaries, not in recalled frequency. A person can meet several people briefly and still accumulate little time with them. Someone else may have fewer encounters that last for an afternoon.

Neither measure tells us whether the company was wanted, whether the conversation was intimate or whether the relationship felt supportive. The study measured contact, not friendship quality or loneliness. Treating those concepts as synonyms would write more into the data than the surveys collected.

The work-network explanation is plausible, not proven

The authors suggested that men may depend more heavily on work-based networks for social contact outside the home. The diary categories offer a reason for that interpretation. Before retirement, men recorded 209 minutes a day with “non-family/friends”, a category that included colleagues, neighbours, housemates and other people’s children. Women recorded 148.6 minutes. Among retirees, those totals fell to 51.3 minutes for men and 45.1 for women.

Work may therefore have supplied more of men’s incidental contact. Yet the category did not separate a close colleague from a stranger in the same building, and the analysis did not ask whether friendships from work survived. The explanation remains a reading of the pattern rather than a tested mechanism.

Still, the social machinery of a job is easy to underestimate. It creates recurring contact at almost no organisational cost to the individual. After leaving, a conversation that once happened automatically may require a message, a mutually free afternoon, transport and a place to meet.

This overlaps with a recent Silicon Canals article on retirement and career identity. A job title can carry status and purpose, but the institution around it also supplies names, routines and small acknowledgements. Those functions often become visible only when the structure is gone.

The study cannot show that retirement caused the split

The 2006 surveys were cross-sectional. They compared retired people with non-retired people at one point rather than observing an individual’s social life before and after leaving work. Age, health, income, location, the reason for retirement, personality and the era in which someone grew up could all influence both retirement status and social habits.

Patulny also noted possible inaccuracies in diary completion, the difficulty of separating retirement from age and low explanatory power in the models. The surveys used binary gender categories and describe Australian life two decades ago. Messaging apps, remote work, later retirement and changing household roles may alter how the same question looks now.

Later evidence makes the picture less tidy. An Australian National University doctoral analysis of longitudinal HILDA data found that men and women who retired during the observation period increased their frequency of social activities, while people who were already retired tended to decline over time. The result used a different measure and design, but that is precisely why it matters.

A 2024 longitudinal comparison of Australia, China and the United States offers another caution. In Australia, retirees had a slightly higher predicted prevalence of loneliness than non-retirees, but the difference was not statistically significant; voluntary retirees had the lowest loneliness and involuntary retirees the highest. Country, choice and measurement changed the pattern.

The missing item in retirement planning

Retirement is usually planned as a financial event because money can be counted. The social infrastructure of work is harder to put on a spreadsheet, even when it has been quietly organising much of a person’s week.

A 2024 Australian Institute of Family Studies report still described retirement as an important transition point. It also cited newer HILDA evidence that older men were more likely than older women to report unsupportive relationships and community disengagement. Those averages do not make men naturally unsociable or women naturally gifted at connection. They suggest that paid work, household roles and learned expectations can distribute the labour of maintaining relationships unevenly.

The 2006 finding is most useful as a question to carry into newer research: when a workplace disappears from someone’s life, which relationships remain without it?