Where did all my friends go after I retired?

You didn’t imagine it. Somewhere in the second year the phone got quieter, and the people you spent thirty years sitting beside turned into people you used to know.

What’s strange is that you can’t point to when it happened. There was no falling-out and nobody moved away. There was a leaving do, and then a stretch of lunches, and then at some point the lunches stopped being arranged.

Here’s the part that surprises most people: you’re probably wrong about what you lost.

Your friends didn’t go anywhere

The best study of this question followed more than five thousand Dutch workers through the year they retired and out the other side — three waves between 2015 and 2023, with each person compared against their own earlier answers rather than against anyone else’s. That design matters. It isn’t asking whether retirees see people less often than workers do; it’s asking what happened to you when you stopped going in.

Contact with friends went up. Not down. Contact with neighbours went up by more than any other relationship measured. Contact with brothers and sisters, with children, with grandchildren — all up a little.

So if the question is where the friends went, the honest answer is that most of them are still there, and you are probably seeing them more often now than you did when you were working.

What emptied wasn’t your friendships. It was the workplace.

The colleagues go — and they go slowly

In the same data, contact with former colleagues also rose when people retired. For a while, everyone kept it up. And then it came apart.

Not gone on your last day. Gone about two years later — which is exactly why it never registered as a loss. Nothing was taken from you on any particular Tuesday. It just quietly stopped being replaced.

Worth knowing that the older research on this disagreed about the speed. Studies through the 2000s and 2010s generally found workplace ties cut promptly at retirement, not two years later. Nobody disputes the direction. The argument is about whether it takes months or years, and either way it happens.

The delay is the disorienting part

For that first stretch, everything holds. There are lunches. There are messages about who’s left and who’s taken over your old accounts. Someone’s in town and you get a drink. All of it feels like proof that the friendships were real and that they travelled.

They were real. What they weren’t was self-sustaining.

What seems to have been generating them was thirty years of unplanned proximity — the same building, the same hours, the same people, whether or not anyone had the energy to arrange anything. You didn’t maintain those friendships. The building did. When the building went, they kept running on stored momentum, and stored momentum runs out at about the two-year mark.

That is a reading of the evidence rather than a proven mechanism — the study measures the fade, not the reason for it. But the strongest hint sits in the same table. The relationship that grew most after retirement was neighbours. Proximity is what replaced proximity.

Two groups the average doesn’t describe

Everything above is an average, and two sets of people sit outside it. Both are the opposite of what you’d guess.

If you’re on your own, you gain the least. You would expect a person with no partner at home to have the most room for other people. In this data the increase in contact with both friends and former colleagues was weaker for people who weren’t partnered, not stronger.

And keeping up with work is largely something women do. Among the men in this study, the rise in contact with former colleagues after retirement wasn’t statistically distinguishable from no rise at all. The whole effect was carried by women. If you’re a man who assumed you’d stay in touch with the people from work, it isn’t a character flaw that you didn’t — on average, the men in this study didn’t either.

Why “just call people” doesn’t work

Because it asks intention to do a job that proximity was doing. Every week you have to decide to call, and some weeks you won’t, and the friendship degrades at the rate of the weeks you don’t.

What worked before was a container: the same people, the same time, recurring, running whether or not you felt like it. That’s what a workplace is. A standing Thursday is a container. A promise to get lunch sometime is not.

The researchers themselves suspect this is where the real action is. Their reading of their own results is that the larger social changes in retirement probably happen through new contacts formed in the activities people take up, rather than through existing friendships being maintained harder.

It’s groups, not people

There’s a second finding worth having, from a British study that followed 424 people through retirement against 424 matched people who didn’t retire, and checked on all of them six years later. What predicted how well the retirees did wasn’t how many friends they had. It was how many groups they belonged to — a choir, a club, a congregation, a committee — and specifically the ones they held onto or joined after they stopped working. The researchers put the size of the effect on a par with regular exercise.

Treat that as a direction rather than a dose. It’s an observational study, its authors say plainly that it can’t establish cause, and the relationship almost certainly runs both ways — being well enough to join things is part of what gets you into them. What it does line up with is everything above: a group is a container, and a container is what a workplace was.

When it’s more than missing people

None of this makes loneliness rare. It’s common in this age group, whether or not anyone has retired.

And it isn’t only a matter of comfort. A large meta-analysis of social relationships and mortality found social isolation carries a risk on the scale of other health risks people take seriously, including smoking.

But notice what the retirement data does not say. It doesn’t say retirement makes people lonely on average — on average, people see more of others, not less. What it says is that the increase isn’t evenly distributed, and the people who don’t get it are identifiable in advance. That’s a better warning than a blanket one, because you can act on it.

Where to start

Not with a list of clubs. With an honest look at which containers you have left.

Work was supplying a great deal more than an income, and the social side of it is the part almost nobody plans for — partly because it doesn’t fail immediately. It degrades quietly, on a two-year delay, in the one area where you have no instrumentation and no advisor asking about it.

Social Life is one of the five PHASE™ dimensions, alongside Purpose, Health, Activities and Everyday Life, and it’s consistently the one people underestimate most. The PHASE Into Retirement™ assessment takes about fifteen minutes and tells you which of the five is thinnest for you rather than leaving you to guess. You can see a complete sample report first — free, no signup.

Your friends are still there. What you’re missing is the building that used to put you in a room with them.