Many believe that happiness rebounds after midlife - a view supported by the widely cited "U-shaped curve of well-being" popularized by the book The Happiness Curve: Why Life Gets Better After 50. But a new study suggests the story is more complicated. Rather than a universal rise in mental health after age 50, researchers found that the rebound is concentrated among one group in particular: unemployed men.
To investigate, scientists from USC Dornsife College of Letters, Arts, and Sciences, working with colleagues at Vrije Universiteit Amsterdam, the University of Amsterdam, and Erasmus University Rotterdam, analyzed data on unemployed men 50 and older across 10 European countries.
The findings, recently posted online ahead of publication in the Journal of Labor Economics and presented at the IZA Institute of Labor Economics, reveal a stark mental health gap. At age 50, when most men are still working, unemployed men were 23 percentage points more likely to report symptoms of depression than their employed peers. For context, the study suggests unemployed men in midlife were more than twice as likely to report symptoms of depression as those who had lost a spouse.
By age 65, however, when most men are retired and the societal expectation to work has faded, that mental health gap disappears entirely.
"It's a striking shift," said Titus Galama, director of the Center for the Study of Human Capital based at USC Dornsife's Center for Economic and Social Research (CESR). "It shows how powerful social expectations around work are - and how much mental health can improve when those expectations lift."
Work norms - not biology - fuel the midlife slump
Mental health improves for unemployed men as they approach retirement - not because of aging or more leisure time, but because of shifting social expectations. In midlife, joblessness often carries stigma. But as more peers stop working, that stigma fades and mental health improves.
The study suggests that the U-shaped curve of well-being may not be a universal feature of aging, but rather a psychosocial phenomenon shaped by social roles and identity. For unemployed men, the expectation to work appears to weigh more heavily on mental health than financial hardship.
Galama and his co-authors, including lead author Coen van de Kraats of Erasmus University Rotterdam, weren't initially studying retirement norms. They were analyzing trends in health, mental well-being and socioeconomic factors over the course of life when they noticed something unexpected: a distinct U-shape in the mental health of unemployed men - with a sharp decline in midlife followed by a rebound near retirement.
That pattern may help explain why the well-being curve appears in some datasets but not others. The rise in mental health after age 50, the researchers argue, is driven largely by this relatively small subgroup. When averaged into the general population, their improvement can create the illusion of a broader, universal rebound.
Earlier theories suggested the rebound was biological - linked to brain chemistry or how people report emotions as they age. But the researchers argue it's psychosocial. The mental health recovery in later life appears to stem from easing social expectations around work, not biology or leisure time.
"Being jobless in midlife isn't just about lost income," van de Kraats said. "It's about lost identity."
How they tracked men's mental well-being
To understand what drives shifts in mental health, the scientists analyzed data from the Survey of Health, Aging, and Retirement in Europe (SHARE), a long-running study that tracks health, economic status and social trends across the continent. They measured symptoms of depression using the EURO-D scale, a clinically validated tool.
The team also took advantage of a natural experiment created by differences in retirement laws across 10 European countries, age groups and time periods. These policy variations allowed them to isolate the effects of retirement expectations from other factors such as age or income.
Their findings were striking: Mental distress among unemployed men dropped by 18 percentage points in the year after reaching their country's early retirement age - and by 37 points five years later. This improvement wasn't tied to better finances, health or increased leisure time. It coincided with a shift in self-perception: As more peers retired, these men began to see themselves as retired, too.
"As society's expectations shift with age," van de Kraats said, "some men find emotional relief not by changing their employment situation, but by no longer feeling they've failed to meet it."
By contrast, the team found little or no mental health improvement among employed men, unemployed women, or those receiving disability benefits - suggesting the effect was unique to unemployed men redefining themselves as retired.
Rethinking the happiness curve
To test their theory, the researchers compared their European findings to data from the U.S. Health and Retirement Study. The American results echoed the European ones, reinforcing the idea that the so-called U-shaped curve of well-being may not be universal.
Because the curve reflects population averages, it can mask meaningful differences across subgroups. In fact, the dramatic mental health gains among unemployed men after age 50 skew the average upward, creating the impression of a general rebound.
Next steps in studying happiness and aging
"I am excited that our analyses have led to a theory that can be empirically tested," Galama said.
Next, the researchers plan to explore whether unemployment affects mental health differently in countries with high versus low unemployment rates, or where formal retirement institutions are absent.
They also plan to expand their research to other groups, including younger unemployed adults, to determine whether shifting work norms contribute to rising levels of distress.
One area of particular interest: the potential connection between work-related social pressure and rising rates of suicide and substance abuse among unemployed men, a phenomenon known as "deaths of despair." The team plans to examine this trend more closely in the U.S.
About this study
In addition to Galama and van de Kraats, the study's co-authors include Maarten Lindeboom at Vrije Universiteit Amsterdam, Monash University, the Tinbergen Institute and the IZA Institute of Labor Economics as well as Zichen Deng at the University of Amsterdam.
The paper is available at IZA (Discussion Paper No. 17586, December 2024) the Tinbergen Institute (2022), and the Center for Health Economics at Monash University (2023-2025).
The research was supported by the U.S. National Institute on Aging, the European Union's Horizon 2020 program and the Dutch National Science Foundation.
