The friendship recession.

In May 2021, the Survey Center on American Life asked 2,019 American adults how many close friends they had, not counting relatives. The question was not new. A national survey had asked Americans the same thing in 1990, and the comparison is what made the results travel. In 1990, 55 percent of men said they had at least six close friends. Thirty-one years later the figure was 27 percent. The share of men who said they had no close friends at all went from 3 percent to 15 percent.

Daniel Cox, who directs the center, gave the pattern a name in an essay that summer. He called it a friendship recession, and the phrase has stuck because it describes something most men over thirty recognize without needing the chart. The group chat still exists. The wedding party has scattered across four time zones. The last time anyone spoke on the phone for an hour was a crisis. Nobody decided to stop having friends. It happened the way a savings account drains: by small withdrawals that never felt like decisions.

The decline was not only among men. Across the whole sample, 49 percent of Americans in 2021 reported three or fewer close friends, compared with 27 percent in 1990, and 12 percent reported none. But the details of the men’s responses are the more striking part. Only 21 percent of men said they had received emotional support from a friend in the past week. Only a quarter said they had told a friend they loved him in that time. Among young men, 36 percent said their parents were the first people they turned to with a personal problem. Among unmarried men who were not in a relationship, one in five reported no close friends.

Two years later, in May 2023, the United States Surgeon General, Vivek Murthy, issued an advisory titled Our Epidemic of Loneliness and Isolation. It is an unusual document, because it treats a feeling as a public-health hazard and backs the claim with mortality data. Murthy’s office reported that about half of American adults showed measurable levels of loneliness before the pandemic. It cited research finding that loneliness and social isolation raise the risk of premature death by 26 and 29 percent respectively, and it compared the mortality effect of weak social connection to smoking up to fifteen cigarettes a day. It also reported a 29 percent higher risk of heart disease and a 32 percent higher risk of stroke among the poorly connected.

The most concrete figure in the advisory came from the American Time Use Survey. In 2003, Americans spent about sixty minutes a day in person with friends. By 2020 the figure was twenty minutes. For people aged fifteen to twenty-four, the drop was from roughly 150 minutes a day to 40, a decline of nearly 70 percent. The pandemic sharpened the fall, but the slope was already there. A generation that grew up with a controller in hand, often playing in the same room as a brother or a neighbor, has been spending less and less time in the same room as anyone.

The explanations offered by Cox and others are dull, which is a reason to believe them. Men marry later than they used to, and for many men a wife or partner has quietly served as social secretary, keeper of the calendar and the person who remembers birthdays. Fewer men attend religious services, and a congregation was once one of the few places an adult could meet the same people every week without trying. Work has changed as well. People switch jobs more often, many work from home, and hours have crept upward for professionals. The Survey Center found that 54 percent of Americans with close friends had met at least one of them through their own or their spouse’s workplace, so a remote job removes a reliable supply.

There is also the matter of how men tend to be friends. Researchers have long observed that male friendship is often organized around a shared activity rather than conversation, the friendship of the pickup game, the fishing trip and the long drive. That pattern is not a deficiency. It does mean that when the activities vanish, the friendships have nothing to stand on. A man who stops playing in the Tuesday league does not usually replace it with a Tuesday phone call. He replaces it with nothing, and the friendship becomes a thing he remembers rather than a thing he has.

Time is the other constraint, and it is less flexible than affection. Jeffrey Hall, a communication researcher at the University of Kansas, published a study in 2018 in the Journal of Social and Personal Relationships estimating how long it takes to turn an acquaintance into a friend. His figures were about 40 to 60 hours for a casual friend, 80 to 100 hours for a friend and more than 200 hours for a close one. “You can’t snap your fingers and make a friend,” he said at the time. Two hundred hours is roughly five weeks of full-time work. College provided those hours for free. A first job often did too. After thirty, nothing provides them unless someone schedules them.

It is tempting to blame the phone for all of it, and the advisory does ask readers to examine their relationship with technology. The honest answer is that the evidence runs in both directions. Pew Research Center reported in 2024 that 47 percent of American teenagers who play video games said they had made a friend online through gaming. Many men in their thirties and forties keep up old friendships through a headset on a Thursday night, and those friendships are real. What a screen does less well is supply the unplanned hour, the dinner that runs long or the ride home that turns into the conversation that mattered. Contact can be maintained remotely. Depth is harder to build that way, and depth is what the surveys are measuring when they ask who someone would call in a crisis.

The Surgeon General’s advisory ends with recommendations for governments, employers, schools and technology companies, and they are sensible. It proposes investing in what it calls social infrastructure, such as parks, libraries and community programs, and it asks platforms to design for connection rather than for time spent. Those changes will take years. The individual version is simpler and harder, because it asks for a change in routine rather than a change in policy.

The first part of it is recurrence. Friendships among adults are rarely made by a single great evening. They are made by the same people in the same place at the same time, repeatedly, until familiarity turns into trust. A standing game night, a monthly dinner with a fixed date, a weekly run with the same two people: these are unglamorous structures, and they work for exactly that reason. The men’s shed movement, which began in Australia as a health initiative for older men and now has more than 900 sheds there, adopted the motto “Shoulder to Shoulder” on the observation that many men talk more easily beside each other than across a table. The workbench does the organizing, and the conversation follows.

The second part is initiative. In most lapsed friendships, both men are waiting for the other to call, and both would be glad to hear from him. The cost of a text proposing a specific date is a few seconds and a small risk of silence. The cost of not sending it compounds, and the Survey Center’s numbers are what that compounding looks like across a country.

The third part is admitting that the thing matters. Only a quarter of men told a friend in the past week that they loved him, and nobody needs to start with that sentence. It is enough to say out loud that the Thursday call is important and to keep it when a work deadline argues otherwise. Friendship among grown men has always depended on some excuse for being together: a war, a factory floor, a team or a parish. Most of those excuses have gone. What remains is the decision to make one on purpose and to keep showing up for it.