How unsuspecting economists (and apes) found the happiness curve
No one went looking for a happiness curve, because no one expected to find one.
The idea of a difficult emotional passage in midlife, followed by growing tranquility in older age, is not new, as any visitor to the National Gallery’s Thomas Cole exhibit will attest. But the idea of a midlife crisis as such, something unique and specific that happens to people in middle age, is surprisingly new. The phrase itself dates back only to 1965, when a Canadian-born psychoanalyst named Elliott Jaques coined it for an article titled “Death and the Midlife Crisis,” published in the International Journal of Psychoanalysis. Perhaps not surprisingly, Jaques was then in his late forties.
“I first became aware of this period as a critical stage in development,” Jaques wrote in the article, “when I noticed a marked tendency towards crisis in the creative work of great men in their middle and late thirties.” With the peculiar combination of shrewd insight and reckless conjecture which is so characteristic of the psychoanalytic method, Jaques expanded his observation into a general theory. “The midlife crisis,” he wrote, “is a reaction which not only occurs in creative genius, but manifests itself in some form in everyone.”
A reaction to what? Beginning in the middle thirties, Jaques hypothesized, the fact of death, heretofore abstract and distant, becomes concrete and personal. “The individual has stopped growing up, and has begun to grow old. A new set of external circumstances has to be met. The first phase of adult life has been lived. Family and occupation have become established (or ought to have become established unless the individual’s adjustment has gone seriously awry); parents have grown old, and children are at the threshold of adulthood.” The implications of one’s impending end can no longer be suppressed or evaded. This, he wrote, “is the central and crucial feature of the midlife phase.”
What followed was a fusillade of Freudian jargon that was already going out of style in the 1960s. (“The inner world is unconsciously felt to contain the persecuting and annihilating devoured and destroyed bad breast, the ego itself feeling in bits.”) What it boiled down to, though, was that in middle age we are forced to resign ourselves to all that we cannot be or do. “Important things that the individual would have liked to achieve, would have desired to become, would have longed to have, will not be realized. The awareness of on-coming frustration is especially intense.”
More than half a century later, a lot of scientific water has flowed under the bridge, and there is a lot to say about what Jaques got right and wrong, a topic I’ll come back to in later chapters. Suffice to say, for now, that the popular culture seized upon Jaques’s basic concept, or at any rate his pithy phrase, and ran with it, quickly transforming it into a cliché.
But psychologists could not find evidence that he was right. Sure, some people have difficulties in midlife; but that is hardly an interesting finding. When psychologists went looking for something special about the middle years, they came back empty-handed. “Objectively, scholars don’t find anything very different going on,” Elaine Wethington, a sociologist and professor of human development at Cornell University, told me. In 2000, in an article titled “Expecting Stress: Americans and the ‘Midlife Crisis’” (published in the journal Motivation and Emotion), she found that about a quarter of people she randomly surveyed reported having “personal turmoil and sudden changes in personal goals and lifestyle, brought about by the realization of aging, physical decline, or entrapment in unwelcome, restrictive roles.” (Women and men were equally likely to report a midlife crisis.) Yet, she noted, “Epidemiological study of psychological distress in adulthood does not suggest that midlife is a time of out-of-the-ordinary distress, for either men or women.” People often said they experienced a midlife crisis, but Wethington suspected that that’s because they often go through difficult stuff in and around their forties. “There are people who believe that the midlife crisis exists,” she said. “It’s a term that makes sense to them in respect to their biographies. They perceive that somewhere in a very long range of midlife, they indeed had a crisis.” And so the idea of midlife crisis, regardless of its scientific validity, might be useful to people as a way to think about their lives. “If we believe it is real, it is real and has consequences,” Wethington told me. But it’s more like folk wisdom or a social convention than solid science.
Others in psychology have little patience for the idea of a distinctive crisis in midlife. “For the most part we didn’t see evidence of it in the data,” Carol Ryff, the director of the Institute on Aging at the University of Wisconsin, told me. In any case, like many psychologists, she sees little point in generalizing about the whole of life or all of humanity, because what are much more interesting and important are the courses and contours of particular people’s lives. “In the final analysis, you’re not talking about real people when you tell these big, generic stories,” she said. “This is an arena where everybody would be helped a lot by not thinking of this or that as something that happens to everybody or doesn’t happen to everybody. It’s better to think in a textured way about for whom, and under what conditions.”
That fits with psychologists’ job description, which is to study the emotions and development of individuals—usually in small groups, and often in laboratory conditions or controlled environments. In small groups and particular cases, the effects of individuals’ circumstances and personalities determine how people are doing. Isolating the effect of just one statistical cause of happiness or unhappiness seems beside the point, because humans aren’t statistics and overall risk factors tell us little about our own lives. Midlife stressors are real, of course. Ask anyone who is caring for her parents and her kids at the same time. But that is not the same as a predictable, definable syndrome linked to age. And so, by the early 2000s, mainstream psychology had decided that the notion of a midlife crisis was unsupported and uninteresting. The profession moved on.
To this day, a lot of people assume that’s the end of the story. “The Midlife Crisis Is a Total Myth,” read one headline in 2011, over an article by Robin Nixon on the website Live Science. “Worried About a Midlife Crisis? Don’t. There’s No Such Thing,” ran another headline, in 2015, above an article on Psychology Today’s website by Susan Krauss Whitbourne, a respected psychologist. In her article, Whitbourne made a set of seemingly final assertions: that there are “virtually no data to support the assertion that the midlife crisis is a universal experience” (which is true); that it’s usually not a crisis anyway (which is also true); and that she herself, in a study of nearly five hundred people in their thirties, forties, fifties, and sixties (a fairly big sample by the standards of psychology), had found no evidence of midlife crises.
But there turns out to be a different way of thinking about midlife and happiness, which needed to be discovered by a different kind of thinker.
* * *
When I first met Andrew Oswald, in 2015, he was sixty-one years old. He was a bit above average height, balding and graying, but with a slender, athletic build honed by daily two-hour walks. His bearing was erect, his manner reserved, his voice soft, his conversation sharply analytical. He described himself as a “buttoned-up Englishman,” and I could see what he meant. Although we met on a late spring day in his office at the University of Warwick, a campus where formality was neither expected nor observed, he wore a dark blue suit with purple and orange stripes, set off by a cream-colored necktie, brown shoes, and a skinny-brim fedora hat. He likes dressing up, he told me. The office contained several chairs, many books, and a whiteboard covered with scrawled equations—apparently for show, since Oswald assured me he does all his work at home. Breaking the otherwise predominant impression of Vulcan reserve was the laughter with which he periodically punctuated his sentences.
Oswald was born in Bristol, England, in 1953. His father was a formidable academic psychiatrist who moved the family to Edinburgh and Australia and set high expectations for his son. Oswald exceeded them, marrying his first girlfriend, obtaining an Oxford doctorate, writing a dissertation that became mildly famous, and having two children, all while still in his twenties. “I was driven, to put it at its mildest,” he recalled.
In the 1970s, Oswald was fascinated by the miserable economy he saw around him in pre-Thatcher Britain, and by the toll it took on ordinary people. “I came into economics because of the tremendous unemployment and inflation and turbulence in Britain in that era,” he said, adding, with a laugh, “I decided to solve that.” At the time, strange though it seems now, British economists were debating whether unemployment was really a serious problem for the unemployed. Maybe it was often voluntary, the result of people preferring leisure to work. Or maybe it just wasn’t so bad. “The Right, with a capital R, argued then, and have always argued, that people more or less chose unemployment because they had low skills and they were offered benefits and that was the rational thing for them to do,” Oswald said. “This issue has been central to economics for more than a hundred years: whether you should think of unemployment as an equilibrium or as an absolute human disaster.” Oswald thought it was probably an absolute human disaster, a suspicion which steered him to the field of labor economics, the study of jobs and wages—but with an emphasis, in Oswald’s case, on how jobs and wages affect subjective wellbeing. His first published paper, in 1979, was about “relative concern,” the phenomenon—at that point much less well known than it is today—of evaluating one’s own income based on others’. Trade unions, he noticed, were obsessed with relative income: keeping their members on par with or ahead of others. “It just seemed to me intuitively obvious that this is what drove humans.”
In those days, his interest in subjective factors marked him and his work as eccentric. He didn’t care. He told himself what he still tells young scholars: “If everyone likes your work, you can be sure it’s not important. I had a really obdurate streak. I was pretty sure I knew what mattered.”
Also unusual then, though not today, was his belief that data trumps theory. “I’d grown up in the Oxford tradition, where implicitly we were left with the view that we didn’t need data.” He had been trained to suppose that clever economists did mathematics rather than grubby empirical work, a view he now calls a “dangerous and awful thing.” Armed with equations, the young Oswald arrived for a fellowship at Princeton University, presented his mathematical models, and was greeted frostily. “They said, ‘This is fine, Andrew, but where’s the evidence?’ This was hugely educational and painful. What do you mean, where’s the evidence?” Discombobulated, Oswald reassembled himself. “I realized that really clever people used evidence.” He became a “big-data” guy years before the term existed.
Perhaps it was inevitable that a scholar attracted to big data, contrarian stories, and subjective wellbeing would find happiness economics irresistible. “I was concerned about how unemployment had affected people. We discovered data sets, huge data sets, where people were asked about their happiness. It seemed natural to try to understand the patterns.” At that point, Easterlin’s pioneering work was still mostly forgotten, but the profession’s complete lack of interest in subjective wellbeing only made the topic more enticing. “I knew no one would like it, and I thought there was a good chance it would never be published. But it seemed fascinating to me. The issues were obviously of first-order importance. We’re trying to understand human happiness. What’s the role of economic factors and other factors? It’s possibly the most important question in all of social science.”
As Joshua Wolf Shenk has observed in his book Powers of Two: How Relationships Drive Creativity, a lot of the greatest creativity is the result of creative dyads, partnerships in which two very different people complement each other and become a sort of super-thinker or super-creator. John Lennon and Paul McCartney, of the Beatles, are an iconic example; or Steve Jobs and Steve Wozniak, the founders of Apple Computer; or Thomas Jefferson and James Madison, whose political partnership married the ideas of individual freedom and constitutional order. Though each member of the dyad was exceptionally talented, their combination sparked a chemical reaction that created something fantastic.
Around the time he discovered big data sets on happiness, Oswald met his own creativity reagent: David Blanchflower. “He’s a genius,” Oswald said. “A genius with data.”
* * *
At 10 a.m. one June morning, not long after meeting Andrew Oswald, I arrived at Dartmouth College’s economics department, made my way along a creaky hallway, and encountered Oswald’s opposite number—and opposite seemed an understatement. Beckoning me into his office, Blanchflower, then sixty-three, was a big-boned, heavyset man with graying, tousled brown hair, and thick hands that swallowed mine whole when we shook. He had a room-filling, boundingly jovial personality. That day he was dressed in summer mode, wearing a floral print shirt (torn near the tail and not wholly tucked in), along with khaki shorts and sandals; but he seemed to be the sort of person who is often in summer mode, regardless of the season. Conspicuous in his cluttered office was a sofa, where he reclined for a good portion of our interview. Even more conspicuous were three side-by-side computer screens on his desk.
Before I could ask my first question, Danny, as he’s called, launched our interview by announcing that half an hour before I arrived he had opened a data set which he had never before examined. He had run some quick analyses and found the typical U-shaped relationship between life satisfaction and age. “Pretty cool, eh?” he said, as I squinted over his shoulder at rows of digits on the monitor. “I’d literally never looked at it. But there it is!”
This particular data set contained a collection of thirty-seven countries, most of them developed countries like the United States and Denmark, but with a sprinkling of other places like China, Latvia, and Turkey. “This is a really strange bunch of countries,” Blanchflower said, which is why he had not bothered with it before. The data came from a survey in which people were asked, “If you were to consider your life in general, how happy or unhappy would you say you are, on the whole?”
I asked him to show me how he crunched the numbers. “There are only twelve hundred observations for the U.S.,” he said, peering at the data. Not enough for meaningful results. So he selected the countries in Western Europe. The man’s fingers almost literally blurred over the keyboard as he bounced his cursor from screen to screen, pulled up data tabs, specified equation variables, ran regressions. At every point, he accompanied himself with imprecations: “Where is that bloody thing?”; “Oh, come on, Danny!” (“Talking to myself,” he mumbled. “Best way to do it!”) It occurred to me that he was probably doing in a few minutes what might take some analysts hours. Sure enough, a U-shaped curve popped out, its bottom in the early to middle fifties. Then we looked at Eastern Europe; a U again, with the bottom later in middle age, and not as much of a rise afterward as in Western Europe. (“Eastern Europeans are all miserable.”) Next up: developing countries and another U, this time bottoming in the early fifties. He added and removed various statistical controls, or used none at all, but the story rarely changed. Then, clearly enjoying himself, and perhaps also relishing my hopeless struggle to keep up with him in my notes, he started in on stress and anxiety, pulling up a British data set with 305,000 people. It showed that U.K. life satisfaction forms a U shape bottoming at age forty-nine, while anxiety and stress peak at about the same age.
“There are hundreds and hundreds of data files I’ve looked at,” he said, when we came up for air. “Hundreds of them. Basically, I do what you’ve just seen. You get these kinds of plots. I know these people who say you don’t see them. But how hard was that to see?”
Blanchflower wasn’t always a genius with data. His youth, in Brighton, on the south coast of England, was haphazard. No one in the family had attended university, and young Danny did not seem likely to be the first. When he was thirteen, his parents were called in to school because he was at the bottom of the bottom class. In high school, he got more interested and did well enough to be asked to take the exams for Oxford, but instead he went to Leicester University, a middle-tier British institution, where he majored in economics, didn’t work hard, and performed more than adequately with little effort. There followed a period when he was “a bit of a hippie,” doing odd jobs—working as a bouncer in California and as a roadie for a rock band, among other things—and then quitting them when he had earned a little money to travel, at one point trekking overland to Afghanistan. “I kind of went sideways for a while,” he said.
But along the way he discovered an affinity for economics, and particularly for statistics. Finding himself in a dead-end job teaching basic economics at the British equivalent of a community college, he had a “light-bulb-flashing moment” (“It was like, Holy shit, I didn’t know I was that smart!”) and, at twenty-eight, went off to do a master’s degree in economics. At that point, he made up for lost time, completing the master’s and then, in under two years, a PhD. This was in 1970s Britain, where strikes and unemployment left many talented young people out in the cold. His dissertation, like Andrew Oswald’s, was on unions and wages, and his inclination, like Oswald’s, was to listen to data and pay attention to the everyday world. “I call it the economics of walking around,” he told me. “I liked the economics. But I really wanted to try to understand youth unemployment.”
Perhaps it was inevitable, then, that he and Oswald would meet, which they did in the mid-1980s, somewhere in London. They soon agreed to collaborate, not knowing that they would go on to write more papers together than they have bothered to count. “He and I have probably written hundreds of papers together,” Blanchflower said. “We’re complete complements. I do the data, he writes the first draft.” A lot of their work was on wages and labor markets, including one paper in 1988 in which the two of them, seemingly goading each other to new heights of contrarianism, give the academic equivalent of the middle finger to conventional economics by announcing, “This paper contains no mathematical model … nor any econometric analysis.” That in-your-face disclaimer caused consternation among economists who saw it, which Blanchflower and Oswald relished. “We were sticking it up everyone’s nose in 1988.” By the late 1980s, the two of them and another like-minded economist, Andrew Clark, were holed up together at Dartmouth, creating a little pod of activity that focused on Richard Easterlin’s nearly forgotten subject of happiness.
Oswald got interested in happiness first, then tried to interest Blanchflower, who initially was indifferent. “He was trying to persuade me there was something in the happiness data. I said: ‘Yeah, yeah.’ But when you look at it, you find the same unbelievable stability in the happiness data as in the wage data.” To a big-data guy, stable, recurring patterns are a sign that something is going on. So in 1993 Oswald and Andrew Clark and a few others organized a conference on happiness economics. “We advertised it hugely, all over the school,” Oswald recalled. “Put out a hundred chairs. Essentially nobody came.” The few speakers who turned up gave papers only vaguely related to happiness, if at all.
The profession’s indifference, of course, only made the subject more appealing to the contrarian duo, who dived obsessively into the data sets, looking for patterns. One of the things they tested for, not because they expected anything but because the data were there, was age. The same pattern kept cropping up. A 1994 paper by Oswald and Andrew Clark (published in the Royal Economic Society’s Economic Journal), on whether unemployment causes unhappiness (answer: yes, and powerfully), contained the sentence: “There is a U shape in mental wellbeing with respect to age.” In a 1996 paper (in the Journal of Occupational and Organizational Psychology), Oswald, Clark, and a psychologist named Peter Warr found that job satisfaction is “U-shaped in age.”
At that point, the result seemed an intriguing oddity. There was no particularly good theory to explain it and not much by way of confirmation, and so the researchers did not pay much attention to it. “We did what people do, which is just proceed,” Oswald told me. “You’re in the fog of research. That’s how I think about it. You’re just out there.” But the age-happiness relationship persisted, and not just in their own work. Carol Graham, for example, used quite different countries and data sets, but found similar patterns. In 2004, Blanchflower and Oswald had enough data and confidence to pronounce that age, all by itself, plays a role as a determinant of happiness. In “Wellbeing over Time in Britain and the USA,” which was published in the Journal of Public Economics, they found that marriage is very good for happiness, that unemployment is very bad, that life satisfaction had stagnated in Britain and declined in the United States (though the happiness of American blacks had risen), and that relative income matters. They also announced that age had its own, independent effect on life satisfaction: “The exact effect of age upon reported happiness is of interest. It is U-shaped.”
The age effect, they said, held in both the United States and the United Kingdom, after adjusting for key variables like marital status, education, and employment. It was true of both men and women. It did not seem to be the result of changes in social or economic conditions over time, because the same pattern turned up across generations. “Something systematic appears to be at work,” they wrote. “No explanation is available even in the psychology literature.”
The 2004 paper, a clarion public declaration that something was going on with age, made a serious splash. Four years later, in the journal Social Science & Medicine, they published a magnum opus on age and happiness, “Is Wellbeing U-Shaped over the Life Cycle?” Their answer, based on large data sets spanning dozens of countries and including hundreds of thousands of people, among them more than half a million in Europe and America alone, was yes. “We show that wellbeing reaches its minimum around the middle of life,” they wrote. “The regularity is intriguing. The U shape is similar for males and females, and for each side of the Atlantic Ocean (though its minimum is reached a little later among American men).” A couple of dozen or so countries did not show a U pattern, but they were mostly developing countries where the number of people sampled was small. In statistics, small samples make it difficult to find patterns amid the “noise” created by the wide variety of individuals’ specific circumstances. If you imagine a sample of only three people, for example, one might be unemployed at age twenty-five, another might happily remarry at age forty-five, and the third might get cancer at age sixty-five—effects which, in that tiny sample, would completely swamp any effect of age in and of itself.
By 2008, Blanchflower and Oswald were able to cite almost two dozen other papers finding the U. For more evidence, they decided to look at mental health. If life satisfaction was low in middle age, that might imply that depression would be higher, so they obtained a database of about a million Britons between the ages of sixteen and seventy. Sure enough, they found an inverse U—a hill, as it were—with the probability of depression peaking in the mid-forties. In 2012, deepening their case, they looked at data counting prescriptions for antidepressants in twenty-seven European countries. Same result: a hill with the probability of using antidepressants peaking in the mid-forties. The following year, they came up with data on the use of mental-health drugs in two American states, New Hampshire and New Mexico: “Here the highest probability of consuming medication occurs in the age bracket 45–49. Again, a midlife peak is visible in the data.”
In 2010, The Economist, the British newsmagazine, published an article titled “The U-Bend of Life,” which made the phenomenon more widely known. Besides drawing upon the work of Blanchflower and Oswald, the article displayed a graph from a then-new paper by an American psychologist named Arthur A. Stone, of the University of Southern California, and three other authors (Joseph E. Schwartz, Joan E. Broderick, and Angus Deaton). Titled “A Snapshot of the Age Distribution of Psychological Wellbeing in the United States,” the paper drew on a Gallup survey of more than three hundred thousand Americans and found the U shape, bottoming in the late forties and early fifties. Even without statistical adjustments, the pattern was visible to the naked eye.
We’ll come later to what might be going on here. Suffice to say, for now, that the researchers were finding that something is going on—something which earlier research about midlife had missed. Life satisfaction tends to decline gradually after early adulthood, bottom out in middle age, then gradually rebound after. Andrew Oswald, for his part, is not shy about the potential significance of the data. More or less the first thing he said to me during our first conversation was: “I view this as a first-order discovery about human beings that will outlive us by hundreds of years.”
* * *
Not everyone was convinced. There was a problem, potentially, with the U story, one worth recounting because it pushed the research in such an unexpected direction. Remember, the U is a big-data phenomenon. Blanchflower, Graham, Oswald, and the others find it by analyzing data sets spanning tens of thousands or preferably hundreds of thousands, and sometimes millions, of people: oceans of people so large that researchers can use powerful statistical techniques to fish out patterns. The kind of data they are looking at, huge international surveys of happiness, did not exist in the 1960s, when Elliott Jaques floated his midlife-crisis hypothesis. And data aren’t people, which is the reason psychologists weren’t convinced by it. If you want to know how people live their lives, how their satisfaction dips and rises, you need to use much smaller samples, you need to differentiate between individuals, and you usually need to look in detail at their lives. You can’t do big data on personal experience. Or so it seemed.
The U faced another challenge. If in 1975 you had asked a million Americans of different ages how satisfied they were with their lives, you might have retrieved an impressively large data set—but it would have been a snapshot, a statistical photo of how people felt in 1975. Knowing that, say, thirty-year-olds and sixty-year-olds were happier than forty-five-year-olds in 1975 might merely tell you that the three generations had experienced different conditions in their lives. Maybe the forty-five-year-olds, having all grown up during the Great Depression, were scarred by hard economic times, making them generally less positive about life. The example is not hypothetical: the Depression made my father pessimistic and insecure about money, with a proclivity for hoarding. To know that the U curve is really a result of age rather than reflecting what statisticians call a cohort effect, you ideally need to compare people who were born in many different years but who had the same generational experience. Which, of course, is impossible. Or you need to follow individuals throughout the course of their lives—but tracking people through the course of their lives, of necessity, takes a lifetime, and many of the subjects will drop out or disappear, and keeping up with the rest costs a fortune. Even tracking people for a year is pretty hard. Partly as a result, when a few researchers had found and looked at lifetime data, the U didn’t reliably turn up.
In 2002, a young graduate student at the University of Warwick approached Oswald about taking him on as a PhD student. Nick Powdthavee was a Thai immigrant to Britain and had an idea about studying “Buddhist economics.” Instead, Oswald steered him to a data set on the wellbeing of South Africans. Sure enough, he found the U. By the time I met Powdthavee, he was in his late thirties and had published dozens of articles, plus a book on happiness economics. “I’ve always found it,” he told me, when I had coffee with him in a basement café at the London School of Economics and Political Science, where he was then a professor (he subsequently returned to Warwick). “In almost every data set I’ve looked at, there’s this U.”
But, again, the South Africa data offered only a snapshot. Powdthavee wanted a motion picture. So he, Oswald, and a third scholar (Terence Cheng) tackled the problem head-on. They managed to find four data sets in three countries (Australia, Britain, and Germany) that tracked people over time, and to each they then applied a statistical technique that allowed them to look at what they called “within-person changes” in life satisfaction. In all four data sets, they found the U, bottoming at various points in the forties. Here was evidence that the U-shaped pattern is something actually experienced by individuals as they aged. Not all individuals, of course, but enough to validate the pattern. “It exists,” Powdthavee told me, adding, “It’s fundamentally important.”
But how fundamental? Is this a biological thing? Or a cultural thing? A result of outside pressures or of inner design? With those questions in mind, and acting on nothing more than a hunch, Oswald emailed a man named Alexander Weiss.
* * *
Whole careers and important discoveries sometimes turn on quirks. Alex Weiss has poor night vision. An American, he is a comparative psychologist at the University of Edinburgh, in Scotland. In plainer English, he tries to learn more about human psychology by studying the psychology of animals. He did a master’s degree on sea anemones and leopard sharks, and then set out to do a PhD on the psychology of moths, but his night blindness scotched that idea. While he was looking for an alternative, a supervisor asked him to join in some work on chimpanzees.
If you are human, your personality and your general level of happiness are heavily influenced by five basic personality traits, which are, in turn, strongly influenced by your genes. The big five, as they are often called, are neuroticism, extroversion, openness to experience, agreeableness, and conscientiousness. The upbeat, positive traits among the big five often cluster together and reinforce each other to create upbeat, positive personalities, a phenomenon which Weiss and various of his colleagues call covitality. To a large extent, we get our personality structure and basic happiness level from our parents: in humans, subjective wellbeing (happiness, broadly defined) and personality structure not only track closely with each other, but are both quite significantly heritable (to the tune of roughly 50 percent). But how important is the biological component? One place to seek insight is in humans’ closest relatives.
Along with two collaborators, R. Mark Enns and James E. King, Weiss decided to look at chimps in zoos. Lo and behold, when he looked at dominance—a major factor in both chimp personality and wellbeing, and a trait in chimps that maps fairly well to extroversion and low neuroticism in humans—Weiss found that it was both strongly heritable and strongly tied to wellbeing. That result and others like it mirrored the patterns in humans, suggesting a strong biological role in happiness, and in the personality traits that give rise to happiness. “That inspired me,” Weiss said, when we spoke about his primate research. “We can learn a lot about human personality and wellbeing from studying chimpanzees and other apes.”
In the course of his work on primate psychology and personality, Weiss had accomplished two useful things. First, he and James King had developed an assessment of chimp and orangutan happiness. Of course, apes can’t tell us how happy they are, but they express a lot of emotion, of which keepers, who tend to know their charges very well, are quite reliable reporters. Asked, for instance, to assess the apes’ positive or negative mood, how much pleasure they get from social interactions, and how successful they are at achieving their goals, multiple raters gave consistent answers, and their answers matched up with objective indicators measuring traits like physical health. Weiss thus found himself in possession of a lot of information about how apes feel.
Second, Weiss had also learned that, for chimps’ personalities, age matters, and in much the same way that it matters for humans. “Human personality development is characterized by individuals’ becoming more introverted, less competitive, and less emotional, and having greater behavioral controls,” he and James King wrote. “This pattern is largely preserved in chimpanzees.” The finding that age shapes personality in chimps is revealing, because it suggests that some of our personality development—as well as our personality structure—is wired into higher-primate biology. “Chimpanzees aren’t out there getting jobs,” Weiss told me. “You have this very different situation, yet this same basic trajectory is going on. If you adjust for the fact that chimpanzees and orangs are living fifty to fifty-five years, you find the magnitude of these age differences is super, super similar to what we see in humans.” It might not be so surprising, then, to find other similarities between humans’ and apes’ emotional development.
One day, out of the blue, an email arrived from Andrew Oswald, an economist Weiss had never heard of. “He’s like, ‘I see you’ve done this work on wellbeing in chimpanzees and orangutans. My work has found this U shape underlying age and wellbeing, and I’m wondering if you’ve thought to look at this in your apes.’” He hadn’t. The two agreed to meet to talk it over, and soon afterward Weiss began plowing through data on 336 chimpanzees and 172 orangutans, mostly in the United States and Japan but also in Australia, Canada, and Singapore. (The apes lived in zoos, research centers, and a sanctuary.) The results were clear. “Sure enough,” Weiss told me, “there’s good evidence for this U shape in our apes.” The apes hit bottom at an age whose equivalent in humans is between forty-five and fifty.
The resulting paper, “Evidence for a Midlife Crisis in Great Apes Consistent with the U-Shape in Human Wellbeing,” was published in Proceedings of the National Academy of Sciences in 2012. “Our results,” wrote Oswald, Weiss, and a handful of other collaborators (including some primatologists), “imply that human wellbeing’s curved shape is not uniquely human and that although it may be partly explained by aspects of human life and society, its origins may lie partly in the biology we share with closely related great apes.”
The paper created a sensation. Midlife crisis in monkeys! “It was a bit overwhelming at times,” Weiss told me. No one was more surprised than Andrew Oswald, who had not expected his wild hunch to pan out. He wasn’t convinced until the third set of data came in, a moment still fresh in his mind when we met several years later. “I always remember the desk I was sitting at three or four times in my life when I thought, This is going to outlive me. It’s either wrong or it’s absolutely fundamental to humans.”
Good empiricist that he is, Oswald said he will breathe a sigh of relief when someone replicates the apes result. Still, when I asked how he thinks matters stand, he replied: “My betting is this is a completely universal human phenomenon.” By “universal” he meant the tendency is universal, not that everyone will have a painful time in middle age. “I think the underlying process is the same in all humans. I could turn out to be wrong and I’m keeping an open mind, but that’s what the evidence suggests.”
It was the ape study that finally riveted my attention to the happiness curve. Everything suggested that my chronic fortysomething discontent was not about my circumstances, or even, in a sense, about “me,” my conscious, reasoning self. And the apes seemed to seal the deal. I hadn’t necessarily made bad decisions or let myself down or curdled as a human being. I didn’t even need to fully understand why I felt as I did, any more than apes do. In fact, if evolution has for some reason wired into us a tendency to experience discontent in midlife, we might expect not to understand why. When nature sets us up with built-in physiological and psychological processes, she rarely provides built-in insight into how they work. All of which would help explain my compulsive but nonsensical daydreams about walking out on a very good writing job.
* * *
“I’m pretty sure I’ve gone through the U, by the way,” Oswald remarked to me at one point, as if in an afterthought. His marriage split up when he was in his forties. (“I wouldn’t wish that on my worst enemy.”) At that point, the work on happiness did not seem likely to be particularly significant. “I was conscious that the second half of life didn’t look very good. I thought there’d be a general deterioration.”
Blanchflower had his own midlife difficulties. “What did I want out of life? I wanted to be famous. And I got to be a full professor at an Ivy League school, and we wrote The Wage Curve [his and Oswald’s influential book on labor markets], and I was only forty. And we both went: Oh! That’s a bit earlier than I thought! I think that was a midlife crisis. That was like: Do something else.” I asked if he had felt trapped at that time, and Blanchflower chuckled that he hadn’t, but his wife had felt trapped with him. He responded to the stress by buying a snowmobile and a boat and by doubling down on work. “A lot of it was that we just worked all the time. I would work and I would spend time with my kids. I certainly didn’t spend time with my marriage.” His marriage, like Oswald’s, broke up traumatically, and he got hit with a cancer diagnosis. The result was a period which he said can be summed up with a single word: pain.
About the time Andrew Oswald turned fifty, he entered a relationship with his current wife, who turned out to be as right for him as the previous match had been wrong. His perspective on life has changed, and his outlook has improved. “You’re more forgiving of yourself,” he told me. “It’s possible to concentrate on the bits of what you’ve done that work well. You don’t have to obsess about all the bad stuff.”
Blanchflower experienced a similar turnaround. In his early fifties, he told me, “life got better.” His divorce finally came through, after seven excruciating years, and his cancer went into remission. He remarried, happily; regained his health; acquired a column in a prominent newspaper; became a member of the Bank of England’s board of governors. And he became a more relaxed person. “I’m less and less inclined to be a boring academic. I have nothing to prove whatsoever.” He fishes, a hobby he talks about with an economist’s rigor. “I go down about seven thirty at night as the sun goes down. Fishing is better if it’s calm, but if it’s calm, the bugs come, so you have a trade-off.”
Did Danny’s and Andrew’s remarriages improve their mind-sets? Did their mind-sets improve their marriages? Reflecting later on my conversations with the two of them, it occurred to me that neither I nor they would ever understand how the changing slope of the happiness curve and the changing contours of their lives and choices had interacted. No data set can solve that mystery. I do know that I encountered two happy men whose lives had rounded a bend. The first time I spoke to Andrew Oswald, I mentioned that I was in my early fifties and that my contentment seemed to be on the rise. He exclaimed, “Just wait until you’re sixty!”